Gray Statistical Characteristics Of Sonar Image And Its Application In OTSU And FCM Segmentation

In sonar technology,the accuracy of sonar image segmentation results directly affects the subsequent processing steps and the final understanding of sonar images,so the sonar image segmentation plays an important role in sonar technology.However,because the sonar image is affected by the seafloor noise,it has the problems of small resolution,blurred boundary,and being susceptible to noise interference.Therefore,there is no reliable and feasible method for sonar image segmentation.This paper mainly studies the grayscale statistical features of sonar images,and uses the OTSU algorithm and FCM clustering algorithm to achieve sonar image segmentation.This paper first summarizes the domestic and foreign research status of sonar image grayscale statistical features,and gives the calculation formulas and physical meanings of partial grayscale statistical features of sonar images.Secondly,the existing OTSU image segmentation algorithm is studied.For the traditional OTSU algorithm,the mean value of grayscale is used to measure the parameters.This paper proposes an improved one-dimensional OTSU image segmentation algorithm.Its main idea is to introduce the mean square error that can reflect the degree of image grayscale dispersion.In the algorithm,the new inter class variance is constructed with the mean square error of gray and the mean value of gray.For the construction of two-dimensional histogram in two-dimensional OTSU algorithm,this paper uses the grey value of gray statistical feature of sonar image,average value of neighborhood gray value,median value of neighborhood gray value and maximumvalue of neighborhood gray value to define different two-dimensional histograms to achieve two-dimensional OTSU dual-threshold sonar image segmentation.Finally,for the classical FCM clustering image segmentation algorithm,neither the uneven intensity of the image intensity nor the pixel spatial information problem is taken into account.In this paper,the gray statistical features of sonar images are introduced into the FCM clustering algorithm,which is based on the three-dimensional vector representation of the gray and spatial information of sonar images,and is different from the other two-dimensional vector representation based on the image gray and spatial information.