[Analysis of Influence of Pushroom Erros on The Data Reconstruction of Computational Imaging Spectrometer].

The technology of computational spectral imaging changes the traditional imaging modalities by introducing a coded aperture in the optical path to achieve transformation of the targets spectral imformation, then we can get the spectral data cube by reverse transform. This paper introduces the principle of a push-broom imaging coded aperture computational spectral imager. In practical applications, the matching error between the push speed and the frame rate can affect the accuracy of the spectral data reconstruction. The error terms are deduced based on the model of pushroom, the influences of matching errors to the spectral data reconstruction are anlyzed. And the second spectra derivative and strehl ratio are introduced as the evaluation parameters to, respectively, evaluate the reconstructed spectral and spatial image in the data simulation analysis. It showed that, when the accumulated error of a complete set of data is more than one pixel, the shading dramaticly area's reconstructed results are relatively poor, but the relatively homogeneous regions are affected small; when cumulative error does not exceed half pixel, strehlretio of each channel were above 0.9, and the lower of the spectral energy, the channels strehl ratio smaller, so the more the ranks of the coding template, the higher the platform's stability required.

[The Snapshot Imaging Spectrometer with Image Replication Based on Wallaston Prism].

In this paper, a kind of snapshot imaging spectrometer is designed by using image replication based on Wollaston prism. The system includes telephoto lens, collimator lens, Wollaston prism, the imaging lens and compensation filters. With this optical system, we can obtain two-dimensional information at different wavelengths of the same target through one time exposure. When Light beam is passing Wollaston prisms, in order to make sure that the different wavelengths will not overlap, the Wollaston prism needs to be designed with relatively large beam splitting angle. Then the incidence angle of the imaging lens is relatively large. It will increase the difficulty of the imaging lens design. We design, analyze the principle and the characteristics of the snapshot imaging spectrometer using image replication based on Wollaston prism, and design a complete set of imaging spectrometer system. The structure is complex; the aperture of the optical system must be well matched. In order to make sure that we can obtain good image quality through the single lens, but also the whole system, we design the telephoto lens which is imaging telecentric structure, the collimator lens which is objective telecentric structure. We use the multiple structures in ZEMAX to optimize the 16 spectral bands. At 56 lp·mm(-1), the MTF is close to the diffraction limit, the RMS of the SPT is in the Airy disk, proving good image quality.

[Fourier Transform Spectrometer Based on Rotating Parallel-Mirror-Pair].

In the temporally-modulated Fourier transform spectroscopy, the translational moving mirror is difficult to drive accurately, causing tilt and shear problems. While, a rotational moving mirror can solve these problems. A rotary Fourier transform spectrometer is recommanded in this paper. Its principle is analyzed and the optical path difference is deduced. Also, the constrains for engineering realization are presented. This spectrometer consists of one beamsplitter, two fixed mirrors, one rotating parallel mirror pair, a collimating lens, a collecting lens, and one detector. From it's principle, this spectrometer show a simple structure, and it is assembled and adjustmented easily because the two split light are interfered with each other after reflected through the same plane mirror; By calculating the expression of it's optical path difference, the spectrometer is easy to realize large optical path difference, meaning high spectral resolution; Through analyzing it's engineering design constraints and computer simulation, it is known that the spectrometer should get the high resolution sample by high-speed spinning motor, so it is easy to achieve precise motion control, good stability, fast measurement speed.

[Manufacture tolerance analysis of solid Mach-Zehnder interferometer in large aperture static imaging spectrometer (LASIS)].

The principle and instrumental structure of large aperture static imaging spectrometer (LASIS) were briefly described in the present paper, the principle of the Mach-Zehnder imaging spectrometer was introduced, and the Mach-Zehnder interferometers' working way in the imaging spectrometer was illustrated. The structure of solid Mach-Zehnder interferometer was analyzed, and discussion was made based on the requirements of field of view (FOV) in image space and single sided interferogram with a small portion around zero path difference (ZPD). The additional optical path difference (OPD) created by manufacturing and matching tolerance of two asymmetrical pentagonal prisms will lead to the displacement of shearing and OPD nonlinearity. It was showed that the additional OPD from non-common optical path structure of solid Mach-Zehnder spectrometer implies more requirements on the manufacture of this element, compared with Sagnac interferometer, for the matching tolerance of two asymmetrical pentagonal prisms to br lower than 0.02 mm. The recovery spectrum error caused by the OPD nonlinearity is lower than 0.2% and can be ignored.

[Multi-DSP parallel processing technique of hyperspectral RX anomaly detection].

To satisfy the requirement of high speed, real-time and mass data storage etc. for RX anomaly detection of hyperspectral image data, the present paper proposes a solution of multi-DSP parallel processing system for hyperspectral image based on CPCI Express standard bus architecture. Hardware topological architecture of the system combines the tight coupling of four DSPs sharing data bus and memory unit with the interconnection of Link ports. On this hardware platform, by assigning parallel processing task for each DSP in consideration of the spectrum RX anomaly detection algorithm and the feature of 3D data in the spectral image, a 4DSP parallel processing technique which computes and solves the mean matrix and covariance matrix of the whole image by spatially partitioning the image is proposed. The experiment result shows that, in the case of equivalent detective effect, it can reach the time efficiency 4 times higher than single DSP process with the 4-DSP parallel processing technique of RX anomaly detection algorithm proposed by this paper, which makes a breakthrough in the constraints to the huge data image processing of DSP's internal storage capacity, meanwhile well meeting the demands of the spectral data in real-time processing.

[Effect evaluation of optical magnification errors for coded aperture spectrometer].

In the present paper, we simulated the system optical magnification error's influence on the quality of reconstructed image, analyzed the variant of the coded aperture with different optical magnification, then proposed an accurate curve of image quality and optical magnification in 128 x 128 code template, which provide important references for the design and development of push-broom coded aperture spectrometer.

[Impact and improvement of the mechanical shutter on large-array filter-type multispectral imaging system].

The present paper analyzes the impact of mechanical shutter on the spectral image acquisition and processing of large-array filter-type multispectral imaging system. The final image quality relies highly on the mechanical shutter due to the fluctuation at exposure time. The conventional mechanical shutter's structure and driving method was analyzed to find out the key fact for its poor stability. An improved method of mechanical transmission and circuit driving was proposed. Laboratory experiments showed that with the improved design strategy, the maximum rate of change between adjacent exposures was reduced from 15.05% to 0.96%, which is a great improvement of the exposure time stability. Field test was also carried out and the results show that the combined color images are closer to the realistic targets and no abrupt color change iwas found. It's of great significance for practical application in multispectral image process, interpretation and target identification.

[Imaging quality evaluation of computational imaging spectrometry].

As a novel imaging spectrometry, computational imaging spectrometry (CIS) has the advantages of high throughput, snapshot imaging etc. However, there is little research on imaging quality evaluation of CIS system. In the present paper, a quantitive evaluation method for imaging quality of CIS system was presented. ISO 12233 chart was used as the objective source, and then imaging and reconstruction of the spatial-spectral information was provided. Calculating modulation transfer functions (MTFs) for the reconstructed images was considered as the criterion of the imaging quality evaluation of CIS system. The result shows that MTFs for single-frame sampling decrease rapidly with the aliasing spectral number increasing. When the number of the aliasing spectra is 9, MTF for the reconstructed image decreases by 50% compared to the original scene. This work helps better understand the pros and cons of CIS system and arrange the aliasing spectral number reasonably to reconstruct the object scene precisely.

[Interference hyperspectral data compression based on spectral classification and local DPCM].

In order to get a high compression ratio, according to the spatial dimension correlation and the interference spectral dimension correlation of interference hyperspectral image data, the present article provides a new compression algorithm that combines spectral classification with local DPCM. This algorithm requires spectral classification for the whole interference hyperspectral image to get a classification number matrix corresponding to the two-dimensional space and a spectral classification library corresponding to the interference spectra first, then local DPCM is performed for the spectral classification library to get a further compression. As the first step of the compression, the spectral classification is very important to the compression effect. This article analyzes the differences of compression effect with different standard and different accuracy of classification, the relative Euclidean distance standard is better than the angle standard and the interference RQE standard. Finally, this article chooses an appropriate standard of compression and achieves the combined compression algorithm with programming. Compared to JPEG2000, the compression effect of combined compression algorithm is better.

[Design of the airborne prism dispersive imaging spectrometer system based on Offner relay configuration].

Considering the development of the need for small volume and light weight of spectral imaging systems on airborne platform, the characters of several main imaging spectrometers was analyzed and the present paper focuses on introducing imaging spectrometers featuring a grating or a prism disperser. A compact hyperspectral imaging system based on Offner relay configuration is reported. Combining the characters of the concentric Offner relay, and knowing the system design parameters, two novel kinds of spherical spectrometer systems using dispersive elements in divergent and convergent beam were designed. Moreover, the system MTF, spot diagram and the curves of smile and keystone were analyzed. The results show that these two kinds of Offner spectrometer achieve the purpose of a compact form of remote sensing system with excellent imaging quality close to diffraction limit. At the same time, decreasing the smile and keystone to a small (0.1) fraction of a pixel assures uniformity of acquiring spectral data.

[Modeling and correction of the effect of detector lateral fringe error in the large aperture static imaging spectrometer].

When the large aperture static imaging spectrometry (LASIS) was used in spaceborne remote sensing, the spectrum could not be well recovered from the interferogram with detector lateral fringe error. In order to obtain the more accurate spectrum, the error must be corrected. According to the imaging principle analysis of LASIS, the imaging model with lateral fringe error was given, and the error correction method was presented. Finally, the error correction method was validated with the data acquired by LASIS system. The results indicated that the error was commendably corrected by the method mentioned above, and the accuracy of recovered spectrum was evidently improved.

[Design of a dual-channel Mach-Zehnder lateral shearing interferometer for the large aperture static imaging spectrometer].

Large aperture static imaging spectrometry (LASIS) is a kind of joint temporally and spatially modulated Fourier transform imaging spectrometry. In such instruments, lateral shearing interferometer is a key element, the most frequently used type of which is the Sagnac interferometer. In this configuration, one half of the light entering the interferometer backtracks and causes a great decrease in energy efficiency. The present paper proposes a modified Mach-Zehnder lateral shearing interferometer structure to tackle this problem. With the ability to produce the same lateral shear, it features the advantage of dual channel output. We present a ray tracing procedure to induce the general expression of the lateral shear as well as analyze the contributions of error sources to the shear accuracy. The results serve as a new idea for the design of large aperture static imaging spectrometers and can be used to instruct the design and optimization of this kind of imaging spectrometer.

[Research on endmember extraction algorithm based on spectral classification].

Spectral unmixing is an important task for data processing of hyperspectral remote sensing, which is comprised of extracting the pure spectra (endmember) and calculating the abundance value of pure spectra. The most efficient endmember extracting algorithms (EEAs) is designed based on convexity geometry such as pure pixel index (PPI), N-finder algorithm (N-FINDR). Most EEAs choose pure spectra from all pixels of an image so that they have disadvantages like slow processing speed and poor precision. Partial algorithms need reducing the spectral dimension, which results in the difficulty in small target identification. This paper proposed an algorithm that classifies the hyperspetral image into some classes with homogeneous spectra and considers the mean spectra of a class as standard spectra for the class, then extracts pure spectrum from all standard spectra of classes. It reduces computation and the effect of system error, enhancing the speed and precision of endmember extraction. Using the least squares with constraints on spectral extraction and spectral unmixing, by controlling the band average value of the maximum spectral redundant allowance to control the number of endmembers, does not need to reduce the spectral dimension and predetermine the number of endmembers, so compared to N-finder algorithm, such algorithm is more rational.

[Advance in interferogram data processing technique].

Fourier transform spectrometry is a type of novel information obtaining technology, which integrated the functions of imaging and spectra, but the data that the instrument acquired is the interference data of the target, which is an intermediate data and couldn't be used directly, so data processing must be adopted for the successful application of the interferometric data In the present paper, data processing techniques are divided into two classes: general-purpose and special-type. First, the advance in universal interferometric data processing technique is introduced, then the special-type interferometric data extracting method and data processing technique is illustrated according to the classification of Fourier transform spectroscopy. Finally, the trends of interferogram data processing technique are discussed.

[Comparison of two types of Fourier transform imaging spectrometry].

The comparison between "large aperture static imaging spectrometry (LASIS)" designed by author and "method for simultaneously measuring the spectral intensity as a function of wavelength of all the pixels of a two dimensional scene" (MSMSI) designed by Israeli researchers is given. The difference of the interferometers and imaging systems in LASIS and MSMSI is discussed. The nonlinear curve of optical path difference of MSMSI is simulated. Inspired by the operation of MSMSI, the time modulated Fourier transform imaging spectrometry based on lateral shearing interferometer was presented. The optical path differences of different field of view were given under the condition of regularly rotating the lateral shearing interferometer.

[Spectrum recovery methods for nonuniform sampling interferogram].

The interferogram acquired by imaging Fourier transform spectrometer (IFTS) can't be used directly and must be recovered. The spectrum recovery processes based on fast Fourier transform (FFT) is the traditional method which is used widely. For some IFTS, the nonuniform sampling of the interferogram is often occurrs. When the aliasing is neglected, the spectrum recovered by traditional method is often distorted. When the spectrum recovery processes based on Fourier transform are used, the precision of the recovered spectrum can be ensured, but the real-time processing requirement can't be satisfied. In order to acquire the precise recovered spectrum of the nonuniform sampled interferogram, the interpolation method and nonuniform fast Fourier transform (NUFFT) method were adopted. For the oversampled interferogram and partial undersampled interferogram, the spectrum recovery methods based on interpolation and NUFFT were presented respectively, and the applicability of these two methods is given. Finally, the computer simulation was performed, and the results indicate that NUFFT method is preferable to interpolation method not for undersampled interferogram but for oversampled interferogram.

[Comparison of correction methods for nonlinear optic path difference of reflecting rotating Fourier transform spectrometer].

The principle of reflecting rotating Fourier transform spectrometer was introduced in the present paper. The nonlinear problem of optical path difference (OPD) of rotating Fourier transform spectrometer universally exists, produced by the rotation of rotating mirror. The nonlinear OPD will lead to fictitious recovery spectrum, so it is necessary to compensate the nonlinear OPD. Three methods of correction for the nonlinear OPD were described and compared in this paper, namely NUFFT method, OPD replace method and interferograms fitting method. The result indicates that NUFFT was the best method for the compensation of nonlinear OPD, OPD replace method was better, its precision was almost the same as NUFFT method, and their relative error are superior to 0.13%, but the computation efficiency of OPD replace method is slower than NUFFT method, while the precision and computation efficiency of interferograms fitting method are not so satisfied, because the interferograms are rapid fluctuant especially around the zero optical path difference, so it is unsuitable for polynomial fitting, and because this method needs piecewise fitting, its computation efficiency is the slowest, thus the NUFFT method is the most suited method for the nonlinear OPD compensation of reflecting rotating Fourier transform spectrometer.

[Techniques for pixel response nonuniformity correction of CCD in interferential imaging spectrometer].

Based on in-depth analysis of the relative radiation scaling theorem and acquired scaling data of pixel response nonuniformity correction of CCD (charge-coupled device) in spaceborne visible interferential imaging spectrometer, a pixel response nonuniformity correction method of CCD adapted to visible and infrared interferential imaging spectrometer system was studied out, and it availably resolved the engineering technical problem of nonuniformity correction in detector arrays for interferential imaging spectrometer system. The quantitative impact of CCD nonuniformity on interferogram correction and recovery spectrum accuracy was given simultaneously. Furthermore, an improved method with calibration and nonuniformity correction done after the instrument is successfully assembled was proposed. The method can save time and manpower. It can correct nonuniformity caused by other reasons in spectrometer system besides CCD itself's nonuniformity, can acquire recalibration data when working environment is changed, and can also more effectively improve the nonuniformity calibration accuracy of interferential imaging

[Research on spatially modulated Fourier transform imaging spectrometer data processing method].

Fourier transform imaging spectrometer is a new technic, and has been developed very rapidly in nearly ten years. The data catched by Fourier transform imaging spectrometer is indirect data, can not be used by user, and need to be processed by various approaches, including data pretreatment, apodization, phase correction, FFT, and spectral radicalization calibration. No paper so far has been found roundly to introduce this method. In the present paper, the author will give an effective method to process the interfering data to spectral data, and with this method we can obtain good result.

[Spectrum recovery method for interferometric data].

The interferometric data acquired by Fourier transform spectrometer was a betweenness which couldn't be used directly. For the successful application of the interferometric data, spectrum recovery processing must be adopted. The conventional method based on inverse Fourier transform was frequently used for spectrum recovery. According to the principle analysis of the interferometric data, the mixing model was induced, and the spectrum recovery method based on the mixing model of the interferometric data was presented. Finally, the computer simulation for these two methods mentioned above was implemented, and the recovered spectra were compared with the standard spectrum. The simulation results indicated that the spectral precision of the new method was better than the conventional method, which provided a new scheme for spectrum recovery.