Moving Small And Dim Target Detection Under Complex Background

Detect and track moving small and dim target under complex background is a key technology in Infrared Surveillance Systems (IRSS). A large number of background and the specialty of the dim small target-far away from the target,low SNR,strong clutter lead to the problem hard to deal with. Now, there are several methods to detect the small target, for example Wavelet Analysis,training of the neural network can get content result. But the calculated amount and real-time require is hard to achieve. They are the key problems in this field. So, the research of detecting small and dim target under complex background have upper applied value.This paper analyzes some methods which are used to detect the target,about the target under complex background or with low SNR.We divide the problem into two sides:one is the method to suppress the background,the other is the method to detect the small target in image sequence. This paper analyzes the traditional method to suppress the background.After that, put forward a concept called "Degree of Complex Infrared Radiation Background "and bring the Weighted Information Entropy into this task. We check the validity of the concept,and choose a suitable high-pass filter, make a connection between the concept and the cut-off frequency. The method can change the parameter of the filter according the different scene and obtain a content result.About the detecting method,the paper use neighborhood detecting method to detect the target using the continuity of the orbit. With the contrast to the traditional method,we can find the this proposal have good commonality and content result.The whole paper uses MATLAB to finish the simulation, we do lots of experiments with different algorithm, give their advantages and shortcoming, and some views to improve the algorithm.