Intelligent Identification of Remnant Ridge Edges in Region West of Yongxing Island,South China Sea  

作　　者：WANG Weiwei GUO Jing CAI Guanqiang WANG Dawei 

机构地区：[1]College of Information and Control Engineering,China University of Petroleum,Qingdao 266580,China [2]Guangzhou Marine Geological Survey,Guangzhou 510760,China [3]Laboratory of Marine Geophysics&Geo-Resources,Institute of Deep-Sea Science and Engineering,Chinese Academy of Sciences,Sanya 572000,China

出　　处：《Journal of Ocean University of China》2018年第1期118-128,共11页中国海洋大学学报（英文版）

基　　金：supported by the National Natural Science Foundation of China (Nos. 41576049, 41666002);the Key Research Projects of Frontier Science of the Chinese Academy of Sciences (No. QYZDB-SSW-SYS025);Geological projects of China Geological Survey (Nos. GZH 201400210, DD20160140);the Natural Science Foundation of Hainan (No. ZDYF2016215);the Key Science and Technology Foundation of Sanya (Nos. 2017PT13, 2017 PT14)

摘　　要：Edge detection enables identification of geomorphologic unit boundaries and thus assists with geomorphical mapping. In this paper, an intelligent edge identification method is proposed and image processing techniques are applied to multi-beam bathymetry data. To accomplish this, a color image is generated by the bathymetry, and a weighted method is used to convert the color image to a gray image. As the quality of the image has a significant influence on edge detection, different filter methods are applied to the gray image for de-noising. The peak signal-to-noise ratio and mean square error are calculated to evaluate which filter method is most appropriate for depth image filtering and the edge is subsequently detected using an image binarization method. Traditional image binarization methods cannot manage the complicated uneven seafloor, and therefore a binarization method is proposed that is based on the difference between image pixel values; the appropriate threshold for image binarization is estimated according to the probability distribution of pixel value differences between two adjacent pixels in horizontal and vertical directions, respectively. Finally, an eight-neighborhood frame is adopted to thin the binary image, connect the intermittent edge, and implement contour extraction. Experimental results show that the method described here can recognize the main boundaries of geomorphologic units. In addition, the proposed automatic edge identification method avoids use of subjective judgment, and reduces time and labor costs.Edge detection enables identification of geomorphologic unit boundaries and thus assists with geomorphical mapping. In this paper, an intelligent edge identification method is proposed and image processing techniques are applied to multi-beam bathymetry data. To accomplish this, a color image is generated by the bathymetry, and a weighted method is used to convert the color image to a gray image. As the quality of the image has a significant influence on edge detection, different filter methods are applied to the gray image for de-noising. The peak signal-to-noise ratio and mean square error are calculated to evaluate which filter method is most appropriate for depth image filtering and the edge is subsequently detected using an image binarization method. Traditional image binarization methods cannot manage the complicated uneven seafloor, and therefore a binarization method is proposed that is based on the difference between image pixel values; the appropriate threshold for image binarization is estimated according to the probability distribution of pixel value differences between two adjacent pixels in horizontal and vertical directions, respectively. Finally, an eight-neighborhood frame is adopted to thin the binary image, connect the intermittent edge, and implement contour extraction. Experimental results show that the method described here can recognize the main boundaries of geomorphologic units. In addition, the proposed automatic edge identification method avoids use of subjective judgment, and reduces time and labor costs.

关 键 词：GEOMORPHOLOGY mapping edge detection image processing 

分 类 号：P7[天文地球—海洋科学]