一种基于边缘检测及纹理分析的水坝图像分割算法

Dam Image Segmentation Based on Edge Detection and Texture Analysis

针对边缘检测算法的局限性及灰度水坝图像的特点,提出了一种基于边缘检测及纹理分析相结合的灰度图像分割算法,首先利用边缘检测算法对待处理图像进行边缘检测得到图像的粗分割,然后在原灰度图像中对得到的边缘位置点进行纹理分析,去除检测到的非目标对象的边缘从而得到分割图像,即细分割。将该算法应用到河坝监测系统中,实验证明该算法达到了很好的分割效果。

Image segmentation is the foundation of the image processing,pattern recognition and computer vision etc.Among the algorithms,the image segmentation based on edge detection has already been the topic of image segmentation.Edge detection can provides the edge points of the region segmentation fast and correctly,but the complexity of the surrounding in which the object being processed lies gives a lot of 'false edges'.In this paper,image segmentation based on edge detection and texture analysis is proposed.First,it processes the gray image by edge detection,called rough segmentation,and then it analyzes the texture of the edge points in the original gray image,called fine segmentation.It applies our algorithm to the system of dam surveillance and get fine results.