Hyperspectral data are the spectral response of landcovers from different spectral bands and different band sets can be treated as different views of landcovers, which may contain different structure information. Therefore, multiview graphs ensemble-based graph embedding is proposed to promote the performance of graph embedding for hyperspectral image classification. By integrating multiview graphs, more affluent and more accurate structure information can be utilized in graph embedding to achieve better results than traditional graph embedding methods. In addition, the multiview graphs ensemble-based graph embedding can be treated as a framework to be extended to different graph-based methods. Experimental results demonstrate that the proposed method can improve the performance of traditional graph embedding methods significantly.

In automatic target recognition systems based on the use of inverse synthetic aperture radar (ISAR) images, it is essential to obtain unbiased and accurate scaled two-dimensional target images in the range-cross range domain. To accomplish this, the modulus of the target effective rotation vector, which is generally unknown for noncooperative targets, must be estimated. This letter proposes an efficient method for estimating the cross-range scaling factor and significantly improving cross-range resolution based on the second-order local polynomial Fourier transform. The estimation requires solving a series of one-dimensional optimizations of a kurtosis objective. Simulations show the proposed approach to be effective and able to accurately estimate the scaling factor in the presence of noise.

Terrestrial laser scanning (TLS) is a noninvasive technique to monitor surface conditions and morphological characteristics of structures and has been successfully introduced to the regular inspection and maintenance of metro tunnels. To accurately analyze the deformation and structural conditions of a metro tunnel, nonliner points (e.g., outliers and accessories) should be detected and eliminated. Nevertheless, the accessories are attached very closely to the liner and cannot be thoroughly eliminated by three-dimensional (3D) geometric information. This study proposes to separate the liner and accessories by combining TLS geometric and radiometric information. A refitted mobile Faro Focus^3D X330 system is used for data collection of a new-built metro tunnel in Hangzhou, China. The results show that the corrected intensity data are an effective physical criterion and a complementary data source to remove accessories that cannot be eliminated by geometric data. After the removal of accessories by geometric and radiometric data, the remaining liner points can accurately reflect the actual structural and deformation conditions of metro tunnels.

The airborne frequency-modulated continuous-wave synthetic aperture radar presents an enormous technical challenge on the design of data storage system due to its characteristics of high-data rate, small size, light weight, and low-power consumption. There are two main problems for the high-speed storage under the miniature requirement. One is the unpredictable response time of the flash translation layer in the CompactFlash card. The other is the relatively long response time of the file system. This paper designs a data storage system in a real-time signal processor. Two techniques called configurable buffer structure and FPFQA (FAT pre- and FDT quasiallocation) are presented to overcome these two problems. The evaluated performance indicates that the size, power consumption, and weight meet the miniature requirement, while the function of the high-speed data storage with approximately 121  MB/s storage speed and real-time file management are realized.

There are many similarities between Brillouin optical time-domain analysis and phase-sensitive optical time-domain reflectometer in system topological structure. A multifunctional distributed optical fiber sensing system based on their similarities and the combination of their topological structure is presented. The system can monitor strain and temperature, as well as detect and locate the intrusions or disturbances along the sensing fiber. Suppose that X is the width of the optical pulse incident into the sensing fiber with the unit of nanosecond and Y is the minimum detectable distance with the unit of meter between two intrusions that are being detected simultaneously. The experiment results show that y changes linearly with x according to the equation: y=0.103x−0.452. By optimizing device parameters of the system, strain measurement accuracy of 3.17  μϵ and the temperature measurement accuracy of 0.45°C have been realized when the optical pulse width is 30 ns. On the other hand, two intrusions have been detected and located simultaneously with this multifunctional system. The results show that the pulse width and the minimum distance between two intrusions (L_min) have a linear relationship. The theoretical analyses and the experiment results show it is applicable to use this multifunctional system for strain and temperature monitoring, and intrusion detecting.

Owing to the rapid development of earth observation technology, the volume of spatial information is growing rapidly; therefore, improving query retrieval speed from large, rich data sources for remote-sensing data management systems is quite urgent. A global subdivision model, geographic coordinate subdivision grid with one-dimension integer coding on 2ⁿ-tree, which we propose as a solution, has been used in data management organizations. However, because a spatial object may cover several grids, ample data redundancy will occur when data are stored in relational databases. To solve this redundancy problem, we first combined the subdivision model with the spatial array database containing the inverted index. We proposed an improved approach for integrating and managing massive remote-sensing data. By adding a spatial code column in an array format in a database, spatial information in remote-sensing metadata can be stored and logically subdivided. We implemented our method in a Kingbase Enterprise Server database system and compared the results with the Oracle platform by simulating worldwide image data. Experimental results showed that our approach performed better than Oracle in terms of data integration and time and space efficiency. Our approach also offers an efficient storage management system for existing storage centers and management systems.

The morphology of road cut slopes, such as length and high slopes, is one of the most prevalent causes of landslides and terrain stability troubles. Digital elevation models (DEMs) and orthoimages are used for land management purposes. Two flights with different orientations with respect to the target surface were planned, and four photogrammetric projects were carried out during these flights to study the image orientation effects. Orthogonal images oriented to the cut slope with only sidelaps were compared to the classical vertical orientation, with sidelapping, endlapping, and both types of overlapping simultaneously. DEM and orthoimages obtained from the orthogonal project showed smaller errors than those obtained from the other three photogrammetric projects, with the first one being much easier to manage. One additional flight and six photogrammetric projects were used to establish an objective criterion to locate the three ground control points for georeferencing and rectification DEMs and orthoimages. All possible sources of errors were evaluated in the DEMs and orthoimages.

Target detection is an important issue in hyperspectral remote sensing image processing. This paper proposes a method for hyperspectral target detection using data field theory to simulate the data interaction in hyperspectral images (HSIs). We then build a data field model to unify spectral and spatial information. Furthermore, a support vector detector based on a data field model is proposed. Compared with traditional methods, our method achieves superior performance for hyperspectral target detection, and it describes a target class with a more accurate and flexible high potential region. Moreover, in contrast to traditional hyperspectral detectors, the proposed method achieves integrated spectral–spatial target detection and shows superior robustness to signal-noise-ratio decline and spectral resolution degradation. The experimental results show that our method is more accurate and efficient for target detection problems in HSIs.

An optimal radar waveform-design method is proposed to detect moving targets in the presence of clutter and noise. The clutter is split into moving and static parts. Radar-moving target/clutter models are introduced and combined with Neyman–Pearson criteria to design optimal waveforms. Results show that optimal waveform for a moving target is different with that for a static target. The combination of simple-frequency signals could produce maximum detectability based on different noise-power spectrum density situations. Simulations show that our algorithm greatly improves signal-to-clutter plus noise ratio of radar system. Therefore, this algorithm may be preferable for moving target detection when prior information on clutter and noise is available.

This paper presents a point cloud optimization method of low-altitude remote sensing image based on least square matching (LSM). The proposed method is designed to be especially effective for addressing the conundrum of stereo matching on the discontinuity of architectural structures. To overcome the error matching and blur on building discontinuities in three-dimensional (3-D) reconstruction, a pair of mutually perpendicular patches is set up for every point of object discontinuities instead of a single patch. Then an error equation is built to compute the optimal point according to the LSM method, space geometry relationship, and collinear equation constraint. Compared with the traditional patch-based LSM method, the proposed method can achieve higher accuracy 3-D point cloud data and sharpen the edge. This is because a geometric mean patch in patch-based LSM is the local tangent plane of an object’s surface. Using a pair of mutually perpendicular patches instead of a single patch evades the problem that the local tangent plane on the discontinuity of a building did not exist and highlights the edges of buildings. Comparison studies and experimental results prove the high accuracy of the proposed algorithm in low-altitude remote sensing image point cloud optimization.

To obtain a complete representation of scene information in high spatial resolution remote sensing scene images, an increasing number of studies have begun to pay attention to the multiple low-level feature types-based bag-of-visual-words (multi-BOVW) model, for which the two-phase classification-based multi-BOVW method is one of the most popular approaches. However, this method ignores the information of feature significance among different feature types in the score-level fusion stage, thus affecting the classification performance of the multi-BOVW methods. To address this limitation, a feature significance-based multi-BOVW scene classification method was proposed, which integrates the information of feature separating capabilities among different scene categories into the traditional two-phase classification-based score-level fusion framework, realizing different treatments for different feature channels in classifying different scene categories. Experimental results show that the proposed method outperforms the traditional score-level fusion-based multi-BOVW methods and effectively explores the feature significance information in multiclass remote sensing image scene classification tasks.

In this work, the issue of robust waveform optimization is addressed in the presence of clutter to improve the worst-case estimation accuracy for collocated multiple-input multiple-output (MIMO) radar. Robust design is necessary due to the fact that waveform design may be sensitive to uncertainties in the initial parameter estimates. Following the min–max approach, the robust waveform covariance matrix design is formulated here on the basis of Cramér–Rao Bound to ease this sensitivity systematically for improving the worst-case accuracy. To tackle the resultant complicated and nonlinear problem, a new diagonal loading (DL)-based iterative approach is developed, in which the inner optimization problem can first be decomposed to some independent subproblems by using the Hadamard’s inequality, and then these subproblems can be reformulated into convex issues by using DL method, as well as the outer optimization problem can also be relaxed to a convex issue by translating the nonlinear function into a linear one, and, hence, both of them can be solved very effectively. An optimal solution to the original problem can be obtained via the least-squares fitting of the solution acquired by the iterative approach. Numerical simulations show the efficiency of the proposed method.

For synthetic aperture radar (SAR), ground moving target (GMT) imaging necessitates the compensation of the additional azimuth modulation contributed by the unknown movement of the GMT. That is to say, it is necessary to estimate the Doppler parameters of the GMT without a priori knowledge of the GMT’s motion parameters. This paper presents a Doppler parameter and velocity estimation method to refocus the GMT from its smeared response in SAR image. The main idea of this method is that an azimuth reference function is constructed to do the correlation integral with the azimuth signal of the GMT. And in general, the Doppler parameters of the presumed azimuth reference function are different from those of the GMT’s azimuth signal since the velocity parameters of the GMT are unknown. Therefore, the correlation operation referred to here is actually mismatched, and the processing result of is shifted and defocused. The shifted and defocused result is utilized to get the real Doppler parameters and the velocity parameters of the GMT. One advantage of this method is that it is a nonsearching method. Another advantage is that both the Doppler centroid and the Doppler frequency rate of the GMT can be simultaneously estimated according to the relationships between the Doppler parameters and the smeared response of the GMT. In addition, the velocity of the GMT can also be obtained based on the estimated Doppler parameters. Numerical simulations and experimental data processing verify the validity of the method proposed.

We investigate multiple-input multiple-output (MIMO) synthetic aperture radar (SAR) scheme with an ultra-low side-lobe ratio performance for using complete complementary sequences (CCS) waveform. It is shown that the CCS can be applied in a MIMO SAR system to obtain high resolution in range direction. In addition, the azimuth channel scheme with the multiple subsequences of CCS is also considered. First, we introduce the concept of CCS and establish the transmitter model based on CCS in MIMO SAR system. Then, we propose the corresponding imaging algorithm for employing the CCS to accurately focus the raw data of the MIMO SAR system and derive the range cell migration correction. Considering the signal complexity, we make use of the CCS pairs to analyze the MIMO SAR imaging algorithm. Finally, simulation results demonstrate the validity and feasibility of the proposed MIMO SAR scheme and the potential advantage of CCS.

In order to suppress multiple mainlobe interferences and sidelobe interferences simultaneously, a mainlobe interference suppression algorithm is proposed. In this algorithm, the number of mainlobe interferences is estimated through a matrix filter at first. Then, the eigenvectors associated with mainlobe interference are determined and the eigen-projection matrix can be calculated. Next, the sidelobe-interference-plus-noise covariance matrix is reconstructed through eigenvalue replacement procedure. Finally, we can get the adaptive weight vector. Simulation results demonstrate the effectiveness of the proposed method when multiple mainlobe interferences exist.

Long-time coherent integration is an effective means to improve the radar detection ability of high-speed maneuvering targets with jerk motion. However, the range migration (RM) and Doppler frequency migration (DFM) have a great impact on the integration performance. To overcome these problems, a unique method, called the second-order keystone transform modified integrated cubic phase function (SKT-MICPF), is proposed. In this method, the velocity compensation and SKT are jointly employed to correct the RM. After the RM correction, the azimuth echoes of a range cell where a target is located can be modeled as a cubic phase signal (CPS), whose chirp rate (CR) and quadratic CR are related to the target’s radial acceleration and jerk, respectively. Thereafter, an effective parameters’ estimation algorithm for CPS, called MICPF, is proposed and applied to compensate the DFM. After that, coherent integration and target detection are accomplished via the fast Fourier transform and constant false alarm rate technique, successively. Compared with the improved axis rotation discrete chirp Fourier transform, the SKT-MICPF achieves close detection performance, but greatly reduces the computational complexity. The results of simulation and real radar data demonstrate the validity of the proposed algorithm.

Biomass is one significant biophysical parameter of a forest ecosystem, and accurate biomass estimation on the regional scale provides important information for carbon-cycle investigation and sustainable forest management. In this study, Landsat satellite imagery data combined with field-based measurements were integrated through comparisons of five regression approaches [stepwise linear regression, K-nearest neighbor, support vector regression, random forest (RF), and stochastic gradient boosting] with two different candidate variable strategies to implement the optimal spatial above-ground biomass (AGB) estimation. The results suggested that RF algorithm exhibited the best performance by 10-fold cross-validation with respect to R² (0.63) and root-mean-square error (26.44  ton/ha). Consequently, the map of estimated AGB was generated with a mean value of 89.34  ton/ha in northwestern Zhejiang Province, China, with a similar pattern to the distribution mode of local forest species. This research indicates that machine-learning approaches associated with Landsat imagery provide an economical way for biomass estimation. Moreover, ensemble methods using all candidate variables, especially for Landsat images, provide an alternative for regional biomass simulation.

“Subpixel-based downsampling” is an approach that can implicitly enhance perceptible image resolution of a downsampled image by managing subpixel-level representation preferably with individual pixel. A subpixel-level representation for color image sample at edge region and color image representation is focused with the problem of directional filtration based on horizontal and vertical orientations using colorimetric color space with the help of saturation and desaturation pixels. A diagonal tracing algorithm and an edge preserving approach with colorimetric color space were used for color image enhancement. Since, there exist high variations at the edge regions, it could not be considered as constant or zero, and when these variations are random the need to compensate these to minimum value and then process for image representation. Finally, the results of the proposed method show much better image information as compared with traditional direct pixel-based methods with increased luminance and chrominance resolutions.

A scheme is presented for ground moving target indication with multichannel synthetic aperture radar (SAR). After the effects of different phase centroids are compensated, the images from different channels are used to form an interferogram. Then the probability distribution function of the interferometric phase is estimated as a function of the interferometric magnitude, thus the detection threshold is derived as a function of the interferometric magnitude. Since the detection threshold is magnitude dependent, this scheme has a better performance than the conventional schemes, especially for slow, weak moving targets. The effectiveness of this scheme is verified by both simulated and real SAR data.

Accurate extraction of urban impervious surface data from high-resolution imagery remains a challenging task because of the spectral heterogeneity of complex urban land-cover types. Since the high-resolution imagery simultaneously provides plentiful spectral and spatial features, the accurate extraction of impervious surfaces depends on effective extraction and integration of spectral–spatial multifeatures. Different features have different importance for determining a certain class; traditional multifeature fusion methods that treat all features equally during classification cannot utilize the joint effect of multifeatures fully. A fusion method of distance metric learning (DML) and support vector machines is proposed to find the impervious and pervious subclasses from Chinese ZiYuan-3 (ZY-3) imagery. In the procedure of finding appropriate spectral and spatial feature combinations with DML, optimized distance metric was obtained adaptively by learning from the similarity side-information generated from labeled samples. Compared with the traditional vector stacking method that used each feature equally for multifeatures fusion, the approach achieves an overall accuracy of 91.6% (4.1% higher than the prior one) for a suburban dataset, and an accuracy of 92.7% (3.4% higher) for a downtown dataset, indicating the effectiveness of the method for accurately extracting urban impervious surface data from ZY-3 imagery.

The elevation image quality of tomographic synthetic aperture radar (TomoSAR) data depends mainly on the elevation aperture size, number of baselines, and baseline distribution. In TomoSAR, due to the restricted number of baselines with irregular distributions, the elevation imaging quality is always unacceptable using the conventional spectral analysis approach. Therefore, for a given limited number of irregular baselines, the completion of data for the unobserved virtual uniform baseline distribution should be addressed to improve the spectral analysis-based TomoSAR reconstruction quality. We propose an L_q(0<q≤1) regularization-based unobserved baselines’ data estimation method for TomoSAR, which uses the geometric imaging relationship between the observed and unobserved baseline distributions. In the proposed method, we first estimate the transformation matrix between the acquisitions and the data of virtual uniform baseline distribution by solving an optimization problem, before calculating the data for virtual baseline distribution based on the acquisitions and the transformation matrix. Finally, the elevation reflectivity function is recovered using the spectral analysis method based on the estimated data. Compared with the reconstructed results only based on the limited irregular acquisitions, the image recovered using the dataset with a virtual uniform baseline distribution can improve the elevation image quality in an efficient manner.

An inverse synthetic aperture radar (ISAR) high-precision compensation method is proposed based on coherent processing of intermediate frequency direct sampling data. First, the compensation of high-speed movement is performed by a modified linear frequency modulation matched filter during the pulse compression. The motion trajectory in the down-range direction is then reconstructed by compensation of window sampling difference of each pulse. Modified envelope correlation is applied to calculate the range profile shift between each pulse and the first one. Polynomial fitting is adopted to accurately estimate the motion characteristics. Subsequently, coherent processing is applied by combining range alignment and initial phase compensation. The migration through range cells correction can be then realized by keystone transform to the highly coherent data. Consequently, ISAR images with high quality are achieved. Experimental results on simulated and real data have demonstrated the validity of the proposed method.

The most challenging aspect of through-wall imaging is the presence of the wall, which suppresses the detection and localization to the obscured scatterers. Moreover, wall parameters are not known in practice; how to detect the target without prior knowledge of the wall is becoming important. Therefore, an approach is proposed to solve this problem in this paper. First, an effective clutter mitigation method based on singular value decomposition is proposed to achieve the target scattering fields. After a number of data pairs that consist of the target position and its scattering fields have been collected, the through-wall detection problem can be resolved by extracting a nonlinear relationship between them. In this way, the presence of the wall is automatically included in the nonlinear relationship, which is obtained through a training phase using a support vector machine. For a detection task, the position of the target can be estimated from this nonlinear relationship. The whole detection procedure does not require the prior knowledge of the wall. Also, it is shown that the proposed method is effective. Moreover, the impacts of training samples and signal-to-noise ratio on detection accuracy are analyzed.

By exploiting the sparsity of radar target image, it is hopeful to obtain a high-resolution target image in multiple-input-multiple-output (MIMO) radar via a sparse representation (SR) method. However, for the three-dimensional (3-D) imaging, the conventional SR method has to convert the 3-D problem into the one-dimensional (1-D) problem. Thus, it will inevitably impose a heavy burden on the storage and computation. A multidimensional smoothed L0 (MD-SL0) algorithm is proposed based on the conventional smoothed L0 algorithm. The proposed MD-SL0 can directly apply to the multidimensional SR problem without transforming to the 1-D case. As a result, a MIMO radar 3-D imaging method via MD-SL0 is achieved with high computation efficiency and low storage burden. Finally, the effectiveness of the method is validated by the results of comparative experiments.

This paper aims to implement an iterative fuzzy edge detection (IFED) method on blurred satellite images. Some degradation effects such as atmospheric effects, clouds and their shadows, atmospheric aerosols, and fog remarkably decline the quality satellite images. Hence, some processes such as enhancement and edge detection in satellite images are challenging. One group of methods that can deal with these effects is fuzzy logic methods. Therefore, IFED method was applied in this work on the subimages of the Ikonos, Landsat 7, and SPOT 5 satellite images, contaminated by aforementioned effects. Such as most FED methods, IFED has two components: enhancement and edge detection. In this context, a six-step iterative method, using the if-then-else mechanism, was implemented on the images to perform fuzzy enhancement, and subsequently, edge detection was done. To evaluate the merit of the enhancement and select the best number of iterations, edge gray-value rate criterion was applied. The peak signal-to-noise ratio (PSNR) is applied for the quantitative evaluation of the IFED method. The results of IFED, in comparison with some prior edge detection methods, showed higher PSNR values and a high performance in the edge detection of the earth features in the blurred satellite images.

Image smear, produced by the shutter-less operation of full-frame charge-coupled device (CCD) sensors, greatly affects the performance of target detection, the centering accuracy, and visual magnitude estimation. We study the operation principle of full-frame CCDs, analyze the cause and properties of smear effect, and propose a smear removal algorithm for star images of full-frame CCDs. The proposed method locates the smears and extracts the rough profiles of the smeared stars by finding the conditional extrema. Then Gaussian fitting is applied to accurately extract the stars, in order to maintain the integrity of star images while minimizing the smear effect. The extraction of smears and stars requires parameters such as the size of the CCD, the integration time and the readout time, as well as the estimation of background noise. We assess the performance of our scheme with real observed data. The experimental results show that the proposed scheme improves the average signal-to-noise ratio of the images by about 22%, presenting better smear removal performance compared with several published methods. The limitation of the proposed algorithm includes the difficulty of distinguishing between two very close stars displaying the gray level of a single peak and overestimation of the background noise may also influence the performance of the algorithm.

Doppler radar is a cost-effective tool for moving target tracking, which can support a large range of civilian and military applications. A modified linear predictive coding (LPC) approach is proposed to increase the target localization accuracy of the Doppler radar. Based on the time-frequency analysis of the received echo, the proposed approach first real-time estimates the noise statistical parameters and constructs an adaptive filter to intelligently suppress the noise interference. Then, a linear predictive model is applied to extend the available data, which can help improve the resolution of the target localization result. Compared with the traditional LPC method, which empirically decides the extension data length, the proposed approach develops an error array to evaluate the prediction accuracy and thus, adjust the optimum extension data length intelligently. Finally, the prediction error array is superimposed with the predictor output to correct the prediction error. A series of experiments are conducted to illustrate the validity and performance of the proposed techniques.

Land cover classification based on remote sensing imagery is an important means to monitor, evaluate, and manage land resources. However, it requires robust classification methods that allow accurate mapping of complex land cover categories. Random forest (RF) is a powerful machine-learning classifier that can be used in land remote sensing. However, two important parameters of RF classification, namely, the number of trees and the number of variables tried at each split, affect classification accuracy. Thus, optimal parameter selection is an inevitable problem in RF-based image classification. This study uses the genetic algorithm (GA) to optimize the two parameters of RF to produce optimal land cover classification accuracy. HJ-1B CCD2 image data are used to classify six different land cover categories in Changping, Beijing, China. Experimental results show that GA-RF can avoid arbitrariness in the selection of parameters. The experiments also compare land cover classification results by using GA-RF method, traditional RF method (with default parameters), and support vector machine method. When the GA-RF method is used, classification accuracies, respectively, improved by 1.02% and 6.64%. The comparison results show that GA-RF is a feasible solution for land cover classification without compromising accuracy or incurring excessive time.

A spectrum reconstruction algorithm based on space–time adaptive processing (STAP) can effectively suppress azimuth ambiguity for multichannel synthetic aperture radar (SAR) systems in azimuth. However, the traditional STAP-based reconstruction approach has to estimate the covariance matrix and calculate matrix inversion (MI) for each Doppler frequency bin, which will result in a very large computational load. In addition, the traditional STAP-based approach has to know the exact platform velocity, pulse repetition frequency, and array configuration. Errors involving these parameters will significantly degrade the performance of ambiguity suppression. A modified STAP-based approach to solve these problems is presented. The traditional array steering vectors and corresponding covariance matrices are Doppler-variant in the range-Doppler domain. After preprocessing by a proposed phase compensation method, they would be independent of Doppler bins. Therefore, the modified STAP-based approach needs to estimate the covariance matrix and calculate MI only once. The computation load could be greatly reduced. Moreover, by combining the reconstruction method and a proposed adaptive parameter estimation method, the modified method is able to successfully achieve multichannel SAR signal reconstruction and suppress azimuth ambiguity without knowing the above parameters. Theoretical analysis and experiments showed the simplicity and efficiency of the proposed methods.

As far back as early 15th century during the reign of the Ming Dynasty (1368 to 1634 AD), Gomantong cave in Sabah (Malaysia) has been known as one of the largest roosting sites for wrinkle-lipped bats (Chaerephon plicata) and swiftlet birds (Aerodramus maximus and Aerodramus fuciphagus) in very large colonies. Until recently, no study has been done to quantify or estimate the colony sizes of these inhabitants in spite of the grave danger posed to this avifauna by human activities and potential habitat loss to postspeleogenetic processes. This paper evaluates the transferability of a hybrid optimization image analysis-based method developed to detect and count cave roosting birds. The method utilizes high-resolution terrestrial laser scanning intensity image. First, segmentation parameters were optimized by integrating objective function and the statistical Taguchi methods. Thereafter, the optimized parameters were used as input into the segmentation and classification processes using two images selected from Simud Hitam (lower cave) and Simud Putih (upper cave) of the Gomantong cave. The result shows that the method is capable of detecting birds (and bats) from the image for accurate population censusing. A total number of 9998 swiftlet birds were counted from the first image while 1132 comprising of both bats and birds were obtained from the second image. Furthermore, the transferability evaluation yielded overall accuracies of 0.93 and 0.94 (area under receiver operating characteristic curve) for the first and second image, respectively, with p value of <0.0001 at 95% confidence level. The findings indicate that the method is not only efficient for the detection and counting cave birds for which it was developed for but also useful for counting bats; thus, it can be adopted in any cave.

NASA has been planning a hyperspectral infrared imager mission which will provide global coverage using a hyperspectral imager with 60-m resolution. In some practical applications, such as special crop monitoring or mineral mapping, 60-m resolution may still be too coarse. There have been many pansharpening algorithms for hyperspectral images by fusing high-resolution (HR) panchromatic or multispectral images with low-resolution (LR) hyperspectral images. We propose an approach to generating HR hyperspectral images by fusing high spatial resolution color images with low spatial resolution hyperspectral images. The idea is called hybrid color mapping (HCM) and involves a mapping between a high spatial resolution color image and a low spatial resolution hyperspectral image. Several variants of the color mapping idea, including global, local, and hybrid, are proposed and investigated. It was found that the local HCM yielded the best performance. Comparison of the local HCM with <10 state-of-the-art algorithms using five performance metrics has been carried out using actual images from the air force and NASA. Although our HCM method does not require a point spread function (PSF), our results are comparable to or better than those methods that do require PSF. More importantly, our performance is better than most if not all methods that do not require PSF. After applying our HCM algorithm, not only the visual performance of the hyperspectral image has been significantly improved, but the target classification performance has also been improved. Another advantage of our technique is that it is very efficient and can be easily parallelized. Hence, our algorithm is very suitable for real-time applications.

For better using of inverse synthetic aperture radar (ISAR) images of ship targets, it is more desirable to select a proper imaging time to obtain high quality top-view or side-view images. However, optimum imaging time selection is not robust enough for the restriction of traditional geometric feature extraction methods. In our study, we propose a method based on the geometric features and gradient maximization. First, we select the imaging instant from radar echoes by the centerline and mainmast of the ship. In this part, we propose a geometric features extraction method to improve the robustness of instant selection in different scenarios. Then, an image gradient maximization is employed to estimate the period for ISAR imaging. Finally, experimental results of both simulated and real signals are provided to demonstrate the effectiveness and practicability of the algorithm.

For the large-area snow depth (SD) data sets with high spatial resolution in the Altay region of Northern Xinjiang, China, we present a deterministic ensemble Kalman filter (DEnKF)-albedo assimilation scheme that considers the common land model (CoLM) subgrid heterogeneity. In the albedo assimilation of DEnKF-albedo, the assimilated albedos over each subgrid tile are estimated with the MCD43C1 bidirectional reflectance distribution function (BRDF) parameters product and CoLM calculated solar zenith angle. The BRDF parameters are hypothesized to be consistent over all subgrid tiles within a specified grid. In the SCF assimilation of DEnKF-albedo, a DEnKF combining a snow density-based observation operator considers the effects of the CoLM subgrid heterogeneity and is employed to assimilate MODIS SCF to update SD states over all subgrid tiles. The MODIS SCF over a grid is compared with the area-weighted sum of model predicted SCF over all the subgrid tiles within the grid. The results are validated with in situ SD measurements and AMSR-E product. Compared with the simulations, the DEnKF-albedo scheme can reduce errors of SD simulations and accurately simulate the seasonal variability of SD. Furthermore, it can improve simulations of SD spatiotemporal distribution in the Altay region, which is more accurate and shows more detail than the AMSR-E product.

Conventional munitions are not guided with sensors and therefore miss the target, particularly if the target is mobile. The miss distance of these munitions can be decreased by incorporating sensors to detect the target and guide the munition during flight. This paper is concerned with a precision guided munition equipped with an infrared (IR) sensor and a millimeter wave radar (MmW). Three-dimensional flight of the munition and its pitch and yaw motion models are developed and simulated. The forward and lateral motion of a target tank on the ground is modeled as two independent second-order Gauss–Markov processes. To estimate the target location on the ground and the line-of-sight (LOS) rate to intercept it, an extended Kalman filter is composed whose state vector consists of cascaded state vectors of missile dynamics and target dynamics. The LOS angle measurement from the IR seeker is by centroiding the target image in 40 Hz. The centroid estimation of the images in the focal plane is at a frequency of 10 Hz. Every 10 Hz, centroids of four consecutive images are averaged, yielding a time-averaged centroid, implying some measurement delay. The miss distance achieved by including image processing delays is 1.45 m.

Improving forecasts of salinity from coastal hydrodynamic models would further our predictive capacity of physical, chemical, and biological processes in the coastal ocean. However, salinity is difficult to estimate in coastal and estuarine waters at the temporal and spatial resolution required. Retrieving sea surface salinity (SSS) using satellite ocean color radiometry may provide estimates with reasonable accuracy and resolution for coastal waters that could be assimilated into hydrodynamic models to improve SSS forecasts. We evaluated the applicability of satellite SSS retrievals from two algorithms for potential assimilation into National Oceanic and Atmospheric Administration’s Chesapeake Bay Operational Forecast System (CBOFS) hydrodynamic model. Of the two satellite algorithms, a generalized additive model (GAM) outperformed that of an artificial neural network (ANN), with mean bias and root-mean-square error (RMSE) of 1.27 and 3.71 for the GAM and 3.44 and 5.01 for the ANN. However, the RMSE for the SSS predicted by CBOFS (2.47) was lower than that of both satellite algorithms. Given the better precision of the CBOFS model, assimilation of satellite ocean color SSS retrievals will not improve CBOFS forecasts of SSS in Chesapeake Bay. The bias in the GAM SSS retrievals suggests that adding a variable related to precipitation may improve its performance.

Multiresolution segmentation and rule-based classification techniques are used to classify objects from very high-resolution satellite images of urban areas. Custom rules are developed using different spectral, geometric, and textural features with five scale parameters, which exploit varying classification accuracy. Principal component analysis is used to select the most important features out of a total of 207 different features. In particular, seven different object types are considered for classification. The overall classification accuracy achieved for the rule-based method is 95.55% and 98.95% for seven and five classes, respectively. Other classifiers that are not using rules perform at 84.17% and 97.3% accuracy for seven and five classes, respectively. The results exploit coarse segmentation for higher scale parameter and fine segmentation for lower scale parameter. The major contribution of this research is the development of rule sets and the identification of major features for satellite image classification where the rule sets are transferable and the parameters are tunable for different types of imagery. Additionally, the individual objectwise classification and principal component analysis help to identify the required object from an arbitrary number of objects within images given ground truth data for the training.

Tharparkar is an arid region in the southeastern province of Sindh, Pakistan, and experienced drought as a regular phenomenon in the past. The complex nature of drought and sparsely located network of met stations handicapped reliable spatial and temporal analysis of drought severity across Tharparkar. Freely available tropical rainfall measuring mission rainfall satellite data and moderate-resolution imaging spectroradiometer normalized difference vegetation index (NDVI) satellite data fulfilled this gap and were used to generate drought indices. Commonly used NDVI and NDVI anomalies pose problems when compared with standardized meteorological drought indices such as standardized precipitation index (SPI) and standardized precipitation and evapotranspiration index (SPEI) for drought characterization. This study compared standardized vegetation index (SVI) with traditionally used, i.e., SPI and SPEI, for modeling drought severity in the arid and fragile agro-ecosystem of Tharparkar. SVI significantly correlated with standardized meteorological drought indices (SPI and SPEI) and revealed vegetation dynamics under rainfall and temperature variations. Weighted overlay analysis in geographical information systems depicted an accurate onset of the 2014 drought. This study provides useful information for drought characterization that can be used for drought monitoring and early warning systems in data scarce, arid, and semiarid regions.

Spectral unmixing is an important procedure to exploit relevant information from remotely sensed hyperspectral images. Each pixel spectrum is unmixed to some pure constitutions, endmembers, and their fractional values and abundances. The aim of this study is to improve neural network (NN)-based unmixing methods, which consist of linearly extracting endmembers, and nonlinearly estimating of abundances. In this seminonlinear method, we use fractional endmembers as inputs and pixel spectrum as output in a multilayer perceptron. Two types of samples are used as training data: (1) the most similar samples to each endmember (core of class) and (2) the most dissimilar samples to all endmembers (border of classes). After training of the network, an optimization step is proposed to model pixel spectrum forwardly. This step starts with initial abundances and optimizes them to obtain a desired pixel spectrum. Application of this method on Cuprite data shows a promising reconstructed image with an average root-mean-square error (RMSE) value of 0.0084. To evaluate the presented algorithm, it is compared with one linear and two nonlinear unmixing methods. The average RMSE values and study of error distribution showed that the proposed method can be accounted as a better selection.

Until now, there have been only a few studies that have made estimates of the woody aboveground biomass (AGB) in an area of agroforestry using remote sensing technology. The woody AGB density was estimated using individual tree analysis (ITA) that incorporated tree species information using a combination of airborne light detection and ranging (LiDAR) and compact airborne spectrographic imagery acquired over a typical agroforestry in northwestern China. First, a series of improved LiDAR processing algorithms was applied to achieve individual tree segmentation, and accurate plot-level canopy heights and crown diameters were obtained. The individual tree species were then successfully classified using both spectral and shape characteristics with an overall accuracy of 0.97 and a kappa coefficient of 0.85. Finally, the tree-level AGB (kg) was estimated based on the ITA; the AGB density (Mg/ha) was then upscaled based on the tree-level AGB values. It is concluded that, compared with the commonly used area-based method combining LiDAR and spectral metrics [root mean square error (RMSE)=19.58  Mg/ha], the ITA method performs better at estimating AGB density (RMSE=10.56  Mg/ha). The tree species information also improved the accuracy of the AGB estimation even though the species are not well diversified in this study area.

Satellite remote-sensing techniques face challenges in extracting vegetation-cover information in desert environments. The limitations in detection are attributed to three major factors: (1) soil background effect, (2) distribution and structure of perennial desert vegetation, and (3) tradeoff between spatial and spectral resolutions of the satellite sensor. In this study, a modified vegetation shadow model (VSM-2) is proposed, which utilizes vegetation shadow as a contextual classifier to counter the limiting factors. Pleiades high spatial resolution, multispectral (2 m), and panchromatic (0.5 m) images were utilized to map small and scattered perennial arid shrubs and trees. We investigated the VSM-2 method in addition to conventional techniques, such as vegetation indices and prebuilt object-based image analysis. The success of each approach was evaluated using a root sum square error metric, which incorporated field data as control and three error metrics related to commission, omission, and percent cover. Results of the VSM-2 revealed significant improvements in perennial vegetation cover and distribution accuracy compared with the other techniques and its predecessor VSM-1. Findings demonstrated that the VSM-2 approach, using high-spatial resolution imagery, can be employed to provide a more accurate representation of perennial arid vegetation and, consequently, should be considered in assessments of desertification.

Two-wavelength (1064/532  nm) lidar observations of long-range transported Saharan dust present in the atmosphere over Sofia, Bulgaria, during a 4-day dust intrusion event in winter 2010, are reported. Aged desert aerosols are detected at altitudes up to 4 km above the sea level, within and above the boundary layer as mixed with other aerosols—representing the particular case under consideration. Optical, microphysical, and dynamical properties of dust aerosols are obtained and analyzed. Special attention is paid to retrieving and vertical profiling of dust backscatter-related Ångström exponents (BAEs), as well as to determining their frequency-count distributions. Obtained BAE values in the range 0.3 to 0.6 (±0.2) indicate domination of coarse particles in the near overmicron size range. Reasonability of coarse-mode-dominated dust size composition is substantiated, based on measurement and transportation-history analysis. The performed frequency-count statistics reveals dust BAE distributions asymmetrically extended to multimode distribution shapes, resulting from dust mixing with finer local aerosol fractions. Peculiarities and patterns of the aerosol dynamics at different stages of dust-loading event are revealed and discussed.

A scheme is presented in this paper for ground moving target indication of multichannel synthetic aperture radar (SAR) systems. “Dominant-velocity” is chosen as a valuable metric to describe the velocity map of the observed scene and the “dominant-velocity image” (DVI) can be generated via the developed spatial spectral processing technique. The mean μ and the standard deviation σ of each dominant-velocity are estimated from its neighborhood. Two different methods are proposed to derive the detection threshold in the (μ,σ) plane: one is a nonparametric histogram approximation approach and the other is a parametric polynomial curve-fitting approach. The proposed ground moving target indication approach is a multistage one: the first stage implements the preliminary detection in the (μ,σ) plane and a clustering technique is utilized to indicate potential moving targets, while the second stage implements the fine detection via a velocity estimation method based on maximum signal-to-interference ratio for the tagged targets. Finally, the effectiveness of the proposed method is verified by both simulated and real airborne SAR data.

Raw data simulation of synthetic aperture radar (SAR) is useful for system designing, mission planning, and testing of imaging algorithms. According to the two-dimensional (2-D) frequency spectrum of the fixed-receiver bistatic SAR system, a rapid raw data simulation approach is proposed. With the combination of 2-D inverse Stolt transform in the 2-D frequency domain and phase compensation in the range-Doppler frequency domain, our approach can significantly reduce the simulation time. Therefore, simulations of extended scenes can be performed much more easily. Moreover, the proposed algorithm offers high accuracy of phase distribution, therefore, it can be used for single-pass fixed-receiver bistatic interferometric usage. The proposal is verified by extensive simulations of point targets and extended scene, in which the results indicate the feasibility as well as the effectiveness of our approach. In the end, the accuracy of phase distribution of the proposed algorithm is further examined with simulations of synthetic aperture radar interferometry.

This paper describes a portable hyperspectral goniometer system for measurement of hemispherical conical reflectance factor (HCRF) data for terrestrial applications, especially in the coastal zone. This system, the Goniometer for Portable Hyperspectral Earth Reflectance (GOPHER), consists of a computer-controlled Spectra Vista Corporation HR-1024 full-range spectrometer mounted on a rotating arc and track assembly, allowing complete coverage in zenith and azimuth of a full hemisphere for recording HCRF. The control software allows customized scan patterns to be quickly modified in the field, providing for flexibility in recording HCRF and the opposition effect with varying grid sizes and scan ranges in both azimuth and zenith directions. The spectrometer track can be raised and lowered on a mast to accommodate variations in terrain and land cover. To minimize the effect of variations in illumination during GOPHER scan cycles, a dual-spectrometer approach has been adapted to link records of irradiance recorded by a second spectrometer during the GOPHER HCRF scan cycle. Examples of field data illustrate the utility of the instrument for coastal studies.

Airborne laser scanning survey data were conducted with a scanning density of 4  points/m² to accurately map the surface of a unique central European complex of wetlands: the lower Biebrza River valley (Poland). A method to correct a degrading effect of vegetation (so-called “vegetation effect”) on digital terrain models (DTMs) was applied utilizing remotely sensed images, real-time kinematic global positioning system elevation measurements, topographical surveys, and vegetation height measurements. Geographic object-based image analysis (GEOBIA) was performed to map vegetation within the study area that was used as categories from which vegetation height information was derived for the DTM correction. The final DTM was compared with a model obtained, where additional correction of the “vegetation effect” was neglected. A comparison between corrected and uncorrected DTMs demonstrated the importance of accurate topography through a simple presentation of the discrepancies arising in features of the flood using various DTM products. An overall map classification accuracy of 80% was attained with the use of GEOBIA. Correction factors developed for various types of the vegetation reached values from 0.08 up to 0.92 m and were dependent on the vegetation type.

An energy functional is proposed based on an edge-region active contour model for synthetic aperture radar (SAR) image segmentation. The proposed energy functional not only has a desirable property to process inhomogeneous regions in SAR images, but also shows satisfactory convergence speed. Our proposed energy functional consists of two main energy terms: an edge-region term and a regularization term. The edge-region term is derived from a Gamma model and gradient term model, which can process the speckle noises and drive the motion of the curves toward desired locations. The regularization term is not only able to maintain a desired shape of the evolution curves but also has a strong smoothing curve effect and avoid the occurrence of small, isolated regions in the final segmentation. Finally, the gradient descent flow method is introduced for minimizing our energy functional. A desirable feature of the proposed method is that it is not sensitive to the contour initialization. Compared with other methods, experimental results show that the proposed approach has promising edge detection results on the synthetic and real SAR images.

Leaf area index (LAI) is a key biophysical parameter commonly used to determine vegetation status, productivity, and health in tropical grasslands. Accurate LAI estimates are useful in supporting sustainable rangeland management by providing information related to grassland condition and associated goods and services. The performance of support vector regression (SVR) was compared to partial least square regression (PLSR) on selected optimal hyperspectral bands to detect LAI in heterogeneous grassland. Results show that PLSR performed better than SVR at the beginning and end of summer. At the peak of the growing season (mid-summer), during reflectance saturation, SVR models yielded higher accuracies (R²=0.902 and RMSE=0.371  m² m⁻²) than PLSR models (R²=0.886 and RMSE=0.379 m² m⁻²). For the combined dataset (all of summer), SVR models were slightly more accurate (R²=0.74 and RMSE=0.578  m²m⁻²) than PLSR models (R²=0.732 and RMSE=0.58  m² m⁻²). Variable importance on the projection scores show that most of the bands were located in the near-infrared and shortwave regions of the electromagnetic spectrum, thus providing a basis to investigate the potential of sensors on aerial and satellite platforms for large-scale grassland LAI prediction.

Snow depth parameter inversion from passive microwave remote sensing is of great significance to hydrological process and climate systems. The Helsinki University of Technology (HUT) model is a commonly used snow emission model. Snow grain size (SGS) is one of the important input parameters, but SGS is difficult to obtain in broad areas. The time series of SGS are first evolved by an SGS evolution model (Jordan 91) using in situ data. A good linear relationship between the effective SGS in HUT and the evolution SGS was found. Then brightness temperature simulations are performed based on the effective SGS and evolution SGS. The results showed that the biases of the simulated brightness temperatures based on the effective SGS and evolution SGS were −6.5 and −3.6  K, respectively, for 18.7 GHz and −4.2 and −4.0  K for 36.5 GHz. Furthermore, the model is performed in six pixels with different land use/cover type in other areas. The results showed that the simulated brightness temperatures based on the evolution SGS were consistent with those from the satellite. Consequently, evolution SGS appears to be a simple method to obtain an appropriate SGS for the HUT model.

Qinghai Lake basin and the lake have undergone significant changes in recent decades. We examine MODIS-derived grassland vegetation and snow cover of the Qinghai Lake basin and their relations with climate parameters during 2001 to 2010. Results show: (1) temperature and precipitation of the Qinghai Lake basin increased while evaporation decreased; (2) most of the grassland areas improved due to increased temperature and growing season precipitation; (3) weak relations between snow cover and precipitation/vegetation; (4) a significantly negative correlation between lake area and temperature (r=−0.9, p<0.05); and (5) a positive relation between lake level (lake-level difference) and temperature (precipitation). Compared with Namco Lake (located in the inner Tibetan Plateau) where the primary water source of lake level increases was the accelerated melt of glacier/perennial snow cover in the lake basin, for the Qinghai Lake, however, it was the increased precipitation. Increased precipitation explained the improvement of vegetation cover in the Qinghai Lake basin, while accelerated melt of glacier/perennial snow cover was responsible for the degradation of vegetation cover in Namco Lake basin. These results suggest different responses to the similar warming climate: improved (degraded) ecological condition and productive capacity of the Qinghai Lake basin (Namco Lake basin).

This paper explores the potential of using unmanned aircraft system (UAS)-based visible-band images to assess cotton growth. By applying the structure-from-motion algorithm, the cotton plant height (ph) and canopy cover (cc) information were retrieved from the point cloud-based digital surface models (DSMs) and orthomosaic images. Both UAS-based ph and cc follow a sigmoid growth pattern as confirmed by ground-based studies. By applying an empirical model that converts the cotton ph to cc, the estimated cc shows strong correlation (R²=0.990) with the observed cc. An attempt for modeling cotton yield was carried out using the ph and cc information obtained on June 26, 2015, the date when sigmoid growth curves for both ph and cc tended to decline in slope. In a cross-validation test, the correlation between the ground-measured yield and the estimated equivalent derived from the ph and/or cc was compared. Generally, combining ph and cc, the performance of the yield estimation is most comparable against the observed yield. On the other hand, the observed yield and cc-based estimation produce the second strongest correlation, regardless of the complexity of the models.

A methodology is presented, by which atmospheric aerosol retrievals from a standard, elastic-scatter, lidar can be constrained by using information from coincident measurements from a high spectral resolution lidar (HSRL) or Raman lidar at a different wavelength. As high spectral resolution or inelastic-scattering lidars are now being incorporated coaxially into instruments with traditional, elastic-scatter channels at different wavelengths, a standard approach is needed to incorporate or fuse the diversity of spectral information so as to make maximal use of the aerosol measurements made from the elastic-scatter channel or channels. The approach is evaluated through simulation and with data from the NASA Langley Research Center Airborne HSRL instrument. The generality and extensibility of the method is also explored and discussed in the context of aerosol modeling.

This paper addresses the design of a low-cost, low-complexity, and rapidly deployable wireless sensor network (WSN) for rescue site monitoring after earthquakes. The system structure of the hybrid WSN is described. Specifically, the proposed hybrid WSN consists of two kinds of wireless nodes, i.e., the monitor node and the sensor node. Then the mechanism and the system configuration of the wireless nodes are detailed. A transmission control protocol (TCP)-based request-response scheme is proposed to allow several monitor nodes to communicate with the monitoring center. UDP-based image transmission algorithms with fast recovery have been developed to meet the requirements of in-time delivery of on-site monitor images. In addition, the monitor node contains a ZigBee module that used to communicate with the sensor nodes, which are designed with small dimensions to monitor the environment by sensing different physical properties in narrow spaces. By building a WSN using these wireless nodes, the monitoring center can display real-time monitor images of the monitoring area and visualize all collected sensor data on geographic information systems. In the end, field experiments were performed at the Training Base of Emergency Seismic Rescue Troops of China and the experimental results demonstrate the feasibility and effectiveness of the monitor system.

The potential use of phased array type L-band synthetic aperture radar (PALSAR) data for discriminating distinct physiographic mangrove types with different forest structure developments in a subtropical mangrove forest located in Cananéia on the Southern coast of São Paulo, Brazil, is investigated. The basin and fringe physiographic types and the structural development of mangrove vegetation were identified with the application of the Kruskal–Wallis statistical test to the SAR backscatter values of 10 incoherent attributes. The best results to separate basin to fringe types were obtained using copolarized HH, cross-polarized HV, and the biomass index (BMI). Mangrove structural parameters were also estimated using multiple linear regressions. BMI and canopy structure index were used as explanatory variables for canopy height, mean height, and mean diameter at breast height regression models, with significant R²=0.69, 0.73, and 0.67, respectively. The current study indicates that SAR L-band images can be used as a tool to discriminate physiographic types and to characterize mangrove forests. The results are relevant considering the crescent availability of freely distributed SAR images that can be more utilized for analysis, monitoring, and conservation of the mangrove ecosystem.

Fractional vegetation cover (FVC) is an important variable for describing the quality and changes of vegetation in terrestrial ecosystems. Dimidiate pixel models and physical models are widely used to estimate FVC. Six dimidiate pixel models based on different vegetation indices (VI) and four look-up table (LUT) methods were compared to estimate FVC from Landsat 8 OLI data. Comparisons with in situ FVC of steppe and corn showed that the model proposed by Baret et al., which is based on the normalized difference vegetation index (NDVI), predicted FVC most accurately followed by Carlson and Ripley’s method. Gutman and Ignatov’s method overestimated FVC. Modified soil adjusted vegetation index (MSAVI) and the mixture of NDVI and RVI showed potential to replace NDVI in Gutman and Ignatov’s model, whereas the difference vegetation index (DVI) performed less well. At low vegetation cover, the LUT using reflectances to constrain the cost function performed better than LUTs using VI to constrain the cost function, whereas at high vegetation cover, the LUT based on NDVI estimated FVC most accurately. The applications of DVI and MSAVI to constrain the cost function also obtained improvement at high vegetation cover. Overall, the accuracies of LUT methods were a little lower than those of dimidiate pixel models.

Unmanned aerial vehicle (UAV) remote sensing technology has come into wide use in recent years. The poor stability of the UAV platform, however, produces more inconsistencies in hue and illumination among UAV images than other more stable platforms. Image dodging is a process used to reduce these inconsistencies caused by different imaging conditions. We propose an algorithm for automatic image dodging of UAV images using two-dimensional radiometric spatial attributes. We use object-level image smoothing to smooth foreground objects in images and acquire an overall reference background image by relative radiometric correction. We apply the Contourlet transform to separate high- and low-frequency sections for every single image, and replace the low-frequency section with the low-frequency section extracted from the corresponding region in the overall reference background image. We apply the inverse Contourlet transform to reconstruct the final dodged images. In this process, a single image must be split into reasonable block sizes with overlaps due to large pixel size. Experimental mosaic results show that our proposed method reduces the uneven distribution of hue and illumination. Moreover, it effectively eliminates dark-bright interstrip effects caused by shadows and vignetting in UAV images while maximally protecting image texture information.

This paper discusses the probability of detection enhancement at the fusion area scanned by two radars. The improvement is implemented by three techniques, scan rate modulation, scan-to-scan processing, and applying the probability of detection rules (AND, OR) to make a decision. The simulation results reveal that using AND rules with precalculation of the threshold is better at enhancing the probability of detection and reducing the probability of false alarm. Furthermore, scan-to-scan processing has a considerable influence on reducing false alarms. Moreover, scan rate modulation increases the probability of detection and maintains the probability of false alarm within a permissible limit.

Object-based approaches in the segmentation and classification of remotely sensed images yield more promising results compared to pixel-based approaches. However, the development of an object-based approach presents challenges in terms of algorithm selection and parameter tuning. Subjective methods are often used, but yield less than optimal results. Objective methods are warranted, especially for rapid deployment in time-sensitive applications, such as earthquake damage assessment. Herein, we used a systematic approach in evaluating object-based image segmentation and machine learning algorithms for the classification of earthquake damage in remotely sensed imagery. We tested a variety of algorithms and parameters on post-event aerial imagery for the 2011 earthquake in Christchurch, New Zealand. Results were compared against manually selected test cases representing different classes. In doing so, we can evaluate the effectiveness of the segmentation and classification of different classes and compare different levels of multistep image segmentations. Our classifier is compared against recent pixel-based and object-based classification studies for postevent imagery of earthquake damage. Our results show an improvement against both pixel-based and object-based methods for classifying earthquake damage in high resolution, post-event imagery.

This study aims to find the optimal vegetation indices (VIs) to remotely estimate plant nitrogen concentration (PNC) in winter oilseed rape across different growth stages. Since remote sensing cannot “sense” N in live leaves, remote estimation of PNC should be based on understanding the relationships between PNC and chlorophyll (Chl), carotenoid concentration (Car), Car/Chl, dry mass (DM), and leaf area index (LAI). The experiments with eight nitrogen fertilization treatments were conducted in 2014 to 2015 and 2015 to 2016, and measurements were acquired at six-leaf, eight-leaf, and ten-leaf stages. We found that at each stage, Chl, Car, DM, and LAI were all strongly related to PNC. However, across different growth stages, semipartial correlation and linear regression analysis showed that Chl and Car had consistently significant relationships with PNC, whereas LAI and DM were either weakly or barely correlated with PNC. Therefore, the most suitable VIs should be sensitive to the change in Chl and Car while insensitive to the change in DM. We found that anthocyanin reflectance index and the simple ratio of the red band to blue band fit the requirements. The validation with the 2015 to 2016 dataset showed that the selected VIs could provide accurate estimates of PNC in winter oilseed rape.

Generating a georeferenced mosaic map from unmanned aerial vehicle (UAV) imagery is a challenging task. Direct and indirect georeferencing methods may fail to generate an accurate mosaic map due to the erroneous exterior orientation parameters stored in the inertial measurement unit (IMU), erroneous global positioning system (GPS) data, and difficulty in locating ground control points (GCPs) or having a sufficient number of GCPs. This paper presents a practical framework to orthorectify and georeference aerial images using the robust features-based matching method. The proposed georeferencing process is fully automatic and does not require any GCPs. It is also a near real-time process which can be used to determine whether aerial images taken by UAV cover the entire target area. We also extend this framework to use the inverse georeferencing process to update the IMU/GPS data which can be further used to calibrate the camera of the UAV, reduce IMU/GPS errors, and thus produce more accurate mosaic maps by employing any georeferencing method. Our experiments demonstrate the effectiveness of the proposed framework in producing comparable mosaic maps as commercial software Agisoft and the effectiveness of the extended framework in significantly reducing the errors in the IMU/GPS data.

Clouds’ macrophysical characteristics play an important role in the climate system and dramatically vary because of the diverse climatic and geographic factors in China. We analyze cloud macrophysical characteristics and the differences between subregions in China (18°–54°N, 73°–135°E) from March 2012 to February 2015 based on Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations, including cloud fractions, cloud vertical distribution, and cloud geometrical properties with the perspective of daytime and nighttime. We found that annual single layer, multilayer (ML), and total cloud fractions are 40.4±1.1%, 22.4±0.4%, and 62.8±1.5%, respectively, and clouds are generally located between 6 and 12 km. The cloud fractions in daytime are less than that in nighttime over the south while that of Tibet shows the reverse trend. In the vertical direction, except for Tibet, the clouds in nighttime have larger spatial coverage and are higher in altitude than that in daytime. The regional average values of cloud macrophysical characteristics in the south are highest, followed successively by Tibet, north, and northwest. Cloud geometrical depth and spacing show a gradually declining trend with the increase in layers and decrease of altitude in ML cloud system.

Urban environments are very dynamic phenomena, and it is essential to update urban-related information for various applications. In this regard, remotely sensed data have been utilized widely to extract and monitor urban land use and land cover changes. Particularly, synthetic aperture radar (SAR) data, due to several advantages of this technology in comparison to passive sensors, provides better performance especially in tropical regions. However, the methodological approaches for extraction of information from SAR images are another important task that needs to be considered appropriately. This paper attempts to investigate and compare the performance of different image classification techniques for extracting urban areas using advanced land observing satellite phased array type L-band synthetic aperture radar imagery. Several object- [such as rule based (RB), support vector machine (SVM) and K-nearest neighbor (K-NN)] and pixel-based [decision tree (DT)] classification techniques were implemented, and their results were compared in detail. The overall results indicated RB classification performed better than other techniques. Furthermore, DT method, due to its predefined rules, distinguished the land cover classes better than SVM and K-NN, which were based on training datasets. Nevertheless, this study confirms the potential of SAR data and object-based classification techniques in urban detection and land cover mapping.

The anisotropic scale space (ASS) is often used to enhance the performance of a scale-invariant feature transform (SIFT) algorithm in the registration of synthetic aperture radar (SAR) images. The existing ASS-based methods usually suffer from unstable keypoints and false matches, since the anisotropic diffusion filtering has limitations in reducing the speckle noise from SAR images while building the ASS image representation. We proposed a speckle reducing SIFT match method to obtain stable keypoints and acquire precise matches for the SAR image registration. First, the keypoints are detected in a speckle reducing anisotropic scale space constructed by the speckle reducing anisotropic diffusion, so that speckle noise is greatly reduced and prominent structures of the images are preserved, consequently the stable keypoints can be derived. Next, the probabilistic relaxation labeling approach is employed to establish the matches of the keypoints then the correct match rate of the keypoints is significantly increased. Experiments conducted on simulated speckled images and real SAR images demonstrate the effectiveness of the proposed method.