In recent decades, with the rapid development of remote sensing technology, high resolution remote sensing images have been widely used in various fields due to their characteristics, such as rich spectral information and complex texture information. As a key step in the feature extraction, multi-scale image segmentation algorithm has been a hotspot currently. The traditional image segmentation is based on pixels, which only takes the spectral information of pixel into account, and ignores the texture, spatial information and contextual relation of the objects in the image. The experimental high resolution remote sensing images are from GF-2 and the features of the experimental data are obvious, the edges are clear. By using the statistical region merging (SRM) algorithm, the fractal net evolution approach (FNEA) algorithm and the unsupervised multi-scale segmentation of color images (UMSC) algorithm, this paper analyzes the segmentation effects of three multi-scale segmentation algorithms on the optimal scale and on the same segmentation scale respectively. The experimental results under the optimal scale and the same segmentation scale show that the SRM algorithm outperforms the UMSC algorithm, and UMSC algorithm outperforms the FENA algorithm in multi-scale segmentation.
Overhauser magnetometer, a kind of static-magnetic measurement system based on the Overhauser effect, has been widely used in archaeological exploration, mineral resources exploration, oil and gas basin structure detection, prediction of engineering exploration environment, earthquakes and volcanic eruotions, object magnetic measurement and underground buried booty exploration. Overhauser magnetometer plays an important role in the application of magnetic field measurement for its characteristics of small size, low power consumption and high sensitivity. This paper researches the design and the application of the analog circuit of JOM-3 Overhauser magnetometer. First, the Larmor signal output by the probe is very weak. In order to obtain the signal with high signal to noise rstio(SNR), the design of pre-amplifier circuit is the key to improve the quality of the system signal. Second, in this paper, the effectual step which could improve the frequency characters of bandpass filter amplifier circuit were put forward, and theoretical analysis was made for it. Third, the shaping circuit shapes the amplified sine signal into a square wave signal which is suitable for detecting the rising edge. Fourth, this design elaborated the optimized choice of tuning circuit, so the measurement range of the magnetic field can be covered. Last, integrated analog circuit testing system was formed to detect waveform of each module. By calculating the standard deviation, the sensitivity of the improved Overhauser magnetometer is 0.047nT for Earth’s magnetic field observation. Experimental results show that the new magnetometer is sensitive to earth field measurement.
Snow accumulation has a very important influence on the natural environment and human activities. Meanwhile, improving the estimation accuracy of passive microwave snow depth (SD) retrieval is a hotspot currently. Northeastern China is a typical snow study area including many different land cover types, such as forest, grassland and farmland. Especially, there is relatively stable snow accumulation in January every year. The brightness temperatures which are observed by the Advanced Microwave Scanning Radiometer 2 (AMSR2) on GCOM-W1 and FengYun3B Microwave Radiation Imager (FY3B-MWRI) in the same period in 2013 are selected as the study data in the research. The results of snow depth retrieval using AMSR2 standard algorithm and Jiang’s FY operational algorithm are compared in the research. Moreover, to validate the accuracy of the two algorithms, the retrieval results are compared with the SD data observed at the national meteorological stations in Northeastern China. Furthermore, the retrieval SD is also compared with AMSR2 and FY standard SD products, respectively. The root mean square errors (RMSE) results using AMSR2 standard algorithms and FY operational algorithm are close in the forest surface, which are 6.33cm and 6.28cm, respectively. However, The FY operational algorithm shows a better result than the AMSR2 standard algorithms in the grassland and farmland surface. The RMSE results using FY operational algorithm in the grassland and farmland surface are 2.44cm and 6.13cm, respectively.
Snow cover information has great significance for monitoring and preventing snowstorms. With the development of
satellite technology, geostationary satellites are playing more important roles in snow monitoring. Currently, cloud
interference is a serious problem for obtaining accurate snow cover information. Therefore, the cloud pixels located in the
MODIS snow products are usually replaced by cloud-free pixels around the day, which ignores snow cover dynamics.
FengYun-2(FY-2) is the first generation of geostationary satellite in our country which complements the polar orbit
satellite. The snow cover monitoring of Northeast China using FY-2G data in January and February 2016 is introduced in
this paper. First of all, geometric and radiometric corrections are carried out for visible and infrared channels. Secondly,
snow cover information is extracted according to its characteristics in different channels. Multi-threshold judgment
methods for the different land types and similarity separation techniques are combined to discriminate snow and cloud.
Furthermore, multi-temporal data is used to eliminate cloud effect. Finally, the experimental results are compared with the
MOD10A1 and MYD10A1 (MODIS daily snow cover) product. The MODIS product can provide higher resolution of the
snow cover information in cloudless conditions. Multi-temporal FY-2G data can get more accurate snow cover information
in cloudy conditions, which is beneficial for monitoring snowstorms and climate changes.
With the rapid development of remote sensing technology, the spatial resolution and temporal resolution of satellite imagery also have a huge increase. Meanwhile, High-spatial-resolution images are becoming increasingly popular for commercial applications. The remote sensing image technology has broad application prospects in intelligent traffic. Compared with traditional traffic information collection methods, vehicle information extraction using high-resolution remote sensing image has the advantages of high resolution and wide coverage. This has great guiding significance to urban planning, transportation management, travel route choice and so on. Firstly, this paper preprocessed the acquired high-resolution multi-spectral and panchromatic remote sensing images. After that, on the one hand, in order to get the optimal thresholding for image segmentation, histogram equalization and linear enhancement technologies were applied into the preprocessing results. On the other hand, considering distribution characteristics of road, the normalized difference vegetation index (NDVI) and normalized difference water index (NDWI) were used to suppress water and vegetation information of preprocessing results. Then, the above two processing result were combined. Finally, the geometric characteristics were used to completed road information extraction. The road vector extracted was used to limit the target vehicle area. Target vehicle extraction was divided into bright vehicles extraction and dark vehicles extraction. Eventually, the extraction results of the two kinds of vehicles were combined to get the final results. The experiment results demonstrated that the proposed algorithm has a high precision for the vehicle information extraction for different high resolution remote sensing images. Among these results, the average fault detection rate was about 5.36%, the average residual rate was about 13.60% and the average accuracy was approximately 91.26%.