一种复杂背景下的目标分割算法  

A Method of Object Segmentation in Complex Environment

作　　者：李建萍[1] 付丽琴[1] 韩焱[1] 

机构地区：[1]中北大学现代无损检测工程技术研究中心

出　　处：《弹箭与制导学报》2010年第4期197-200,共4页Journal of Projectiles,Rockets,Missiles and Guidance

基　　金：国家重点基金(60532080);省自然基金(2007012003)资助

摘　　要：自动目标检测是图像精确制导技术的重要研究内容。针对地面复杂背景条件下目标难以检测的问题，提出一种结合灰度形态学滤波与直方图独立蜂搜索的目标分割方法。首先，对复杂背景下的图像进行形态学滤波和增强，增强目标和背景的灰度对比度；然后利用文中提出的图像直方图独立峰搜索方法确定图像阈值．对图像进行分割。实验表明：文中方法是有效的．对复杂背景下的目标图像可取得较好的分割效果。Automatic target detection is an important part of the research on imaging precision guided technology. Since it is difficult to recognize targets in a complex background, a segmentation algorithm based on gray-scale morphology filter and searching independent hump of the image histogram has been proposed. Firstly , morphology methods had been used to filtering the image noise and enhance the gray scale contrast between the object and the background;and then determine the histogram threshold by searching independent hump of the image histogram. The new approach is proved to be feasible and effective by some comparative experiments.

关 键 词：数学形态学 阈值分割 自适应阈值 

分 类 号：TN911.73[电子电信—通信与信息系统]