Study On Deformation-camouflage Target Discovery Based On Image Processing

Digital image processing technology has been applied in the military field extensive, but the Applied Research in the Deformation-camouflage Target Discovery is poor. Deformation-camouflage has become the most important one of camouflage techniques on the modern battlefield. In this paper, the characteristics of Deformation-camouflage are systematically analyzed and then a method based on image processing is proposed to discovery Deformation-camouflage Target. On this basis, on the color space transformation, homomorphic filtering enhancement and image segmentation such as are systematically studied by experiments.Camouflage can be divided into protection-camouflage, deformation-camouflage and imitation-camouflage. Deformation-camouflage is used most widely. The main characteristics of deformation-camouflage have:①Because the color of camouflage spots must comply with the main background spots in the object area, camouflage spots and background spots are similar in color.②Because the design of deformation-camouflage spots seeks to the size of camouflage spots must comply with the main background spots, camouflage spots and background spots are similar in size.③Because deformation-camouflage often adopts some amorphous spots constituted by irregular contour curve and the disposition of camouflage spots must follow certain principles, the camouflage spots can distort the lines and contours of object.The analysis shows it is difficult to find the deformation-camouflage target quickly and accurately. However, it is possible to rapidly and accurately determine the detect image with image processing techniques.HIS color space transformation can divide the image information into three components: H (Hue), I (Intensity) and S (Saturation), then the components can be treated separately. In HIS color system, S (Saturation) is very important to edge detection of the deformation-camouflage image because the same color not necessarily has the same S (Saturation).As deformation-camouflage makes the camouflage spots spraying on the object similar with background spots, however, because of mixing paint and artificial operation limits the color saturation is not the same in the image. If S(saturation) is treated by Sharpen operator, the difference between object and background will be enhanced and this provides conditions for edge detection and object segmentation. One of the difficulties of Deformation-camouflage Target image segmentation is how to detect and extract integrated edge of camouflage spots and background spots. By enhancing the high-frequency component Homomorphic filtering can enhance the reflection component, so that the edge at the junction of different objects is enhanced to detect and extract easily.The image processing experiments are carried by the Roberts operator, Prewitt operator, Sobel operator, LoG operator and Canny operator. The experiments show that: Prewitt operator is more appropriate for the Deformation-camouflage Target Discovery.Therefore, this paper presents the method that pre-process the detect image by HIS color space transformation and homomorphic filtering, then detect edge with Prewitt operator and segment the target image by mathematical morphology techniques. The experimental results show that: The method is a feasible method to fast determine the detect the Deformation-camouflage Target image and find the Deformation-camouflage Target accurately and timely.