基于图像形态滤波的边缘检测

Edge Detection Based on Image Morphological Filtering

为很好地检测出被噪声污染的图像的边缘 ,首先采用CB形态滤波对图像进行预处理 ,再用canny法获得较高的边缘检出率 ,最后采用最小费用搜索法进一步提高了边缘质量 .建立在结构元素轮廓的概念之上 ,以集合延展度为特征进行处理的CB形态滤波保证了滤除噪声的同时保持图像细节 ,canny算法对噪声和弱边缘有更强的适应性 ,而最小费用搜索法对因滤波后遗留噪声引起的虚假边缘有很好的抑制能力 .实例表明这种检测方法能够获得较好的图像边缘 .

In order to obtain true edges from noisy image,three steps are introduced. Firstly contour bougie morphological filtering is adopted with leaving a very little noise that is amalgamated with or close to true image edges. Secondly canny method that is less likely to be “fooled' by noise is applied. Finally LCS(least cost searching) method is used to improve edges' quality by removing false edges which are caused by left noise.