From small scale to medium and large scale 320×240, 640×480 uncooled focal plane array (UFPA), uncooled infrared technology has been developed rapidly in recent years, and 1024×1024 FPAs are expected in the future. The pixel pitch is also developed from 50～25 μm, and the sensitivity is improved largely, so that the thermal imaging system is applied successfully in military affairs. Part model has been armed in military and measured well. In the future, with the developing larger scale and smaller pixel of UFPA, the application of uncooled thermal imaging system will be wider, especially in light thermal weapon,such as light weapon vision, driver vision enhancer and handle thermal system, uncooled thermal imaging system is expected to replace the high price, low reliability, big bulk cooled thermal imaging system gradually in recent years.

Laboratorial measurement method of MRTD is discussed, which is important parameter of characterizing infrared imaging systems. The method for the improvement and alternative of MRTD put forward in international and the performance model are analyzed. The advantage and disadvantage among the MRTD-ID, MTDP and TOD are compared. In addition it is pointed out that TOD method will play an important role in assessing infrared imaging system in the future.

For infrared surveillance system, it is most difficult that the measurements are only angles not distance, which makes target location of fire control system be difficult to realize. In this paper, the outputs of infrared sensors and the angle parameters are used to set up a range estimation model, which provides an important theoretical reference for the research of systems.

Based on the analysis of the present testing principle and methods, a novel full-automatic testing design for the resolution is presented. With the three principal technical designs including programmed resolution chart , Cathegreen collimator and CCD image sampler, a unique testing equipment for the resolution of aiming system of laser range finder is developed, which can all be controlled, processed and displayed by computer. Because of small volume, light weight, high computerization and simple operation, the testing efficiency is improved greatly. It has a good prospect to be popularized and applied.

The situation of the pumping beam departing from the oscillating beam in a fiber-coupled laser diode end-pumped Nd:YAG laser is studied. A method of accurate spatial coupling between the pumping and the oscillating beam by use of CCD is presented. Using this method, the spatial position of pumping beam is adjusted, ideal Gaussian beam with ellipse rate of (0.98) and the M2 factor of 1.01 are obtained after achieving the best spatial coupling between the pumping and the oscillating beam approximately. This technology can be used in other diode end-pumped solid-state lasers(DPSSL) with different laser materials, and it is practical for the improved accuracy.

According to the principle of laser terminal guidance, some important performances of laser radiation indicator are analyzed and the numeralization test method is introduced. Using simulation echo wave and accurate time delay circuit to measure ranging ability ,using CCD to grasp laser facular to measure divergence of laser pulse, the digitized methods are realized. It can measure the energy of every laser pulse, interval, frequency precision of one exposure period. It also gives the sequence chart to describe them. The circuit and light path diagrams are described with their data processing method. For using four-channel structure, host computer communicates with slave single chip microcomputer, the size of modularization is reduced, and the man-machine interface and the reliability of software are improved.

A continue wave laser diode driver controlled by single-chip microcomputer (SCM) is presented. It can work in three operating modes——automatic current control (ACC), automatic power control (APC) and automatic voltage control (AVC). The sweep current can attain the level of milliampere from 0A to 2A, and the temperature could also attain to ±0.1 ℃ with PI control and the thought of integral separation. Complete laser diode protection is provided including independent current limit and power limit settings, output off shorting circuits and a slow turn on sequence. The results show that this driver has a lot of advantages, such as high intelligence, good noise suppression, high stability and convenience.

The gasifying phenomenon in laser etching is analyzed with the theory of heat conduction. According to different materials, appropriate working condition is selected, experiments of engraving, broaching and marking by using high-power laser are operated, and the etching results are compared. The results show that the method is consistent with the theory analysis and can be applied in many fields.

A USB interface firmware applied to the wireless laser communication system is formulated. This system is a USB peripheral that is based on AN2131Q in EZ-USB serial of Cypress. After a USB peripheral is loaded, AN2131Q runs the firmware to manage the USB peripheral and communications with the host. How to use EZ-USB development kit and interrelated software to debug the system is discussed.

Inputting the velocity of moving target to the automatic tracking system of TV or infrared is an effective method to improve the tracking accuracy in electro-optical theodolite. In order to obtain the velocity of moving target, the methods of identifying the delay time of TV or infrared off-target and Kalman filter are designed. Fitting the TV or IR off-target for the code position of electro-optical theodolite and forecasting the velocity of being tracked target can compose compound control system. The simulation and test results show that velocity of moving target can be forecasted by fitting the target-missing quantity for instrument position quantity using the methods and tracking precision can be improved in servo-control system.

Firstly, the difficulties of dim target detection are pointed out in this paper. Secondly, all methods are classified as DBT and TBD according to the order of detection and tracking, the mathematical models based on the two kinds of methods are presented and the mainstream methods of each kind of methods are also introduced. Thirdly, the executive performances of the two kinds of methods are compared. In the end, some promising research directions in this field are proposed .

In view of the difficulties in detecting long-distance target, such as low SNR, relative movement in frames and noise, a preprocessing algorithm using morphological dilation before accumulation and tracking is presented to enhance the target. Wavelet transform is adopted to remove the correlation of 1/f noise in parallel scanned infrared image and keep down the gaussian noise effectively at the same time. Furthermore, the relationship between threshold, detected probability and faulty detected probability of point target in single image is also expatiated. The experimental result indicates that the algorithm can effectively detect the point target with SNR=2.

When the impending antiship cruise missile's motion vector alines with the axis of IR detection system, it is difficult to detect and track it by its trace analysis. A novel spatio-temporal filtering method is presented, which employs wavelet analysis and pixel value changing measurement. The images used in the experiments are composed of some real pictures and man-made point target (signals,) and the method is proved to be with high accuracy and sensitivity and low false alarm rate.

A new edge detection method based on the fusion of soft morphology and Sobel operator is developed. The omnidirectional morphological elements are designed to do recursive soft morphological operation to enhance operation speed and then delete noise. The rule of fusing the edge obtained by morphology and the edge obtained by Sobel operator is established. And the experimental results show that the clear and continuous edge can be got. This method has good robust to noise and can get reliable edge in noisy image.

Because of work at all day and all weather, radiometer can provide the specific information which infrared and optical detectors can′t provide. However it is urgent to improve the image resolution. On basis of studying the fusion algorithm for multisensor, a new approach to fuse the millimeter wave measuring image and optical-based simulated image by wavelet transform is presented. The simulation result shows that fusion image has more complementary information than the measuring image and possesses higher resolution as guidance matching map.

The LBRDF(1.06 μm) ,equivalent optical constant dispersion, specular reflectivity spectrum and total hemisphere reflectivity spectrum of gilded film, plating aluminous adiabatic material and silicic compound for rough surfaces are measured. The methods of specular reflective intensity are utilized to calculate the root mean square of the gilded film and adiabatic material surface roughness. Based on measured data of LBRDF, the poly-parameter genetic optimized algorithms are utilized to create statistical BRDF model of above-mentioned material. According to the ellipsometry measuring the optical constants, the Powell optimum method is used to deduce these data with certain wavelength for adiabatic material in different waveband.

This article illustrates the operation principle of IEEE 1394 Serial Bus and design architecture of IEEE 1394 Link Layer Controller, and gives a logic design of IEEE 1394 Link Layer Controller, function simulations in software and FPGA-based circuit verification. The experimental result has shown that the controller can transfer data at 100 Mbit/s and accord with IEEE 1394 specification.

The concept of parallel-computing based MPI and the implementation of parallel-computing cluster are introduced. The parallel way of image processing arithmetic is presented. Some methods of impoving parallel efficiency are put forward. A matching arithmetic based on intensity-based correlation and the FFT arithmetic are implemented and run on the parallel-computing cluster. The results are analyzed and compared. The expermental data show the relationship of computing efficiency and the quantities of communication. The feasibility scope of parallel computing in image processing is put forward.

The design of a portable IR image processing system using PC-104 plus embedded PC, uncooled IRFPA and FPGA is introduced. In addition, the two main technique problems, two-points NUC algorithm implemented by hardware and ensurement of image frame integrality, are discussed . Furthermore, some application systems, such as IR image detecting system and IR image tracing system, can be realized based on this IR image processing system.

Template updating method is the key of the success of the matching and tracking algorithm. In order to improve the performance of the tracking algorithm using template matching, a new method for tracking rigid objects in image sequences is proposed after analyzing many template updating methods. A Kalman filter is used to update the template dot by dot instead of renewing it as a whole, and make it adapt to changes in object orientation, illumination and occlusion. The new method improves the performance of the image matching and tracking. The experiments testify the algorithm′s efficiency.

SPIHT algorithm is an embedded zerotree coding algorithm known for its high performance. However, it needs a large number of memories and many repeated calculations, so it is more complicated, and can′t implement real-time compression. The improved algorithm modifies the SPIHT algorithm by use of two concepts-minimum threshold and minimum exported bit, changing the original scanning order. Experimental results show that it reduces both the memory requirement and the time consumption, at the same time the PSNR value and human′s visual effect of the reconstructed image are close to SPIHT algorithm.

The reason of information losing in image histogram equalization is analyzed and a new histogram equalization algorithm based on adaptive local gray level modification is proposed. The algorithm can enhance the image contrast and hold the detail simultaneously. Furthermore, the proposed approach is simple and easy to perform. Experimental results show that it is effective.

Feature selection is one of important research areas of data mining. In real world, many systems to be investigated are very complicated, some of them have problems of noise and information-absent. Through several feature selection methods——evaluation of samples classification and evaluation of each element of feature set, some features with great influence to the performance of the complex information-absent system can be found. Then a practical and correct direction to improve this system is pointed out. The improvement has been proved by the experimental results.

An image coding scheme based on multiwavelet transform and vector quantization is proposed. Firstly, image is decomposed by multiwavelet transform, and then the high frequency coefficients are coded by use of the codebook made by the improved LBG algorithm. The scheme fully uses the comparability of different resolution sub-images in the multiwavelet domain. Coefficients in the highest resolution sub-image are indexed with codebook, and coefficients in the lower resolution sub-image use these index information directly under some certain organization. Comparing to the traditional coding method based on single wavelet, the scheme can improve the quality of reconstruction image and reduce the bite rate at the same time.

With the rapid development of computer networks and the further progress of multimedia technologies,copyright protection of digital media has been one of the most important issues in information technology field. An approach for transparent image watermark based on wavelet multiresolution decomposition is proposed in which a binary image is embedded in the original image. The method uses the characterastic of Human Visual System(HVS) to adjust the intensity of the watermark adaptively. A flag matrix is set to record information about embedding watermark and be used for detecting the watermark. Experimental results show that the watermark is invisible and the algorithm is robust to common image processing.

A modified multiscale fusion method combining context based on weight values for image segmentation is introduced. After the implement of the raw multiscale segmentation based on wavelet-domain hidden Markov tree model, weights for individual dyadic square of the image under consideration at each different scale are set. Finally, raw multiscale segmentation results are fused with the context based on the corresponding weight values. Experimental results show that the method can gain a better performance than the existing technique.

A new way of Support Vector Regression (SVR)-a new regression technique based on the structural risk minimization principle is proposed, it has the similar mathematic form as that of support vector classification, and a simplified Sequential Minimal Optimization (SMO) algorithm based on it is given. Compared with the existing algorithms, the simplified one is simpler in mathematic form and much easier to be implemented without loss in space and time efficiency, which is of benefit to the extensive application of SVR.

According to the requirement of photoelectric theodolite which is used for measuring high speed target′s characteristic, a kind of information record system has been carried out. Both of structure and principle of the system have been presented. This system has two characteristics that are high speed, uncompressed digital image data record and synchronous measurement data record. At last, a model has been finished, its total performance shows that this record system can attain high capability and fulfill works of the target′s characteristic measurement.