针对含有复杂地物、边缘密度大的遥感影像,提出了一种基于多结构元素形态学的边缘检测方法。首先,根据遥感影像的特点,用形态学的开闭滤波器去除影像中的噪声点。然后基于多结构元素形态学进行边缘检测,最后,对边缘检测结果进行二值化及细化处理。并同Sobel及Canny算子的检测结果进行了比较。结果证明,该方法能有效地提取边缘、保护细节。

Based on the character of remote sensing image, which has complex terrain and objects and abundant edge, a new morphology edge detection algorithm is proposed. Firstly, according to the characteristic of the remote sensing image, we remove the noise by using open-closing morphology filter; then use morphology edge detection algorithm to detect the edge; finally, we change the results into binary image and thinning it. The algorithm, Sobel algorithm and Canny algorithm are applied in remote sensing image, and their results are compared. The conclusion proved the new morphology edge detection algorithm can well remove the noise under the condition of preserving details of the remote sensing image.