Study Of The Methods For Targets Detection And Position Acquisition Based On Radar Video

Maritime targets detection and position acquisition in sea clutter background have an important application value in both military and civilian fields. In this paper, radar video sequences composed of azimuth and range which are acquired in each scan period of radar are used to detect maritime targets. Methods for targets detection, position acquisition and coupling targets segregation in maritime targets detection system are investigated.In this paper, methods for image background modeling are introduced. Moving targets can be detected according to the differences between current frame and modeled background. For dim targets detection in heavy background clutter, methods of track-before-detect (TBD) are further discussed. TBD approaches avoid the loss of Signal to Noise Ratio (SNR) caused by Constant False Alarm Rate (CFAR). A method combined with fading recursive least-squares and mathematical morphology is proposed by analysing the TBD characteristic of the fading recursive least-squares filter. This method can detect target and its track quickly. Furthermore, it is suitable for engineering application.Azimuth and range of radar video will be used simultaneously to realize target position acquisition, which can overcome the target splitting caused by conventional one-dimensional acquisition. Two methods are presented to segment coupling targets, including the chain code based method and the Possibilistic c-means algorithm. Both the two methods can segment coupling targets caused by several factors, such as sea clutter and so on. Thus, targets position acquisition can be completed, which can provide the foundation and condition for targets tracking latter.