Research For Target Detection Technology Of Navigation Radar Based On SVM

With the development of national economy,the position of ocean is becoming more and more important.Because of its high resolution,long detection distance and full-time operation,X-band navigation radar is often used in real-time detection of ships,ice floes and other objects in the field of marine environment detection,which provides a guarantee for navigational safety of ships.Traditional sea target detection theory is often based on CFAR method of single sea clutter amplitude distribution model.However,it is difficult to match the actual sea situation with single sea clutter amplitude distribution model,and CFAR method of "point-by-point detection" is too large to meet the requirements of real-time detection.Therefore,how to improve the detection efficiency and performance of navigation radar target detection algorithm according to the characteristics of navigation radar echo is still a problem to be solved in current research.Based on the national defense research project "Shipborne X-band Navigation Radar Wave Retrieval Technology",this paper firstly filters the co-frequency interference which affects the performance of target detection in radar image based on the measured data.Then,aiming at the problem of excessive computation of traditional CFAR method,radar target detection is divided into two parts: pre-detection and fine detection.In pre-detection,SVM is used to filter the radar echoes without targets quickly by classifying the radial radar echoes,thus reducing the computational burden.In the process of using SVM classification,the selection of radial radar echo features directly affects the classification effect.Most of the features described in the current research are based on synthetic aperture radar(SAR).Due to the different radar systems,the direct use of these features in X-band navigation radar will inevitably result in performance loss.Aiming at the characteristics of the radial echo of X-band navigation radar under the influence of near-far effect,this paper proposes a feature extraction method based on fitting correlation coefficient.Through a large number of experimental data,it is proved that the proposed feature extraction method can achieve better classification effect in SVM classification in X-band navigation radar.In the fine detection of radar targets,the characteristics of sea clutter amplitude distribution in each region of the measured data are analyzed.The parameters of the commonly used sea clutter amplitude distribution model are estimated by using the maximum likelihood estimation method and the mean square error test method.The model that best matches the characteristics of sea clutter amplitude distribution in each region is obtained.Then a new sea clutter amplitude distribution model is proposed.Detection method of clutter background partition.After deducing the detector thresholds under different distributions,the optimal parameter selection methods of various types of detectors are obtained by experiment with measured data.Finally,the experimental results show that the detection performance of the proposed method is better than that of the target detection method based on the single sea clutter amplitude distribution model,because the local optimal sea clutter amplitude distribution model is used in each region.Finally,the whole detection algorithm and the traditional CFAR algorithm are used to test the radar data under different sea conditions.The results show that,compared with the traditional algorithm,the proposed method can not only greatly improve the efficiency of target detection,but also reduce part of false alarm when a small amount of target edges are lost.