A New Method Study On Detection Of Small Target In Infrared Image Sequence

With the development of missilery and technique, researchers are paying more and more attention to infrared imaging guidance technique. The technique of small infrared target detection is one of the key technologies in the system of infrared image information. So the detection and recognition of small target in infrared images has been the subject of intense investigation now and in the future. The infrared image includes several aspects: great-area continuous background, small infrared target, noise and so on. There are many problems in the target detection, such as low signal to noise ration (SNR), little available information, complicated background information, and great handling capacity of information etc. There are chiefly two methods: the method based on the single image and the method based on the infrared image sequence. The method based on the single image only utilized some essential features of the image, such as the grey level. So, many detection methods are based upon the infrared image sequence. Particle swarm optimization (PSO) algorithm is a Heuristic stochastic global optimization technique, and also PSO algorithm is an evolutionary computation technique based on swarm intelligence method. The particle swarm algorithm searches out the optimal regions in complex searching space through the interaction of individuals in a population of particles. What makes PSO predominant is that it is easy to calculate, simple to imply and used widely in many fields. In this paper, we introduced the classical particle swarm optimization algorithm and several of its variants and some applications of the algorithms. The article puts forward a new detection method of small infrared moving target based on PSO algorithm. The detection includes two steps: image preprocessing and target detection. The images are preprocessed using two-step separation method to pick out a small quantity of possible moving targets. And then PSO algorithm is used to search out the possible trajectories built with the possible targets. Then the true target and trajectory can be detected. The experiment results show that the algorithm can detect the true moving small target effectively.