利用空间填充曲线和灰度分布估计的声图分割  

Research on sonar image segmentation by spatial filling curve and grayscale distribution estimation

作　　者：杨海东[1] 孙正凯[1] YANG Hai-dong;SUN Zheng-kai(Nanchang Institute of Technology,Nanchang 330044,Jiangxi,China)

机构地区：[1]南昌理工学院

出　　处：《声学技术》2019年第6期691-697,共7页Technical Acoustics

基　　金：江西省教育厅2017年度科学技术基金重点项目(GJJ171041)

摘　　要：针对声图中的目标分割问题,提出了一种利用空间填充曲线和灰度分布估计的声图目标分割方法。该方法首先利用空间填充曲线将声图由二维矩阵变换为一维向量;其次在一维空间下进行滤波、灰度分布估计、阈值计算、分割处理;最后将分割后的一维向量逆变换为二维矩阵,得到目标分割结果。实际的声图处理验证了方法的有效性。Aiming at the problem of target segmentation in sonar images,a method based on space filling curve and grayscale distribution estimation is proposed and named as SFC-GDE Method.In this method,the space filling curve is firstly used to transform sonar image from two-dimension matrix to one-dimension vector,and then,the entire sonar image processing process,including filtering,grayscale distribution estimation,threshold calculation and image segmentation,are carried out in one dimensional space.Finally,the segmented one-dimensional vector is transformed into a two-dimensional matrix by inverse transformation to achieve target segmentation.Actual sonar image has been processed to verify the effectiveness of the proposed method.

关 键 词：声图目标分割 空间填充曲线 灰度分布估计 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]