Research On Target Feature Extraction Method With Side Scan Sonar

Unmanned underwater vehicle (UUV) is an effective vehicle to complete underwater tasks and a kind of tools with autonomous work ability. However the premise of this autonomous work ability is autonomous environment detection and environment perception.However,it is limited to access to the underwater target information under complex marine environment. For this problem,it is a more feasible approach to equip UUV with side scan sonar. With higher resolution and farther work effective distance, side scan sonar is one of the main sensors for underwater target detection currently.This paper presents the method of target feature extraction using side scan sonar,mainly includes three aspects:Firstly, according to the working principle of side scan sonar, it is easy to understand the meaning and application of side scan sonar data in JSF file which is a file format storing side scan sonar sample data. And then combined with side scan sonar specification, this paper designs a suitable data analysis method for JSF file. Through the data analysis method, these data are extracted including the latitude and longitude information, attitude information, side scan sonar work information and side scan sonar sample data.Secondly, build up a side scan sonar coordinate system, UUV body coordinate system and north-east coordinate system, and then complete the conversion between them. It is successfully to draw the original side scan sonar image by the conversion model of side scan sonar sound intensity to gray level. Since the side scan sonar is affected by the seabed noise seriously, it is necessary to design an effective image pre-processing method which will be done in three steps. The fist step is to remove Gaussian noise using the Gaussian filtering method. The second step is to compensate the feature in the image using the method of grey enhancement. The third step is to select an appropriate threshold parameter for image binaryzation.Finally, according to the classification of side scan sonar and characteristics of different features, the paper provides a target feature extraction method which is suitable for side scan sonar image which is called Hough transform. The line feature in the sea trials data was successfully extracted by Hough transform. With the real marine experimental data, it has been validated that the proposed method of target feature extraction using side scan sonar is correct and practical.