Target Association And Echo Simulation Of Marine Radar Based On Contour Features

Radar target association is an important part of target tracking.Traditional target association methods are generally based on location information,ignoring other features of the target echo.However,when there are many observation targets and they are close to each other,only using the position information for association may cause mis-tracking.Therefore,it is necessary to extract other features of the target as the auxiliary judgment basis of the location target association method.In addition,the targets in the radar simulator are mostly fixed-shaped figures,and more features are needed to simulate the actual ehanges in shape and size.With the rapid development of the marine radar system,the resolution of the radar is significantly improved,and the target appears as a target image with a certain pixel size on the radar echo image,which provides the possibility to extract more features from the radar echo.In the above background,this paper analyzes the contour features of marine radar target echo images,and the target association algorithm and target echo simulation algorithm based on contour features are proposed.The main contents include:(l)Radar target image extraction.The target iaage is extracted from the acquired radar image by the region growing method,and a method for extracting the same target eontinuous image is proposed.The method can automatically extract the target image in the continuous radar scanning period after selecting the target.(2)Analyze the edge features of the target image.The Radon variation method is used to align different radar target images.The degree to whieh the same target changes over time and the degree of difference in different targets are analyzed.Through the study of echo images of multiple targets,it is found that in a short scan time,the degree of change of the same target is smaller than the difference of different targets,and the change of the same target with time is mainly reflected at the edge,and the difference of different targets is also reflected at the edge.It is explained that features distinguishing different targets can be extracted from the radar target image,and the feature mainly appears as edge features.This result provides a basis for the research of target association and target simulation.(3)Target association based on contour features.Since the features embodying the target differences are mainly distributed on the target edge,a target association algorithm based on prior contour features is proposed.The algorithm uses the DTW distance to measure the similarity of the two contours.Due to the instability of the target,a target association algorithm based on the average contour of multiple prior contours is proposed.The algorithm calculates the contour located in the middle of multiple contours and uses this contour for the target association,which improves the accuracy of the target association.(4)Target echo simulation based on contour features.In this paper,the contour,area and undulation of the target image are modeled separately,and the shape,size and amplitude of the target image are simulated.The simulated target image is eonverted into a target echo waveform in a polar coordinate system.Firstly,based on tthe analysis of the target edge features,the DTW algorithm is used to calculate the target contour correspondence.According to the correspondence relationship,the variation law of the target contour with time is analyzed,and the target simulation algorithm based on the contour feature is proposed to simulate the actual change of the target contour.Aiming at the extreme situation that the target suddenly becomes smaller,the target area echo probability distribution is analyzed.Based on the contour simulation results,the area-based target size correction algorithm is proposed,which makes the target image size change more realistic.Finally,the Swerling model is used to simulate the target amplitude.The simulated target image ean be transformed into an echo waveform in polar coordinates,which can be used as a signal source for teaching demonstrations,algorithm research,and the like.In this paper,300 frames of images were extracted for 20 different targets for analysis experiments.The contour-based radar target association algorithm experiments show that the method of using the contour features for target association has obvious advantages in the case where the two observation target areas and shape complexity are very similar.For the radar target extracted in this paper,the average accuracy of different targets is about 95%,and the more stable the target,the higher the accuracy.Based on the contour and area modeling,using the target simulation algorithm proposed in this paper,only one actual target eontour can be used as a parameter to generate target echo signals of countless scan cycles.For the simulated target image effects,experiments show that the simulated target contour variation,area,and perimeter probability distribution are consistent with the actual situation.