Sonar Image Feature Extraction In The SLAM Technology

This paper mainly studies the image feature extraction which is the front end of SLAM algorithm.Establishes the simulation model of SLAM algorithm based on EKF,by studing the theory about SLAM algorithm.Considering the characteristics of side scan sonar image,the article study the issues related to image feature extraction and image segmentationThe first part of the article is the basic theory of the SLAM algorithm.This paper analyzes the research content and system model of the algorithm,and illustrates the feasibility of the algorithm.The paper analyses the two methods of feature extraction,which is the linear characteristics based on Hough transform and the point characteristics based on SIFT algorithm.This paper studies the characteristics of side scan sonar image.and choose the suitable method in this paper.The following,the article further research side scan sonar image feature extraction method.Paper proposes a way of Image segmentation.By studying the current segmentation algorithms,the article determines the method.The article will take shadow as the starting point for the target boundary.In this section,the paper analyzes the sonar image texture feature.The article detailed analysis of point theory of feature extraction which based on SIFT algorithm.This paper successfully extract the SIFT point characteristics of the image that has split the target,on MATLAB platform.The next article stablishes the model of SLAM algorithm based on EKF.To illustrate the advantages of the algorithm,this paper established two models of SLAM algorithm.They are the model of linear programming and the model of purposeful programming.The last,Using the measurement image of the bottom of the sea,the article calculates the path of the underwater vehicle with the SLAM algorithm..