Firstly, the working temperature changing is studied based on ground phased array radar. Secondly, a still balance theory model is proved by heat conduct theory. Finally, the results of some foreign phased array radars are given in four representative conditions, and the relationships of working temperature with heat spoilage and wind speed are presented.

Based on the radiative transfer and Mie scattering theory, the computing method of analyzing the effects of forward scattering of fog on range-gated imaging system is given. Firstly, the empirical relation of water content and visibility for fog is used to estimate fog drop size distribution. Secondly, the Mie scattering phase function is presented. And then the apparent extinction coefficient can be calculated. Finally, the forward scattering, backscattering and signal irradiance on photo-cathode varying with time and the image contrast between object and background are obtained when the effects of forward scattering are considered. The results demonstrate that the irradiance on object, backscattering and signal irradiance on photo-cathode are increased and the image contrast between object and background is decreased. During the gated time, the irradiance of forward scattering on photo-cathode is much less than backward scattering.

The hypersonic vehicle may be heated and stressed by the turbulent flow field around during supersonic flight. That may influence on the performance of the onboard optical sensor system. So the window must be cooling. The selection of the cooling type, the new infrared window material, the selection of the dome material and the design of the cooling system are presented.

The aero-optical effects generated by the high temperature shockwave flow field around the head of infrared guided hypervelocity flight vehicles will seriously influence the detection, identification and tracking. In order to correct this, some photoelectric correction technologies on aero-optical effects are discussed. Main subjects are adaptive optics technology based on wave-front sensing and self-focusing, compact photoelectric systems with high frequency, frame frequency and frame integrating time of image adaptively varying technology and phase diversity technology. Comparative analysis is also presented on the photoelectric correction technologies mentioned above.

The principle and function of optical range finder are described. According to the defect of the low precision of optical range finder, the design of the step servo based on single chip microcomputer and digital processing equipment is given, which is considered as an improved optical range-finder and it is used to carry out digital calculation and precise control, and the measure precision of optical distance measuring equipment is raised.

Visible image-based detection and tracking system has been used for controlling the firing system to fire impending antiship missile. In this paper, the surface target image is segmented by computing the difference of two adjacent images and then by filtering with wavelet transform. Target detection and tracking is fulfilled by computing the target figures′ consistency and by its centroid linear regression. Experimental results show that this method can be used to detect the antiship cruise missile whose gray scale is very close to the environment and also used widely for other kinds of missile or airplane detection.

A new image recognition algorithm—eigenspace-based approach is proposed. Firstly, recognition system obtains the possible object′s images. Then, all images and image set of each object are compressed to obtain low\|dimensional subspace, called the universal eigenspace and object eigenspace separately. Given an unknown input image, the recognition system projects the image into two kinds of eigenspace. The exact position of the projection in the eigenspace and the distribution formula of samples′ projections determine the object′s species and poses in the image. Experiments are conducted using several objects with complex appearance characteristics. The results show that the proposed approach could keep robustness for the variable poses of object, and could restrain illumination variety.

A target extraction method based on sub-block features is presented to detect certain target. The new feature vector is constructed through computing sub-block features. The feature′s threshold is decided based on Ostu principle to complete the object′s coarse detection. The accurate targe location is obtained by edge detection in this object region finally. The experimental result shows that this method is good at extracting targets with some certain shape and size from complex natural scenes under low contrast and varied illumination conditions. It′s low counting cost and easy for hardware realization.

Based on the performance of the Gabor wavelet on image representation, a face recognition method based on Gabor wavelet and Support Vector Machine is proposed. The Gabor wavelet is used to extract the feature and Support Vector Machine classifier is used for classification. The experimental result shows that the method is effective.

The ability of data concealment is studied adequately. It is proposed that the ability of data concealment can be scaled by use of the degree of edge,the even degree of distribute and luminance. Using the idea of frequency modulation which belongs to communication engineering, frequency modulation watermark which is different from traditional watermark technique is brought up. Constituting baseband signal by selecting the intermediate frequency coefficent of image piece with strong data concealed ability, then, embedding and picking up watermark according to the idea of frequency modulation. This technique can give attention to invisibleness and rubustness, and realize blind examination efficiently.

Based on the idea of layered architecture, an embedded operating system for real-time image processing constructed with ADSP14060 is designed and realized, with the characters of small size, portability and scalability. This system includes not only the functions of task management, resource management and intertask communication on chip, but also the function of intertask communication between chips. The experimental result shows that this system runs with high real-time and strong robusticity and obtains satisfactory effects on parallel image processing.

Two low SNR image sequences energy accumulation algorithms in common use are discussed. One is multi-frames cumulating aiming at slow moving targets, the other is morphological dilation-cumulating aiming at fast moving targets. The hardware implementation based on FPGA of these two algorithms are presented. Experimental result shows that the design has simple structure, clear flow, high running frequency, less periphery components, good expansibility, portability, and has been successfully applied in IR information processor.

The noises and suppression technology in IRFPA CMOS read\|out integrated circuit are analyzed. A new method avoiding the shortage of correlated double sampling method is studied, which can cancel noises without adding power consumption and complexity of silicon technique. In principle, KTC noise can be completely cancelled by this method if the values of capacities are properly selected. Furthermore, the 1/f noise can also be suppressed.

Ti(24 nm)/Al(90 nm) film was deposited on unintentional doped GaN and Si-doped GaN to form metal/semiconductor contacts ,which were annealed at 400～800℃ in the ambient gas of N2.The result shows that Schottky barrier height decreases , ideal factor increases and surface condition of contacts degenerates when annealing temperature increases for unintentional doped GaN. The sample at 600℃ annealing forms Ohmic contacts with low specific contact resistance of 3.03×10~(-4) Ω cm~2. However, the sample for Si-doped GaN with carrier concentration of 5.88×10~(18)cm~(-3) can form Ohmic contacts without annealing, specific contact resistance of the sample is 4.03×10~(-4)Ω cm~2.

A system is introduced, which can automatically measure and analyze a cephalogram. This system includes the patient information management module, the landmarking module, the measuring and analysis module, and the inter-active module. Several image processing methods are used in the course of landmarking, such as wavelet multi-scale analysis and Canny filter. A knowledge-based edge-tracing algorithm also is used to detect edges. Landmarks are detected according to the facial anatomy knowledge. The experimental result shows that the system has good performance.

To seek for new measure way of saddle of bearing the weight, and replace the traditional way with handwork, a kind of untouched measure system by use of the technology of image calibration and measure is proposed, the operation principle is introduced, the method of image processing in this system, and the factors which result in measuring error are analyzed. The experimental results meet the requirement and the measure speed is faster, the method is valid and novel, and it provides new idea for other objects′ measure.

Considering the Guassian distribution of the pump beam and the thermal effect in the gain medium, the coupled rate equations of a LD-pumped passively Q-switched Nd:YVO4 laser at 1.06 μm with Cr4+:YAG are given. The theoretical average output power, pulse width, and pulse repetition of the laser are obtained by the numerical solution of the rate equations. The experimental results accord with the numerical solutions of the coupled rate equations.

Based on available data and theoretical models, the computer simulation of laser radar for direct detection is demonstrated. The basis of the theory model of laser radar for direct detection is built on that the maximal detectable distance is decided by certain detection probability, false-alarm rate, signal noise ratio and threshold noise ratio. In this simulation, it is supposed that the detector noise be denoted by Gaussian probability intensity function, the absorption and scattering of atmosphere and aerosol be interpreted by the U S Standard Atmosphere Model. The simulation results show that this simulation software of direct-detection lidar can give reasonable data under various practical conditions.

Segmentation can be facilitated by knowledge of vision task. In this paper, an approach of segmentation is proposed, which is guided by detection of salient features of the target, and makes the inter-classes variance based thresholding be effective. Rules for local segmentation are defined under consideration of the different distribution of the bulk and engine in IR ship images. The approach is applied to IR ship segmentation successfully, and can be adopted in real-time IR ship detection and recognition.

The task of image segmentation is a crucial step in image processing which is an important application of computer vision. The algorithm presented in this article is composed of two steps, which are the morph-based preprocessing and the ameliorated watershed method. In the first step, the foreground is simplified and the background is kept as the same as possible, and then the ameliorated filed-decision watershed method is used to achieve the segmentation of images in the second step. The experimental result shows that the algorithm is effective.

A simple segmentation method is presented based on the rules of the segmentation result and real-time requirement. Firstly, the moving regions are obtained by the intersections of two neighbor difference images, then small regions that are not accurate are removed. At last, the moving regions are scanned in term of abscissa and ordinate. Experimental result show that the algorithm can get satisfactory result with little time.

The accuracy and validity of the traditional image correlation-matching algorithms are degraded by the presence of the gray change, image clutter, object variation and occlusion. The reason of the degradation is that the method of adding the differences of the pixels′ gray value is used according to the corresponding position as degree of mismatching. The regions formed by image segmentation is a higher-level representation than pixels. Based on the technology of adaptive multi-thresholds image segmentation through potential function clustering, a new algorithm named weighted region template correlation (WRTC) is introduced, which constitutes a method to measure the various membership degree according to different regions. The results show that this new algorithm is robust to more image distortions and has its application value in image matching.

The similarity of the element of matrix after DCT is studied by theory with changing of brightness or contrast, and with transformation of mirroring or transposing. On the other hand, by making full use of the characteristic, an adaptive matching approach based on DCT is proposed. With respect to matching criterion and robustness of noise interference, attention is mainly given to a majority of energy, so it can meet the demand when only matching a few elements of matrix on the upper-left corner after transform. Experimental result proves that the method based on DCT is perfect.

Digital photometrology and the other applications need to mosaic serial images to yield panoramas. The algorithm for mosaicing several frames of images with rotation or distortion are discussesed. In the algorithm, two pairs of control points between two frames for rough positioning are used, and the corresponding points with local-adaptive matching are precisely positioned, and then the data are merged gradually. The experimental results show that the algorithm can solve the rotation and distortion problems in image mosaic to a certain extent, and avoid losing details.

A multi-frame blind deconvolution method is presented for the restoration of the astronomical object which has been degraded mainly by the turbulence and the identification of the point spread function. By using the improved generalized expectation-maximization on the base of the penalized maximum-likelihood estimation method, iterative algorithm is obtained to solve the above problems. Experimental results show that the method is effective and robust.

A novel approach for reconstructing texture-mapped, 3D scene models from a single painting or photograph is presented based on inside characteristics of single view. Gained a sparse set of characteristic parameters in three directions by automatic recognition, the vanishing points and vanishing lines are computed. With these projective attributes, the approach can build the model automatically. The key feature of the approach is an algebraic representation which unifies three typical measurements, and the approach keeps users from the problem of the camera′s internal calibration. We can save the data with VRML format, which gives a convenient way for browsing on internet. The approach has been shown to yield high quality results on a large variety of images.

A region and edge segmentation-based highly scalable coding method on color image has been presented. The algorithm is divided into three sequential steps. The first step is that the color image is segmented as homogeneous texture region or independent object. The second step includes DCT transform of edge pixels and wavelet transform of individual region. The third step is to output layered code stream after the wavelet coefficients is organized by MEZW method.

A new technique of the fusion of a high spatial resolution image and a low-resolution, multisprectral image is developed based on wavelet transform. This method merges the lower frequency components of a multispectral image and its high spatial resolution image by means of local deviation rules with weighting average. And then the merged image is reconstructed by an inverse wavelet transform using the fused approximation and details from the high spatial resolution image. Compared with WT method, the new method presented is clearly better in preserving spectral and improving spatial presentation.

According to the characteristic of hyper-spectrum image and the requirement of realization with hardware, a novel compression scheme based on in-frame decorrelation is presented for hyper-spectrum image. The compression scheme includes two aspects: image matching and in-frame decorrelation, the residual image is compressed based on wavelet transform with bit-plane and adaptive arithmetic coding. The output bit-stream is generated according to the rate-distortion criterion and the experimental result is very satisfactory for hyper-spectrum image compression. The image compression algorithm based on bit-plane and arithmetic coding can also be used to other cases, it′s faster than SPIHT and easy to realize with hardware.