Concealed Weapon Detection Research Based On Image Fusion

Concealed weapon detection, mainly through some scientific and technological means to identify and locate the hidden knives, guns, camouflage explosives and other dangerous items. This, through the image fusion techniques to detect hidden weapons is a more effective detection technique.Against the weapons which is hidden under clothing in the detection of problems, this paper is researching fusion methods based on wavelet transform and color transfer of infrared and visible image. As the visible image is a true reflection of the target scene with high resolution and natural color, while the infrared image is the surface temperature distribution of the reflection, can detect hidden weapons under clothing, but which is short of good contrast and resolution rate. Therefore, the infrared and visible image fusion can complement each other, not only to retain the image fusion infrared images to detect hidden weapons and information, but also able to maintain high-resolution visible images.Lifting Wavelet Transform has many advantages, such as simple structure, easy to implement inverse transform, fast speed and saving storage space. Non-sampling lifting wavelet has translation invariance, which is effectively sparse for the edge of the image features. Wavelet structural change to the updated forecast of the first adaptive lifting wavelet, wavelet structure to solve the first prediction error will be generated with the gradual accumulation of scale problems. Fusion of infrared and visible images, the paper researched a wavelet transform based on lifting wavelet transform domain in the low-frequency subband using the weighted average, high-frequency subband energy measure of a regional integration methods, experiments show that the fused image retains the infrared image weapon characteristics, maintaining the visible image on the scene of the clear details.By lifting wavelet transform fusion images can get the Gray-scale image. through Color transferring makes the gray-scale images is close to natural color visible image. Gray-scale image details is showed out by the colors, thus greatly improving the human eye to distinguish details of the image, thereby more conducive to the understanding of the scene. Color fusion is usually used by color space transformation for color transfer. Based on the color transferred by HSV and LAB space conversion, this paper gets the color transmission using the image fusion transformed by lifting wavelet. At the same time, this method was compared to the image expressed by the PCA-RGB color fusion, and the fused images conversion by the IHS color space, experiments is showing that the color transferring images by HSV and LAB space conversion are better, and the Colors are closer to the visible images. the image can not only get the shape of weapons, features, location and other important information also can get other important information of weapons carriers'appearance, clothing color and so on.