说明：   边缘是图像中最重要的信息，是模式识别、图像分 割以及场景分析的重要基础。边缘检测是图像处理领 域一项重要的技术和任务。图像中边缘通常解释为灰 度值有突然改变，像素梯度具有局部极大值。很多经 典的边缘检测算子如:Roberts, Sobel, Prewitt, Frei- Chen和Laplacian等，均以图像局部区域特征与特定 边缘模式匹配为基础，定义数个方向导数模板(一般为 窗口)，将图像与模板进行卷积运算实现边缘检测。它 们具有实现简单，速度快速，易于应用的特点。当模板 方向与边缘方向相匹配时，能取得非常有效的边缘检 测结果。
(The edge is the most important information in the image, and is an important basis for pattern recognition, image segmentation and scene analysis. Edge detection is an important field of image processing technologies and tasks. Edge in the image is usually interpreted as grayscale values ​ ​ sudden changes in pixel gradient local maxima. Many classic edge detection operators such as: Roberts, Sobel, Prewitt, Frei-Chen and Laplacian, etc., are the image of local regional characteristics specific edge pattern matching based on the definition of a number of directional derivative template (usually a window), the image convolution operation of edge detection with a template. They have a simple, fast speed, easy to apply. When the template direction and the edge direction matches, can achieve very efficient edge detection result.)