In recent years, hyperspectral remote sensing has stepped into a new stage in China. There are some advanced hyperspectral imagers and CCD cameras developed by Chinese institutes and companies. Pushbroom Hyperspectral Imager (PHI) and Operative Modular Imaging Spectrometer (OMIS) have presented the level of airborne hyperspectral imagers in China, which have been developed by the Chinese Academy of Sciences. A narrow band hyperspectral digital camera system (HDCS) was developed and tested in 2000, the center of wavelength of which can be changed to fit different applications. There is also a kind of Fourier Imaging Spectrometer developed in China. Accordingly, Chinese scholars have created a number of models to meet different application problems. Some new models for hyperspectral remote sensing are provided. They are Hyperspectral Data Classification Model, POS Dat Geometric Correction Model, Derivative Spectral Model (DSM), Multi-temporal Index Image Cube Model (MIIC), Hybrid Decision Tree Model (HDT) and Correlation Simulating Analysis Model (CSAM). Some successful applications are provided and evaluated.

Inverse Synthetic Aperture Radar (ISAR) imaging of maneuvering target has received much attention in recent years. In the paper, we first discuss translational motion compensation (TMC), which can be decomposed into two step: envelope alignment and autofocus. There are many effective algorithms for these two step, but most of algorithms are only suitable to steadily flying targets. The analysis of the paper shows that the envelope alignment algorithms for steadily flying targets are still effective for maneuvering targets, but most of autofocus algorithms are not work very well, especially when the maneuverability is great, so this paper also presents a new autofocus algorithm based on coherent integration to solve this problem. Then after TMC, we discuss imaging problem of maneuvering target, in fact, it is a instantaneous spectrum estimation problem, most of instantaneous ISAR imaging methods of maneuvering target are based on time-frequency distribution, and suppose scatterer echo has constant amplitude and linear frequency modulation. In fact, when target has dihedrals and trihedrals component, or scatterer' migration through resolution cells (MTRC) exists, these situation all will cause the amplitude is not a constant. For maneuvering target, scatterer echo's Doppler is time varying, usually, its time-frequency distribution can be regard as linear or part linear. So we propose a instantaneous ISAR imaging method based on multicomponent amplitude modulation and linear frequency modulation(AM-LFM) signal's parameter estimation, this method first estimates the chirp rate of one component by dechirping, then demodulates this component to a sinusoidal signal, separates and cleans this component in Fourier domain, and estimates instantaneous amplitude, then for next component, finally we get all the scattterers instantaneous amplitudes and frequencies, and the instantaneous ISAR images are obtained now.

The paper provides some new developments based on the azimuth synthetic and compressing method to achieve a high azimuth resolution and unambiguous image for forward-looking SAR (F-SAR) with multiple receiving antennas. The first point of the new development is addressed on how to select and optimize the reference functions for azimuth compression. The second new point is addressed on how to configure the parameters such as wavelength and the distances between the receiving antennas. Ans the effect of these parameters to the practical implementation of F-SAR system will be also discussed. The third point is addressed on the methods to fuse the data from different receiving antennas for solving the left/right azimuth ambiguity.

Along with specking imaging characteristic, by wavelet transform, multiresolution for extended object (satellite) speckle imaging is proposed in this paper. Form the simulation, the object feature at different scale is extracted. The satellite features on horizontal and vertical orientation become visible. Only the low-frequency sub-band images are used, the satellite outline is fairly good. Furthermore, for large size images, it may be a method to lessen the computational complexity and compress the images data under the tolerant conditions.

In ideal optical system, one pixel on the image surface corresponds to a unit area on the object surface. However, in actual optical system, the image of a unit area on object will be extended to a spot around the central pixel due to the response of optical system and the process of electrical signal. The spot is also called point extended function. In this paper, the extension of imaging is simulated in laboratory, which shows the brightness variation of the image. In the experiment, an object that corresponds to 1*1 pixel is placed in a uniform background at appropriate distance, and the radiance is measured on the image surface. Then we change the object to 2*2, 4*4, 8*8, 16*16 pixels and the radiance is also measured. It is verified that the image of unit area will be extended to several pixels. After a careful analysis and examination of the data, a new method of estimating the transmission function of optical system is presented based on optical imaging theory. The method is verified in laboratory by measuring blackbody using AGEMA900 thermo-vision imaging device, and can be applied to estimating the transmission function of unknown optical system.

A scanner based diode pumped solid state imaging laser radar is introduced in this paper in detail. Three different methods for the acquisition of range and intensity image are described and compared. The image for the range and intensity can be obtained simultaneously based on a PC-based acquisition card technology. The method based on FPGA is more popular but with low effective ladar range. The theory based on the voltage to time conversion technique is simple, but with low range precision.

The use of radar polarimetry for subsurface sensing is introduced in this paper. We developed a polarimetric borehole radar, which can acquire the full-polarimetric radar profile in a borehole. The measured profile showed that subsurface fracture characteristics can be obtained from the polarimetric analysis. This information is closely related to subsurface fracture characteristics, which cannot directly be imaged by conventional radar.

Polarimetric radar interferometry is a compound technique that has shown the ability to extract geophysical parameters from synthetic aperture radar (SAR) images and its usefulness in terrain classification and surface change detection. A three-dimensional image can be constructed by coherence integrating the backscatter data over the measured frequency band and the two spatial coordinates of the two-dimensional synthetic aperture. Accordingly, we have developed a vector network analyzer based three-dimensional synthetic aperture radar imaging system. The radar system consists of the two-dimensional antenna scanner, which are two double-ridged waveguide horn antennas. This test synthetic aperture radar system is used for monitoring radar targets in an anechoic chamber under ideal conditions. The system operates at frequencies between 50 MHz and 20 GHz. The antenna scanning aperture is about 2 meters by 2 meters. Being applied the rotation of polarization basis, the full-polarimetric images for several targets with various permittivities were acquired and some experimental results are demonstrated. When the radar target is a conducting cylinder, the polarization effect is clear. The differential interferometric SAR has shown the high resolution of less than one centimeter in near field condition. The interferometric polarimetric SAR provides abundant information about the radar target is identified. Moreover, the potentials of three-dimensional synthetic aperture radar system are discussed.

Hand held uncooled Staring Infrared Camera has extensive domestic applications in China. A practical and low cost infrared starring camera has been designed and built. The prototype infrared camera is configured in three separate units: a signal acquiring unit, signal processing unit and the battery supply. The system design method of Uncooled Staring Infrared Camera based on 320 x 240 Uncooled Focal Plane Array (UFPA) is proposed. The paper discusses its performance and reviews the prospect for future developments.

Based on the analysis of microscanning, the model of pixel transfer function (PTF) and its comparison with modulation transfer function (MTF) is introduced. With PTF microscanned images and their comparison with ideal images, simulated images with existing method and non-microscanned images are made and analyzed.

Motion compensation of airborne synthetic aperture radar (SAR) consists of motion compensation based on inertial navigation system (INS) and autofocus algorithms. This paper deeply analyzes both the two aspects on the principle and gives the implementation solution. By the motion compensation based on INS, it gives not only the influence of the precision but also the effect of the data rate and points out that the data rate of the INS is the same important as the precision of INS for better result of motion compensation. Because of its excellent compensation ability and wide compensation scope, phase gradient autofocus (PGA) is used as an effective algorithm to do autofocus work. Based on the principle of PGA, the detail approach and simulation result is presented. Additionally, this paper also points out that the bottle-neck of the real time imaging processing of airborne SAR is not the computing capacity of the processing chip but the amount of memorize of the system. Then it gives an implementation solution for airborne SAR system by using high speed DSPs and huge amount memorize, with which the real time signal processing of the imaging algorithm combined with the motion compensation can be realized.

A new technique to generate microwave radiometric images from the optical images is introduced in this paper. This technique can be applied in generating the source images for purpose of image-matching navigation. For comparison, two microwave radiometric images are presents, one is imaged by a 8mm ground-based microwave radiometer and the other is generated by this technique.

Spectral image sensors provide images with a large umber of contiguous spectral channels per pixel. This paper describes the calibration of spectrograph based spectral imaging systems. The relation between pixel position and measured wavelength was determined using three different wavelength calibration sources. Results indicate that for spectral calibration a source with very small peaks,such as a HgAr source, is preferred to arrow band filters. A second order polynomial model gives a better fit than a linear model for the pixel to wavelength mapping. The signal to noise ratio (SNR)is determined per wavelength. In the blue part of the spectrum,the SNR was lower than in the green and red part.This is due to a decreased quantum efficiency of the CCD,a smaller transmission coefficient of the spectrograph,as well as poor performance of the illuminant. Increasing the amount of blue light,using additional Fluorescent tube with special coating increased the SNR considerably. Furthermore, the spatial and spectral resolution of the system are determined.These can be used to choose appropriate binning factors to decrease the image size without losing information.

Crop physiology analysis and growth monitoring are important elements for precision agriculture management. Remote sensing technology supplies us more selections and available spaces in this dynamic change study by producing images of different spatial, spectral and temporal resolutions. Especially, the remote sensing data of high spectral and high temporal resolution will play a key role in land cover studies at national, regional and global scales. In this paper, Multi-temporal Index Image Cube (MIIC) is proposed, which is an effective data structure for the parameterization of multi-dimensions spectral curve. MIIC is very useful for supporting the dynamic analysis on vegetation phenological and physiological characters. Based on multi-temporal meteorological satellite data and multi-temporal ground spectral measurement data, the temporal characters of different vegetation physiological parameters are contrasted and analyzed from temporal index image cube. In addition, MIIC also has very wide use in hyperspectral remote sensing applications.

For rapid and steady collection of high spectral resolution airborne data, a narrow band multispectral digital camera system (MDCS) was developed and tested in the year 2000. The MDCS was built based on three 1024x1024 pixels, 12bits digitalized area CCD cameras, whose FOV and IFOV are about 20 degree and 0.34 mrad respectively. Precise exposure control and synchronic trigger control are provided in this system, and the problem of collection and recording of large digital image data has been well solved. The center wavelength and bandwidth of the bandpass optical filters in this system can be customized to fit different application. The filter bandwidth can be changed from 10 to 25nm, and the filter center wavelength can be changed from 400nm to 900nm. The 10nm bandpass filters centered at 555, 650, 725nm and 650, 725, 825nm were used for agriculture research in the test phase. High spatial-resolution multispectral images were acquired on December 5, 2000 with the MDCS. At an altitude of approximately 3500 meters, the spatial resolution was 1.2 meter. Image processing was made for improvement of the image quality. The image restoration of motion-blurred image is discussed in the paper.

In this paper, we study the technique of high precision terrain map generation using imaging laser radar. For getting equal distance square mesh terrain map, we use the multinomial function, the spline function, and the even B-spline function to fit the original scanning data, then resample using above results. In order to get real time terrain map, we have set up two model of resampling. One is the adjacent model, another is sub-adjacent model. The simulated experiment was finished on different conditions using different resampling methods. The simulated experimental results indicate that the imaging precision is higher using the spline function resampling than using other one, using the sub-adjacent model resampling has less error than using the adjacent model, but it has a less lag in time than using the adjacent model.

Based on analyzed deficiency of traditional radar signal transmit system, advance optoelectronics technology and In- System Programmable (ISP) technology are used to realize general radar signal fiber transmit system. A special synchronous multiplex/demultiplex circuit is brought forward and realized with in-system programmable Large Scale Integrated (ispLSI). And Radar signal transmit quality is improved obviously.

We analyze superresolution techniques in optoelectronic imaging systems. Sampling is necessity in optoelectronic imaging systems. Aliasing will occur when sampling an object containing frequencies exceeding half the sampling rate, i.e., high frequencies will overlap with low frequencies and significantly degrade image quality. The schemes of expanding the unaliased bandwith and recovering the aliased high frequencies are referred to as superresolution techniques here, including techniques of microscanning (MS), microzooming (MZ) and dual-magnification (DM). They all exploit series of frames of the same object, but different in image acquisition and processing. MS using subpixel shift frames is simple in reconstruction, but complex in shift machine. MZ using different magnification frames is less complex in machinery, but needs data processing in spectrum. DM can achieve the results similar to MS or MZ with fewer undersampled frames, but only suitable to periodic targets. Theoretical comparison and computer simulations about these techniques are presented.

The confocal imaging system has wide applications in medical imaging for its sectioning and suppression of stray light. As we know, the small spot stimulated by laser light or itself in biotissues can provide rich multispectral information. In a conventional confocal imaging system, we sometimes neglect the spectral information or just use a single spectrum. Here we present the combination technology of confocal imaging and multispectral image. We simply use the grating to obtain the spectral distribution of the focus point. So we can acquire the multispectral image of the whole biotissue. The principle of design is analyzed and discussed. The reconstruction method is also discussed.

The multi-spectal or hyper-spectral imager provides three dimensional description for spatial and spectral intensity distribution of objective scenes and it can be a powerful tool for remote sensing in much application. There are several approaches to collection three dimensional data cube of image, and most of them require a scanning mechanism. Therefore those methods are difficult to record both spectral and spatial information of a dynamic scene such as a missile in flight, changing rapidly red tide, and so forth. Computed Tomography Imaging Spectrometer (CTIS) is a new branch application of computed tomography technology. The CTISs with different type configuration have been reported by a few of author. Two main types of CTISs are proposed. One is to use the two dimension gratings like Dammann gratings, the two 1-D gratings placed in orthogonal way, or three 1-D gratings separated by 60 degree. Due to its temporally and spatially non-scanning technique, this type is capable of capturing the flash events and can be used for instantaneous spectral imaging. The main problem of these CTISs is that the diffractive efficiencies are not only depended on the various wavelength but also on the different diffractive orders. That will effect the reconstruction algorithm and its results. These problems lead to reduce the signal to noise and dynamic range of spectral imaging system. The other is to take the approaches such as rotational spectro-tomography, or grating combined with a rotational direct-view prism. The advantages of these approaches are a)high throughput and b)easy to obtain more uniform data of different projection, but its disadvantage is obvious that the moving parts must be adopted. In our work, a principle and configuration of CTIS is indicated, and especially, a novel hybrid diffractive-refractive element is proposed, which is a combination of an array of optical prisms and one dimension holographic gratings. It can provide the uniformity of performance of projection data in different directions. As a result of high throughput and non-scanning parts, it can collect transient spectral imaging. Furthermore this grating is easily fabricated relatively. The capability enables it to be a prospect of applications in variety of application fields. This type of CTIS with the novel hybrid diffractive-refractive element is designed.

A method to obtain the image information from spectrum and reconstruct the image is presented in this paper. The theoretical explanation, experimental results and application to the imaging method are also introduced.

Trichosanthin(TCS), a ribosome inactivating protein extracted from the root tuber of a traditional Chinese medicine herb Tian Huo Fen(THF), possessed abortifacient, anti-tumor and anti-human immunodeficiency virus(HIV) activities. For centuries in China, THF has been used as an effective folk medicine to terminate early and midtrimester pregnancies and to treat ectopic pregnancies, hydatidiform moles and trophoblastic tumor. We observed the changes in reactive oxygen species and intracellular calcium in mouse embryos induced by TCS with confocal laser scanning microscopy in combination with the fluorescene diacetate (DCFHDA) and Fluo-3-AM. The results indicated that TCS induced increase in intracellular calcium and production of reactive oxygen species in mouse embryos , and TCS inhibited the development of mouse embryos effectively. Mouse embryos of different developmental stages before implantation are used in the experiments. This provides new insight into mechanism for abortifacient activity of TCS.

A new technique is developed for the data fusion of multispectral image and panchromatic image. The intensity component of fusion image is modified by combine multispectral information and high-resolution information, which is determined by image edge intensity. The fusion image is reconstructed by means of the inverse IHS transform. Experiment comparison shows that our method performing better in preserving spatial resolutions and color content than that of traditional IHS transform technique and wavelet transform fusion method.

Remote sensing image change detection techniques are widely used in environmental change detection such as landuse change monitor, flood monitor. Many change detection techniques are used in practice today. This paper reports the development of techniques based on artificial neural networks and presents a new method of integrating artificial neural networks (ANN) with gray system theory for remote sensing image change detection. Gray system theory, founded by Professor Deng Julong, can handle undetermined problem .It is effective when the sample datum can not satisfy some distribution. The accuracy of image change detection based on traditional ANN is influenced by some factors such as network architecture, training set. The number of hidden layers and the number of nodes in a hidden layer are not easy to deduce. The traditional neural network architecture which gives the best results for image change detection can only be determined experimentally, and this can be a lengthy process especially for large image. This paper presents a new method that the number of nodes in hidden layers is deduced by using gray correlation analysis in gray system theory. A neural network based change detection system using the backpropagation training is developed. The trained three-layered neural network was able to provide information of changes and detect land-cover change with an overall accuracy of 91.3 percent. Using the same training data, a maximum-likelihood supervised classification produced an accuracy of 85.1 percent. The experimental results by using multitemporal TM imagery and SPOT imagery. Findings of this study demonstrated the potential and advantages of using neural network and gray system theory in multitemporal change analysis.

In this paper an adaptive approach for color image enhancement is proposed. The approach is based on saturation feedback technique and adjusts enhancement coefficient automatically by genetic algorithms. Experimental results on color aerial images show the feasibility of the proposed method.

A constrained linear discriminant analysis (CLDA) approach is presented for hyperspectral image detection and classification. Its basic idea is to design an optimal transformation matrix which can maximize the ratio of inter- class distance to intra-class distance while imposing the constraint that different class centers after transformation are along different directions such that different classes can be better separated. The solution turns out to be a constrained version of orthogonal subspace projection (OSP) implemented with a data whitening process. The CLDA approach can be applied to solve both detection and classification problems. In particular, by introducing color for display the classification is achieved with a single classified image where a predetermined color is used to display a specified class. The real-time implementation is also developed to meet the requirement of on-line image analysis when immediate data assessment is critical. The experiments using HYDICE data demonstrate the strength of CLDA approach in discriminating the small targets with subtle spectral difference.

Imaging data from airborne hyperspectral sensor possess strong spatial correlation within each band as well as strong spectral correlations between the data in that band and the data in its two adjacent bands. These close correlations inherent in the imaging data provide access to evaluating image quality of different channels. According to the above characteristics, a method is described in the present paper for hyperspectral remote sensing imagery quality assessment. The method, which is based on analysis of within_band spatial autocorrelation information and between_band comparative analysis, is developed for that purpose. Convincing result is achieved when this method is applied to imaging data acquired from PHI.

In this paper, a morphological wavelet principle and a SAR image enhancement technique are introduced, and the image enhancement technique performance based on morphological wavelet transformation is proceeded using SAR image data. In wavelet transformation domain, an image enhancement method of 2D Butterworth filter constructed is applied, good enhancement effects are obtained, and an extraction edge and texture study of applying prewitt operator is taken. Because of morphological wavelet transformation good performance, a high efficiency in image enhancement and texture obtaining is gotten. The research results show that the SAR data with Speckle and Pepper and Salt noise is enhanced very well by nonlinear enhancement processing in morphological wavelet domain, and the ocean target characteristics is clearer, such as ship, ship tail and interalwave.

This article introduces the multivariate change detection which is based on the established canonical correlation analysis. It also proposes using post processing of the change detected by the multivariate change detection variables using maximum autocorrelation factor analysis. Differing from traditional schemes, the strategy takes two multivariate satellite images covering the same geographic area acquired at different points in time as a random whole sample and transforms two sets of random variables into one set of new random multivariate by using the so-called canonical transformation introduced in the paper. In doing so, the correlation between spectral bands in the same image and in the two different images is removed out as much as possible that the actual changes in all channels simultaneously can be accurately detected. The strategy is invariant to linear scaling. Therefore, it is insensitive to differences in gaining settings in a measuring device, or to linear radiometric and atmosphere correction schemes. The experimental results show the fact that the presented method is exactly creditable and effective on multivariate change detection of remote sensing satellite data.

Rough sets theory is a new mathematical tool to deal with vagueness and uncertainty. It is a promising soft computation method of intellective information processing. A method for speckle intelligent filtering of SAR image based on rough sets theory is presented in this paper, This method is proved to be practical and efficient by the experiments.

Three main models (linear, logarithmic, and exponential model) are introduced in this paper for monitoring the current soil moisture according to the Apparent Thermal Inertial (ATI) theory based on NOAA AVHRR data. The processing steps of these models are presented in detail as well. The three models are tested and validated by using ground truth data. Then the environment that has caused drought is also discussed in the paper by using ancillary data including DEM and LANDCOVER. The results reveal that all the three models can meet the need of the large scale drought monitoring, and that the application of ancillary data not only makes the quantitative analysis of drought distribution and it causes possible, but also validates the models correctness in a certain extent.

In this paper, we present a novel technique for redshift identification. Redshift is a key parameter of celestial spectrums. In the literature, there are few reports on redshift identification due to either no many people working on the problem or perhaps industrial confidentiality. Our technique is a pseudo-triangle technique. It consists of the following three major steps: firstly, the 3 wavelengths corresponding to the 3 highest intensity values of an unknown spectrum are selected to construct a pseudo-triangle, and the largest angle of this triangle is calculated which is independent of redshift value. Secondly, the obtained angle is used as an index to retrieve the corresponding 3 model wavelengths via a pre-calculated look-up-table, which is composed of all the combinations of all the feature wavelengths of the model spectrum. And finally based on the 3 corresponding wavelengths, the corresponding redshift value is derived. The main characteristic of our technique is its simplicity and efficiency, which is demonstrated by experiments on simulated data as well as on real celestial spectrums. It is shown that the correct identification rate can reach as high as 86.7%. Taking into account the high noisy nature of celestial spectrums, such a result is considered a good one.

The conventional filters can't achieve good effects in reducing speckle for high resolution single-look spaceborne synthetic aperture radar(SAR) images. In this paper an algorithm on efficiently reducing speckle is developed. This algorithm uses multiple structuring elements to replace common structuring elements so as to create an omni-directional multiple structuring elements soft-morphological filter whose weight values can be obtained through the improved impulse Bp neural network(NN) self-adaptive method. The performance of this algorithm is analyzed in detail. Finally, the raw data of RADARSAT is used to demonstrate its efficiency. The result shows that the filter can bring better effect than other filters.

An algorithm for retrieving land surface temperature (LST) was developed by using multiple linear regression analysis and path analysis based on the National Oceanic and Atmospheric Administration's Advanced Very High Resolution Radiometer (AVHRR) remotely sensed thermal infrared data. The analysis results indicated that the brightness temperatures of the AVHRR thermal infrared channels 4 and 5 should be involved in retrieving LST, and showed that it became difficult to disentangle the effects of the brightness temperatures of the AVHRR channels 4 and 5 on LST. Therefore, a split-window algorithm for retrieving LST was developed by introducing the difference of the two brightness temperatures.

Due to the complex signal-dependent nature of the speckle in SAR image, it is more reasonable to use different speckle descriptions and despeckling filters for different kinds of regions. A multi-description despeckling approach is presented in this paper. Based on the local statistics, the x² test is introduced to segment the SAR image into the Gamma-distributed homogeneous regions and the more fluctuant heterogeneous regions. Then a MAP filter is used in the homogeneous regions, and a modified median filter is utilized in the heterogeneous regions. X-band airborne SAR image and synthetic image are used for illustration and comparison.

Landsat TM imagery is being used to monitor landcover change in a number of operational projects in Australia. Most recently, the Australian Greenhouse Office (AGO) has undertaken a project to provide historical monitoring of changes in woody vegetation across the continent. In order to map and monitor landcover changes in a consistent and comparable manner across broad areas, it is highly desirable to have a consistently calibrated numerical base. To this end, the AGO has supported the creation of a rectification and calibration base for Australia, using Landsat TM data; 369 Landsat 7 images from the period July 1999 to September 2000 were purchased from the Australian Centre for Remote Sensing (ACRES). This paper describes the processing and production of the calibration base from these images.

Fuzzy C-means (FCM) is an unsupervised clustering technique and is often used for the unsupervised segmentation of multivariate images. The segmentation is based on spectral information only and geometrical relationship between neighboring pixels is not used. In this paper, a semi-supervised FCM technique is used to add geometrical information during clustering. Geometrical information can be adapted from the local neighborhood, or from a more extended shape model such as the hough circle detection. Segmentation experiments with the Geometrically Guided FCM (GG-FCM) show improved segmentation above traditional FCM such as more homogeneous regions and less spurious pixels.

Traditional standard products are processed and supplied on an individual scene basis. It reduces the complexity of data product generation and is of benefit to product catalogues. In fact, a lot of product users need consecutive multi-scenes. Putting them together is necessary and brings errors. In this paper, a pass processing approach will be discussed. This approach is based on pass model, using a few Ground Control Points (GCPs) to improve the satellite ephemeris and attitude data. It has been proved that this method is more effective than individual one.

In this paper, we make research on the analysis and processing for some data from an in-flight absolutely radiometric calibration experiment of imaging spectrometer. We use two kinds of methods to retrieve the surface reflectance of some targets on the ground and make a comparison with measured results.

Now the developing research of Agent can help operators to do the routine assignments, by which we can economize the precious resources and improve the real-time image analysis of the computers. This paper firstly makes a brief introduction of the Agent conception. Then we make some discussions about the multispectral images of a certain area, which is based on the concept of Agent. The main objects of this paper are inspections of forest (grassland) fire. The purpose of this paper is to propose three stages with which Agent could monitor the wildly areas and make decision automatically, without operators' intervention. First stage, if the value of pixels are more than a given threshold, Agent will give the operators an alarm and notify the operators that there are something happened; Second stage, analyze data and self-learning; Third stage, according to the database and knowledge database, Agents make decisions. As the decisions will be influenced by many factors, so some models, such as heat sources model, weather model, fire model, vegetation model are needed.

A new filter for the multichannel image based on the technology of Fuzzy Neural Network (FNN) is proposed. Multichannel images obtained by remote imaging technology are often corrupted by noises such as gauss noise, impulse noise and speckle noise. Tradition linear filters and some nonlinear filters can't fulfill the task of removing the mixed noises clearly and quickly. The biology vision systems are sense to wide bind of light and their network structure can perform in parallel manner and filter mixed noises, which inspires to the design of FNN image filter. The neural network structure in the presented filter is good at learning from sample data and parallel calculating while the fuzzy mechanism embedded in the network can detect different patterns in order to remove noise and keep details and textures. The filter has three function layers. In the first layer, corrupted image is introduced into five fuzzy sets characterized by membership functions. In the middle layer, fuzzy reasoning hidden in the neural network detects the data pattern and indicated noise pixels. In the output layer, the defuzifieation process is done and the restored image is obtained. A learning method based on the advanced genetic algorithm is adopted to adjust the network parameters from a set of training data. Experimental results shows that the FNN filter is effective in noise canceling and simple to be realized by hardware.

In the paper, according to characteristics of SST image of satellite remote sensing, we study the detection and recognition of Ocean Meso-scale Eddy. In the research, several methods of image processing are used.

We propose a general framework to estimate vegetation parameters from multidimensional remote sensing data using physical models and a moderate amount of ground control data. This framework is exemplarily demonstrated for winter wheat fields imaged by the Daedalus multispectral scanner using physical radiative transfer models, like the SAIL and PROSPECT model, combined with a linear empirical model. Results show the invertibility of the model for leaf area index, chlorophyll content, dry matter, and water content. The attained accuracies of these parameters are compared to the amount of ground control data. The vegetation parameters estimated with help of the devised method are intended to be used to derive information about soil heterogeneities that are important for precision farming.

Multiwavelet transform, a new notion addition to wavelet theory, offers simultaneous orthogonality, symmetry, short support, and vanishing moments, which are not possible with scalar wavelets. In this paper, a novel fusion method for Landsat Thematic Mapper (TM) and Synergetic Aperture Radar (SAR) images using multiwavelet transform is proposed. The approximation subband of the fused multiwavelet transform representation is from that of Landsat TM image, and the detail subbands are constructed by fusing those of Landsat TM and SAR images. The proposed method was compared quantitatively with intensity-hue-saturation, principal component analysis, and wavelet transform. It was found that the presented technique is clearly better than those methods in preserving both spectral and spatial information.

Environment preservations of the forest become world-widely close-up. The remote sensing technologies have been used for observation of the forest. However, the remote sensing by satellite is seriously hindered by atmospheric phenomena and cannot get the detailed data, which are necessary for precise analysis of the state of forest. So the observation system on the ground is needed. In this paper, we develop a multispectral infrared camera system to investigate the continuous spectra of mid-range infrared whose wavelengths cover around 3000nm to 5000nm. We define hue value in the case that the number of primary image is more than three. An algorithm to calculate the hue value from multispectral infrared images is proposed to obtain spectrum change of plants. An experiment of observation using the multispectral infrared camera system when the plant is growing week from cutting water off is shown.

This paper introduces an application of land-cover changes detection based on remote sensing image analysis. The main purpose of this study is to monitor urban expansions and the changes from cultivate land to construction land with multitemporal TM image and SPOT image. Data fusion, image enhancement and texture analysis techniques are used in this case study. The effectiveness of the proposed approach is confirmed by experimental results carried out on remote sensing data of Jinzhou area in Dalian, China.

Remote Sensing data acquired by various sensors during the last 40 years for the Turfan Basin in northwest China were compiled and processed to map and monitor changes of surface features related to hydrological conditions within the basin. Various image processing techniques were used to map surface features directly or indirectly related to former lake levels and groundwater tables for inferring conditions of the past climate. Special attention has been paid to the distribution of open water body, playa salt crust, waterlogged and salinized soils, and desert vegetation. With improved spatial resolution and more spectral bands, data acquired by ASTER sensor were explored to get additional information for discriminating these surface features. The compiled data indicate that Aiding Lake was a seasonal lake as early as in 1961. For the last 40 years, Aiding Lake has existed as a lake with open water only during the spring and early summer months. The size of the lake water body and timing and duration of the lake phase have varied year to year, suggesting that Aiding Lake is highly sensitive to climatic fluctuations and it is an ideal location for detecting possible impacts of global climatic changes.

The change detection is one of the important topics in multi-temporal remotely sensed data and significant effort has gone into development of change detection method. This paper will introduce the method for multi-temporal change detection with multivariate analysis, which is based on the canonical correlation analysis and the orthogonal transformation. Moreover, a case study with NOAA/AVHRR data will also be presented.

Geological structures used to be faint and blur in remotely sensed image as they are usually buried or hidden under the ground. In spite of that, the information of them can be found out in single band or multi-bands of the multi-spectrum data. Geological structures can be considered as image anomalies upon complex background. It is an important approach for geological structure enhancement to enhance the difference between the anomaly and its background in single band or multi-bands. Characteristics of spatial spectral distribution is thus special significant for image processing. Along this way, we improved on two methods, mean-residue (MR) and selective principal component analysis (SPCA), with emphasis on spatial spectral analysis, to enhance geological structures. Applications of the methods to actual TM data have arrived at good results. The keys of the two methods are respectively the determination of filter kernel and the selection of band pair.

Optimum band selection for visual interpretation and classification is an interesting task in conventional remote sensing, and, as an effective means to mitigate the curse of dimensionality, which has assumed growing importance with the availability of hyperspectral remote sensing data. In determining three-channel combination for a informative display in an image-cube and determining feature combination for fast classification, band selection is regarded indispensable in hyperspectral remote sensing. When applied to data acquired from a hyperspectral sensor, which is usually with a set of hundreds of band, however, conventional band selection procedure, of any criterion, becomes not viable with respect to the particularly time consuming. To cope with this pitfall, a method based upon genetic algorithm is proposed in this paper. An experiment, with a 121 band data set, demonstrate the efficiency. For simplification, the algorithm is designed to choose a combination which produces the most informative visual result when used as the top color preference in an image- cube. With little modification in criterion, the algorithm can be used to select features for classification purpose. The corresponding result is also presented in this paper.

Solar eclipse would result in a shadow on the earth which might be observed by orbiting satellite and shown in a satellite image. Even though this rarely happens, a sequence of images contaminated by solar eclipse has been obtained from a geostationary meteorological satellite. The paper describes how the shadows in the imagery series have been eliminated with image processing technique. A 2D model is designed to simulate the shape and variation of the solar eclipse shadow so that the shadow can be cleared to a certain extent. Cloud tracking and pattern recognition techniques are also applied in verifying the algorithm.

Extracting Airport Runways in Infrared aerial images is a task of great important for many applications. Some traditional methods rarely give a complete outline of the desired structures. A new method has been proposed to extract the airport runways. It contains three steps: Firstly, we extract the straight lines form the original infrared aerial images. Secondly, we get rid of some short lines on the basis of rule-based knowledge that we have got form field experts. Thirdly, we confirm whether there is an airport according to the position interrelation among these straight lines and rule-based knowledge. If there is an airport, we shall extract the runway regions from it. Our techniques have been tested on several typical of infrared aerial images that contain airports. The experimental results show that all the airports have been found, and the runways have been extracted entirely. This algorithm also adapts to aerial optical images.

In this paper, a novel unequally spaced sampling method is proposed according to the characteristic of rosette scanning infrared image system to meet the requirement of short processing time and good image quality. The influence of sampling parameters is discussed and the optimum values of the parameters are determined by experiments for this real-time infrared image system.

An evaluating platform for an infrared countermining system was introduced to test the interference and anti- interference performance of each side in the battlefield. The final goal achieved by this platform was to search the best anti-interference algorithm for the infrared guiding system and select the most effective interference mode for the other side. The battlefield environment included the attacker (missile), target (ship), backgrounds (sea and sky) and interference (smoke bomb or flame tracer). In order to select the best anti-interference algorithms for the guiding system, a database which was composed of various recognizing and tracking algorithms was built. To achieve this in hardware, a photoelectric tracker (PET) and a DSP system was used to simulate the infrared guiding missile instead of an actual guider which has only fixed algorithm. There were two evaluating mode, real-time and offline. In real-time mode, the recognizing and tracking algorithm was fixed according to simulate the actual state of the missile. In offline mode, different algorithm could be selected to attain the best anti-interference by replaying the recorded scene.

The Multi-targets and Backgrounds Simulation (MTBS) was a computer model which generated infrared dynamic images detected by the shipborne infrared warning system in battlefield situations. A 3D scene was built. The targets distributed in a hemisphere whose center was the sensor of the infrared warning system. The background ought to be the sky or the sea. It could analyze the infrared characteristic of the targets and backgrounds. Combining the atmosphere attenuation, the energy accepted by the sensor could be determined. All the parameters of the targets and backgrounds were managed orderly. With the 3D scene and battlefield effects, dynamic images could be obtained with computer. MTBS was composed of eight modules linked by a main routine. Each module performed its specific task in the overall simulation framework. MTBS also provided an effective tool to evaluate the performance of infrared warning system in laboratory instead of outfields experiments.

An ocean surface model for synthetic 3-5micrometers and 8~12micrometers infrared imaging application has been developed which is based on the JONSWAP wave spectrum. First, the slope of sea surface was calculated by the JONSWAP wave spectrum. Next, the emissivity of the sea surface was obtained from the slopes. After that, the geometric relation of sun, sky and facet of sea is used to determine whether the radiance of the sun is reflected by this facet or not, and the radiation at detector is obtained. Later on, the noise of detector was simulated by average method. Finally, 3-5micrometers and 8~12micrometers infrared images of sea background is generated. The texture of 3-5micrometers and 8-12micrometers simulated infrared images is statistically analyzed. The statistics includes temperature histogram, standard deviation analysis. The results show that the texture of 3-5um and 8-12um infrared images is similar in many statistics. But to the images that are derived from the high-reflective ocean surface incurred by the sun, the textures have some differences. The similarities and differences of the textures are explained in theory.

With the development of the infrared imaging guiding missile, more and more simulations are required. These simulations often include: the close-loop simulation, the open-loop testing, the performance evaluation of the anti- interference and etc. In order to carry out these simulations, infrared images which include targets, backgrounds and interference are required. A hard-in-loop simulation was presented to generate these dynamic images. The infrared image, as the signal resource of the guider, is composed of four models: target, background, interference and atmosphere. The infrared radiation characteristic of the target, background and interference is the direct factor which influences the infrared images. And the existence of the atmosphere brings on the radiant attenuation. The targets and interference are combined into the backgrounds by the technology of computer image generating (CIG). Two types of schemes of the hard-in-loop simulation are given: open-loop and close-loop. This can achieve the arbitrary changing of the battlefield, accordingly, the battlefield adaptability of the infrared imaging guiding weapons can be assessed in laboratory. It is no other than an important performance that is difficult or not to be assessed in outfield experiments.

In this paper, infrared and low light level image processing theories and techniques in color night vision system are studied in detail. The characteristic of IR and low light level image is described and the significance of IR and low light level image processing in color night vision technology is pointed out. On the bases of these, IR and low light level image processing theories and techniques in color night vision system, including geometric calibration by control point method, image zoom by gray-level interpolation, IR and low light level image noise suppression techniques and gray-level transform by look-up table (LUT) processing, are presented. The theory models and calculating methods of every above image processing technique is described. Through IR and low light level image processing, the quality of the night vision images is improved significantly and the color night vision system is developed successfully.

Automatic vehicle classification (AVC) is the important prerequisite for automatic charging and administering system. In this paper, a new AVC system based on infrared plane detecting technique is introduced, and in order to dispose the data gathered by the infrared detecting system to classify the vehicle, fuzzy pattern recognition technique is used. By this technique, vehicles difficult to be classified by usual method can be classified correctly.

The algorithm and the architecture of the information processing system of infrared imaging guidance are studied. Wavelet transform can decompose the image in multi-scale, so the different frequencies components of an image can be selected. The dim target can be detected by wavelet transform. The dramatic increase in the computational speed provided by DSP offers the possibilities of processing large high-resolution images in real time. The authors have experience in processing high-resolution images (up to pixels, eight bits per pixel) using TI's TMS320C6201 DSP. We have developed a technique, which uses FPGA to perform wavelet transform to decrease the execution time.

With the help of attention mechanism, small target pre-detection can improve the efficiency of detection, and along with the waterline detection, the judgment of whether IR ship targets exist can be made. In this paper, the ability of parallel computation of pre-detection is analyzed. Starting from two points: the balance between communication cost and parallel computation and the efficiency of utilization of the system resources, how to divide the computation task is under consideration, so as to be adapted to pipeline structure and parallel structure to get the best performance. Also, the performance of the algorithm in different size images is investigated. At the end of the paper, experiments are used to prove that the analysis in this paper can provide guidelines of the real-time realization of the algorithm.

In this research work, an experimental investigation was performed for the determination of defects on aircraft materials. Infrared thermographic active approach was used for the location and identification of defects on cracked aluminum panels untreated and repaired with carbon or boron epoxy composite patches. Furthermore, the quantification of these detected defects was attempted, using selected image processing and analysis techniques. Thermographic image analysis using histograms and digital image processing were applied on these obtained thermal images with the intention of calculating the localized defected areas. The results from these image analysis techniques were compared and discussed in terms of developing an integrated approach for the quantification of defects on aircraft components.

In this work, infrared thermography was applied and investigated as a non-destructive tool in the assessment of materials and techniques for the protection of cultural heritage. Diagnostic studies on monuments and historic buildings, situated in Greece, were performed. Long wave infrared thermography was used on restoration and traditional - historic materials concerning architectural surfaces and historic structures for research purposes such as: the assessment of moisture impact to porous stone masonries and the evaluation of conservation interventions (materials and techniques) regarding, consolidation interventions on porous stone masonries, restoration of masonries by repair mortars, and cleaning of facades. The results of this work indicate that thermography can be considered as a powerful diagnostic nondestructive tool for the preservation and protection of cultural heritage.

In this paper, we apply blind deconvolution to infrared images. We concentrate on the important case where the blurring operation is spatially invariant; the point spread function (PSF) is more like a speckle, an out-of-focus, or their combinations with a little motion blur. The information suggestion on PSF is only Gaussian based spectra or shape. The restoration procedure involves estimating the PSF and the true image alternatively. For simplicity, we use gradient-based iterative method to solve minimization problem. The model is not complicated and we make the uncertainty of blind deconvolution into the choice of only a few parameters. Finally, deconvolution results are presented for both synthetic and real data.

The isotherm is an important feature of infrared satellite cloud images (ISCI), which can directly reveal substantial information of cloud systems. The isotherm extraction of ISCI can remove the redundant information and therefore helps to compress the information of ISCI. In this paper, an isotherm extraction method is presented. The main aggregate of clouds can be segmented based on mathematical morphology. T algorithm and IP algorithm are then applied to extract the isotherms from the main aggregate of clouds. A concrete example for the extraction of isotherm based on IBM SP2is described. Study of the result shows that this is a highly efficient algorithm. It can be used in feature extractions of infrared images for weather forecasts.

The new-type infrared thermal imaging technology should be a combination of infrared detecting and real-time processing. The intelligence ,modularity, miniature and real-time are the developmental subjects in infrared image processing In this paper, on the basis of the researching in the architecture feature of nowadays digital image processing system, the development. ideas and technology approaches of infrared image real-time processing are described .The new types of structure and technique of infrared thermal image real-time processing--adaptive time-space sequence flow network construction and feature weighting adaptive real-time processing method are presented firstly. According to the infrared image characteristics, the histogram of infrared image and the correlation inside frame or interframe are analyzed in detail, and some image enhancement methods are taken. These thermal image processing theory and algorithm are realized by hardware circuit. In PCB design, additional circuit noise and digital pulse interference are discussed deeply. On the basis of many experiments, the PCB design regulations with anti-interference are established. Finally, the Modularized Infrared Image Real-time Processor is finished with noise suppression and image enhancement. The design and circuit construction of the modularized infrared image real-time processor are given.

This paper presents a new nonuniformity correction (NUC) algorithm based on scene-matching for linear scanning focal plane arrays, especially for the rapidly developing TDI scanning arrays. The algorithm bases on matching every two adjacent images in which most parts reflect the same scene. The response relationship between detectors can be calculated from the matched part-pairs. Then it can be used to complete the nonuniformity correction and to adapt the drift of the nonuniformity with variations in environmental temperature. The experiments on real image sequences demonstrate the accuracy and robustness of the proposed algorithm.

A new band model was established in this paper. Co₂ CO, H₂O and CH₄ are the main infrared absorption gases in atmosphere. The infrared spectral absorption coefficient of the gases can be expressed in a simple function with the new band model, and the expression can be used to calculate directly. In a certain wavelength range, the mean absorptance and transmittance of the gases were calculated with the new band model, the error was in reasonable precision range by compared with experimental data.

Fourier-transform infrared spectroscopy is an efficient technique for the detection and quantification of molecules in gas mixtures. A passive infrared system was used to detect specific hazardous clouds in the atmosphere and sound an alarm in time. The Fourier transform infrared spectrometer is operated in the 8-12 atmosphere window region of the spectrum. A spectral identified system of hazardous clouds is based on neural network. The network is trained by the use of the backward-error-propagation algorithm. The identical results of the target spectra from the open-air have demonstrated that the detection-integrated system may be used to detect and identify the infrared spectra of hazardous clouds.

Focal plane arrays (FPA) of HgCdTe detectors with responses in the 3-5micrometers or 8-14micrometers bands have been successfully incorporated into infrared imagers. Using a variety of digital signal processing techniques, good imagery has been obtained but the image quality is limited by sampling effects at the focal plane that cause non-isoplanantism and aliasing. The verification of the operation of these detectors is an important step in the design and development process. Modulation transfer function (MTF) is a measurement of how well the system will faithfully reproduce the scene. As the MTF decreases, scene detail associated with those specific spatial frequencies will be reproduced with lower contrast. The development of focal plane arrays with considerable on-chip processing, notably time-delay integration, imposes new and severe requirements on the instrumentation to establish the MTF of the integrated focal plane arrays.1 This paper details a method for measuring the MTF for such arrays. Several methods to derive the MTF from the experimental data were investigated; the preferred approach is presented.

The uncooled IRFPA thermal imaging systems, which are designed to operate at room temperature, will make the sensors less expensive, lighter, and more reliable. The sensor of this infrared FPA imaging system is the 320x240 LW IRCOMS uncooled integrated microbolometer detector. The time sequence required by the detector is described. The timing control circuit of the camera is designed using Altera EPLD.

Some applications of synthetic aperture radar (SAR) need SAR data processing in real time or near real time such as data evaluation and disaster monitoring. Two quick-look algorithms based on FIR filter using in spaceborne SAR are introduced in this paper. The raw data is preprocessed by FIR filter and subsampled with low sample rate, which can reduce data and computation amount to generate low resolution SAR image. The first algorithm designing a two-dimension filter prefilters the raw data and the other method uses two one-dimension filters in the procedure of processing the data with R-D algorithm. The performance of the two algorithms is analyzed. Finally, the raw data of RADARSAT is used to prove the algorithms. The result shows that the algorithms can be used in quick-look imaging effectively.

Based on the 3D noise model, main noise sources of single detector scanning, serial scanning, parallel scanning parallel output and staring systems are analyzed respectively. And it is assumed that all noise is Gaussian distributed. Then models are established. In addition, non-linearity is discussed and simulated. With the method of pixel-based processing, these models are implemented. Finally, the results are shown and discussed.

Based upon the static performance of thermal imaging system, the visual range prediction adapted for the extended source target and the spot target is studied. Two kinds of mathematical model and the equation of operating range are set up for the extended source target and the spot target. Meantime, the paper gives computer simulation results and presents a method of the visual range prediction of thermal imaging system to verify whether the thermal imaging system satisfied the technical requirement of the operating-range. Finally, in this paper, the visual range prediction and the test results are compared.