Research On Infrared Moving Small Target Detecting And Tracking Under Complex Background

Infrared image guided weapons are proved to be very efficient in modern wars because of it’s passive working modern, favorable stability and adaptability. Infrared dim target detecting and tracking method becoming a research hotspot as the core of the infrared guidance technology. There are lots of achievements about Infrared dim target detecting and tracking in recent years. But infrared dim target detecting and tracking in complex background is still a difficult problem because the target imaging of very small area and the effect of background clutter. In this paper,we have conducted the thorough research on infrared dim target detecting and tracking in complex background. The main work is as follows:(1) A new method depends the optimal signal to clutter ratio for dim target detection in single frame. First filtering the image with normalized Laplace of Gaussian (NLOG) filter to gain the contrast of the image and search the local extremum pixels to be the candidate target. Then using multi-scale analysis to calculate the scale of the candidate targets and gain the clutter of the candidate targets. Take the optimal NLOG filter response as the target signal strength of the candidate target and the standard deviation of it’s neighborhood recovery as the target background clutter. Finally, compute the signal to clutter ratio as the objective measure of the target.(2) A novel method for infrared dim target tracking based on Kalman filter and LOG distribution field. We find that target pixels are less than half of the target area when describing the target area with rectangle. Therefore,the target model has been badly influenced by the adjacent background pixels. In order to avoid this bad effect, we use LOG filter to suppress the background gray scale change first and then construct the distribution field for the target. LOG distribution field ensuring the advantage of the traditional distribution field, and can adapt to background changing.(3) Analyzing the imaging process of ground target and figure out image compensation method derived to camera motion when the camera status information and ground height is available. Then analyzing image path after image compensation and extract the length feature of the path to detecting moving target.In summary, this article has conducted thorough research to the infrared moving dim target detection and tracking in complex background, some achievements have been made and has been verified by experiment.