A novel application of hyperspectral instrumentation is described here, where a sheet-of-light method, produced by a low power laser light, is used to compute range measurements, and a hyperspectral instrument is used to acquire spectral information in the visible and near-infrared range of the spectrum. This report addresses two problems which in the literature are generally addressed independently; acquisition and analysis of hyperspectral images, and acquisition and analysis of range information. The bimodal instrument described in this report consists of a LCTF-based hyperspectral system which is used to acquire spectral images in the 450-1100 nm range and a multi-line laser light used to acquire range measurements of a scene. This laser light method can be used to acquire fast range measurements as the scene is partitioned into three sub-ranges therefore reducing the acquisition time by threefold. The methods used to get calibrated hyperspectral measurements (to reflectance values) of the proposed 3D hyperspectral instrument, and range measurements (to mm) are described in this paper. A test case shows the capabilities of this instrument for producing 3D hyperspectral imagery of human skin samples.
The Moderate Resolution Imaging Spectroradiometer (MODIS) reflective solar bands (RSB) cover wavelengths from 0.41 to 2.2μm. They are calibrated on-orbit by a solar diffuser (SD) panel, made of space-grade Spectralon. During each SD calibration a solar diffuser stability monitor (SDSM) is operated concurrently to track the changes of the SD bidirectional reflectance factor (BRF). The SDSM views alternately the sunlight (Sun View) through a fixed transmission
screen and the sunlight diffusely reflected from the SD panel (SD view). A design error in the SDSM system, not discovered until post-launch, has caused significant ripples in the SDSM Sun view responses. Consequently an alternative normalization approach has been developed to remove the ripples in the SDSM Sun view responses and their impacts on the SD degradation analysis. This approach has been successfully used in the SDSM measurements on-orbit. In order to reduce the direct solar exposure onto the SD panel, the MODIS instrument was designed with a SD door that is normally commanded to an "open" position during SD/SDSM observations and to a "closed" position when the calibration is completed. For Terra MODIS launched in December 1999, an SD door related anomaly occurred in May 2003 that led to a decision to set the SD door permanently at the open position. This operational configuration has resulted in extra time of direct solar illumination on the SD plate and therefore a much faster SD degradation rate. In this paper we provide a brief description of the MODIS RSB calibration algorithm and the on-board SD and SDSM system used for the calibration. We examine the Terra MODIS SD degradation rate and its spectral dependency. The results from five years of SDSM observations are summarized in this paper and used to evaluate the SD on-orbit performance and its impact on the MODIS RSB calibration uncertainty. Prior to the SD door anomaly, the SD annual degradation rate was approximately 3% at 0.41μm, 2% at 0.47μm, and 1% at 0.53μm. After the SD door anomaly, the SD annual degradation rate has increased to 10% at 0.41μm, 7% at 0.47μm, and 4.5% at 0.53μm.
Hyperspectral images of the Earth’s surface are increasingly being acquired from aerial platforms. The dozens or hundreds of bands acquired by a typical hyperspectral sensor are acquired either through a scanning process or by collecting a sequence of images at varying wavelengths. This latter method has the advantage of acquiring coherent images of a scene at different wavelengths. However, it takes time to collect these images and some form of co-registration is required to build coherent image cubes. In this paper, we present a method to register many bands acquired sequentially at different wavelengths from a helicopter. We discuss the application of the Phase Correlation (PC) Method to recover scaling, rotation, and translation from an airborne hyperspectral imaging system, dubbed PHyTIS. This approach is well suited for remotely sensed images acquired from a moving platform, which induces image registration errors due to along and across track movement. We were able to register images to within ± 1 pixel across entire image cubes obtained from the PHyTIS hyperspectral imaging system, which was developed for precision farming applications.
Terra MODIS, also referred to as the MODIS Protoflight Model (PFM), was launched on-board the NASA's EOS Terra spacecraft on December 18, 1999. It has been in operation for more than four years and continuously providing the science community quality data sets for studies of the Earth's land, oceans, and atmosphere. It has also served as the primary source of information for the MODIS Land Rapid Response System for observing and reporting on natural disasters, and providing active fire information around the Earth. The MODIS instrument has 36 spectral bands with wavelengths ranging from 0.41mm to 14.5mm: 20 bands with wavelengths below 2.2mm are the reflective solar bands (RSB) and the other 16 bands are the thermal emissive bands (TEB). The RSB are calibrated on-orbit using a solar diffuser (SD) with the degradation of its bi-directional reflectance factor (BRF) tracked by an on-board solar diffuser stability monitor (SDSM). The calibration coefficients are updated via Look-Up Tables (LUTs) for the Level 1B code that converts the sensor's Earth view response from digital counts to calibrated reflectance and radiance. In this paper we review the MODIS RSB on-orbit calibration algorithm and the methodology of computing and updating the calibration coefficients determined from the SD and SDSM data sets. We present examples of the sensor's long-term and short-term stability trending of key RSB calibration parameters using over four years of on-orbit calibration data sets. Special considerations due to changes in instrument configuration and sensor response are also discussed.
The MODIS Flight Model 1 (FM1) has been in operation for more than two years since its launch onboard the NASA's Earth Observing System (EOS) Aqua spacecraft on May 4, 2002. The MODIS has 36 spectral bands: 20 reflective solar bands (RSB) with center wavelengths from 0.41 to 2.2mm and 16 thermal emissive bands (TEB) from 3.7 to 14.5mm. It provides the science community observations (data products) of the Earth's land, oceans, and atmosphere for a board range of applications. Its primary on-orbit calibration and characterization activities are performed using a solar diffuser (SD) and a solar diffuser stability monitor (SDSM) system for the RSB and a blackbody for the TEB. Another on-board calibrator (OBC) known as the spectro-radiometric calibration assembly (SRCA) is used for the instrument's spatial (TEB and RSB) and spectral (RSB only) characterization. We present in this paper the status of Aqua MODIS calibration and characterization during its first two years of on-orbit operation. Discussions will be focused on the calibration activities executed on-orbit in order to maintain and enhance the instrument's performance and the quality of its Level 1B (L1B) data products. We also provide comparisons between Aqua MODIS and Terra MODIS (launched in December, 1999), including their similarity and difference in response trending and optics degradation. Existing data and results show that Aqua MODIS bands 8 (0.412mm) and 9 (0.443mm) have much smaller degradation than Terra MODIS bands 8 and 9. The most noticeable feature shown in the RSB trending is that the mirror side differences in Aqua MODIS are extremely small and stable (<0.1%) while the Terra MODIS RSB trending has shown significant mirror side difference and wavelength dependent degradation. The overall stability of the Aqua MODIS TEB is also better than that of the Terra MODIS during their first two years of on-orbit operation.
This paper presents a comparative analysis of two evolved neural networks for control. Traditionally, the structure of Radial Basis Functions Networks (RBFNs) and Multilayer Feedforward Networks (MFNs) are found by a trial-and-error process. This process consists on finding an appropriate network structure such that the unknown nonlinearities of the plant can be estimated to some desired accuracy. In general, a neural network is composed of two elements: structural and learning parameters. The structural parameters are all those elements that determine the size of the network. The learning parameters are all those elements that determine learning and convergence of the network. The approach presented in this work uses a Genetic Algorithm (GA) to evolve the structure, and uses a gradient descent algorithm to adjust the weights in the network. An analysis of the evolution of RBFNs and MFNs by means of a GA is examined in detail. It is shown that the networks can be encoded in a chromosome for their evolution. Experimental results show the performance of Evolved Radial Basis Functions Networks and Evolved Multilayer Feedforward Networks in the identification and control of a nonlinear plant.
A methodology is described for an airborne, downlooking, longwave infrared imaging spectrometer based technique for the detection and tracking of plumes of toxic gases. Plumes can be observed in emission or absorption, depending on the thermal contrast between the vapor and the background terrain. While the sensor is currently undergoing laboratory calibration and characterization, a radiative exchange phenomenology model has been developed to predict sensor response and to facilitate the sensor design. An inverse problem model has also been developed to obtain plume parameters based on sensor measurements. These models, the sensors, and ongoing activities are described.
A new approach to the extraction of the polygonal approximation is presented. The method obtains a smaller set of the important features by means of an evolutionary algorithm. A genetic approach with some heuristics, improves contour approximation search by starting with a parallel search at various points in the contour. The algorithm uses genetic algorithms to encode a polygonal approximation as a chromosome and evolve it to provide a polygonal approximation. Experimental results are provided.
This paper presents a new approach to the identification and control of dynamical systems by means of evolved radial basis function neural networks (ERBFNs). Traditionally, radial basis function networks (RBFNs) parameters which are used for identification and control are fixed beforehand by a trial-and-error process. This process consists of finding structural and training parameters. Once these parameters are fixed the only parameters that remain to be determined are network weights. In general, the weights are adjusted using a gradient approach so that network output asymptotically follow the plant output. In this paper a new approach to the selection of structural and training parameters is introduced. A hybrid system is proposed which uses an evolutionary algorithm to select optimum structural parameters and uses the LMS algorithm to adjust network weights. In this context, RBFN parameters such as basis function centers, widths and training parameters are chosen at random and adjusted by an evolutionary algorithm, throughout the identification and control process. Experimental results show that the system is able to effectively identify and control dynamical systems.
A new approach to the identification of dynamical systems by means of evolved neural networks is presented. We implement two functional neural networks: polynomials and orthogonal basis functions. The functional neural networks contain four parameters that need to be optimized: the weights, training parameters, network topology and scaling factors. An approach to the solution of this combinatorial problem is to genetically evolve functional neural networks. This paper presents a preliminary analysis of the proposed method to automatically select network parameters. The networks are encoded as chromosomes that are evolved during the identification by means of genetic algorithms. Experimental results show that the method is effective for the identification of dynamical systems.
This paper deals with the problem of robust fast recognition of partially occluded or incomplete views of `flat' objects. Robustness is accomplished through hypothesis confirmation using complementary or supporting information available for the current hypothesis and by model-based hypothesis verification. Classification speed is obtained by pruning the hypothesis hierarchy using simple pruning procedures based on structural properties derived from current object representation. In addition, classification speed is also improved through the use of simple model-based decision making procedures instead of computationally expensive transformations. Normalized Interval Vertex Descriptors (NIVD) are used to represent objects. NIVDs are representations derived from the physical characteristics of an object (vertices and sides) that are easy to obtain, especially for polygon like shapes. They provide not only a compact representation, but they also allow the definition of features that can be used to speed up the classification process. Experimental results of this process are also included.