基于去噪的熵算子边界检测法

图像的熵用来刻划图像的平滑性或均匀性 在图像的局部窗口中计算时 ,如果窗口中存在边界 ,则窗口中的图像不均匀 ,其灰度变化急剧 ,计算出的熵小 ;反之熵大 设定熵的阈值 ,即可判断是否存在边界 由于熵算子对噪声很敏感 ,直接用它进行边界检测 ,效果很差 文中针对这一缺陷 ,提出将熵算子与去噪相结合的边界检测法 ,如果计算出的熵大于阈值 ,要判断是噪声的出现所引起 ,还是边界的出现所引起 ,这样 ,边检测边界边去噪声 用该方法对信噪比较低的图像进行实验 ,得到了满意的效果

An Edge Detection Method of Entropy Operator Based on Noise Removal

Entropy of an image is to express the smoothness or homogeneity of the image.While computing in a local window,if there exist edges,the local image will not be homogenous,variation of the grayness will be sharp and the entropy obtained will be low.Otherwise,the entropy will be high.Given the threshold of entropy,it can be determined whether or not exist edges.Because entropy operator is sensitive to noise,the effect is bad if it is directly used to detect edges.In view of the deficiency,the paper comes up with an edge detection method in which entropy operator is combined with noise removal.If the entropy computed is higher than the threshold,it will be necessary to determine whether it is caused by noise or by edges.Thus edge detection and noise removal can be made at the same time.With this method satisfactory effect has been achieved by experimenting upon image with low ratio of signal to noise.