The Research On The Method Of Improving Radar Signal Detection Performance

Since the 21st century, with the continuous development of the shipping industry, people pay more attention to the security of maritime traffic. Nowadays, the situation is becoming more complicated, in order to ensure the safety of maritime traffic, the performance of radar signal detection need greater efficiency. The main task of marine radar detection is to find the target from the noise, sea clutter and other interference. Besides, radar target recognition is based on the information collected from the target feature. The main task of radar target recognition technology is to distinguish and identify the target clearly, it’s a kind of radar target detection, which can improve the target detection probability and reduce false alarm rate. Therefore, applying the radar target recognition technology to radar signal detection can improve the detection performance of radar signals.The theory of radar signal detection is analyzed in this thesis, including signal detection of optimal criterion, constant false alarm rate and pulse accumulation detection. Then this article analyzes the theory of target recognition and its development history, the feasibility of using target recognition technology to improve the performance of radar signal detection. Finally, two signal detection methods to improve the detection performance of radar signals are presented.(1) The matching detection algorithm based on the characteristics of radar signal envelope. The rising edge of the target envelope’s high order moment is extracted, and it’s used for frame matching and frame accumulation.(2) The matching detection algorithm based on the contour feature of radar target. According to the knowledge of mathematical morphology, the interested region’s radar signal is processed, including filling the inner and smoothing the edge. Then the contour information is used for frame matching and frame accumulation.In this thesis, the proposed radar target detection algorithm based on feature recognition is completed under the Visual Studio 2010 environment. Experimental results show this method is very effective.The research is under the support of the branch project of Zhicheng bridge station and Badong station of Yichang vessel traffic management system (contract label: 2010 Z942).