Research On Infrared Small Target Detection Algorithm

Due to advantages of concealment, the strong ability of anti-interference, the recognition of camouflaged targets and passive detection, infrared detection has been used in the area of aerospace technology military guidance satellite infrared weather map and traffic regulation, especially in military field. With the popularization of the Infra-Red Monitoring and the application of infrared thermography technology, the infrared small target detection technology has been a hot issue in recent decades.An infrared small target usually locates in the complicated and variational background, whose gray composition usually presents big fluctuation and possesses most area. Under that circumstance, the target point is vulnerable to the background to the extent that it will be submerged in the background. That makes the detection of the small target point very difficult. Large amounts of scholars both at home and abroad spend a lot of energy in infrared small target detection technology and propose a lot of effective target detection and tracking algorithms. However, although the infrared target image detection algorithms proposed at the present stage are effective, they are constrained at the same time. This paper proposes three effective modified infrared small target detection and tracking algorithms based on research of predecessors. The main work and innovations are as follows:(1) An algorithm to extract the infrared small target template was proposed, which creates the small target template by using Low/High Pass Filter and Background subtraction technology. And the small target template is used to filter the infrared image to restrain the background of the image and detect targets. Compared with mean-shift algorithm and the fixed weight algorithm, this method performs better at improving image quality, removing noise and background, and improving the signal to noise ratio (SNR)and the signal to noise gain.(2) An artificial bee colony algorithm was proposed to optimize the threshold segmentation function. By means of extracting the best threshold parameter optimization and segmenting target image, the target information is enhanced and the noise and background is also removed. It is effective to restrain infrared image background and improve the SNR. (3) An infrared small target detection algorithm based on LMWIE preprocessing and the ABC adaptive threshold segmentation algorithm was proposed, where the preprocessing with the image information entropy algorithm of the background which combined with ABC adaptive threshold segmentation algorithm, can perform preferably in target detection. Compared with OTSU and Iterative method, it is more effective when it comes to the infrared small target image detection with a Low SNR.