图像边缘检测Laplace算子的改进

The Improvement of Laplace Operator in Image Edgedetection

目的改进原有Laplace算子在检测图像边缘时的精度和清晰度,进一步准确地研究图像的基本特征.方法通过将原有的模板从3×3矩阵形式扩展成7×7矩阵形式.除了对特殊方向的边缘进行检测外,增加对非特殊方向边缘的检测.同时根据各点与中心点的距离远近设置合理的权系数,以提高边缘检测的效果.结果通过比较原有算子和改进算子的实验效果图表明,运用改进的Laplace算子进行边缘检测,边缘更加连续、准确和清晰,有效避免了伪边缘的出现.同时检测出一些原有的Laplace算子检测不出来的边缘,弥补了原有的Laplace算子的缺陷,取得了较好的效果.结论对Laplace算子的改进,提高了图像边缘检测的效果,较好的保持边缘的连通性,准确地检测图像的边缘.

This paper is done in order to improve the precision and clarity of old Laplace operaror in edge detection so that the basic characteristics of image can be studied more precisely.The effect of edge detection can be greatly improved by means of extending the old Laplace operator from the 3×3 matrix to 7×7 matrix,adding some non-special detection direction to some special detection and setting the different coefficient according to the distance of each point from the center point.The experimental comparison of the old Laplace operator with the improved operator has shown that the edge becomes much more continuous,more accurate and clearer,and the fake edge can be avoided effectively when the new Laplace operator is used in edge detection.At the same time,it can detect the edge that the old operator cannot detect and remedy the limitation of the old operator and the result is very good. So it can be concluded that by means of improving the Laplace operator,the effect of edge detection can be boosted,the continuity of edge can be well maintained and the edge can be detected precisely.