The Target Detection Method Based On Tracking Feedback

The main task of ship navigation radar includes measuring target range,bearing,speed,heading and finally providing collision threaten judgment information for the navigator.Target detection and tracking are two key steps in marine radar information processing.The main purpose of this research is to improve the detection performance of marine radar.The core idea is to use information feedback from target tracking procedure to adjust local detection threshold automatically for fluctuating target.The scope of local threshold adjustment is set by the maneuvering evaluation information extracted from the tracking algorithm for moving targets.The main contents of the thesis consist of the following two parts.(1)In order to improve the detection performance in case of target fluctuation,the tracking information is used to adjust local detection threshold.For the target that has been tracked,the magnitude of target radar echo signal may be reduced due to the fluctuation of the signal to noise ratio and cause target missing.In this thesis,when the target does not appears in the association window the feedback information from the target tracking procedure,i.e.,the predicted target position is used to set the centre for the detection window.And then the detection threshold inside the window is reduced and second time detection is carried out inside the window.The target detection probability will increase and false alarm probability will not increase significantly.Since the traditional detection probability and false alarm probability are not enough to describe the performance of the proposed algorithm.This thesis puts forward the global false alarm probability,the overall average false alarm probability and false alarm probability in the window and multi-frame average false alarm probability.The calculation formulas for these parameters are derived.Finally simulation experiments are designed to verify the formulas.Extending the detection window and lowering the local detection threshold will have an impact on the detection performance.Lowering the detection threshold will make the clutter more easily be detected as target(false target)and increase the probability of false alarm.Expanding the detection and tracking window will make other targets tracks or more clutter get into the target detection window and affects the detection of the real target.Therefore,when the two parameters are modified,a limit must be set to reduce the impact to an acceptable extent.In view of this situation,this thesis also puts forward the concept of false target probability in the window and deduces the formula for calculating the probability of false target according to different detection methods.(2)Taking into account the real targets on the sea are generally both fluctuating and maneuvering,on the basis of the fixed size detection window a method for computing the size of detection window and the detection threshold inside the window for fluctuating and maneuvering target is proposed.The size of the detection windows and the detection threshold inside the window are connected with the target mobility in this procedure.The formula for calculating the probability of maneuvering target is derived by using Kalman filter.Finally,the method is applied to the target detection algorithm,and the rationality of the algorithm is verified by simulation experiments.