摘要
为了解决传统Canny方法中采用统一阈值进行边缘检测的弱点,提出了改进的边缘提取方法。将卷积模板扩大为3×3,得到各像素的灰度梯度,然后根据3×3邻域内的平均灰度梯度代替原梯度,结合影像中灰度梯度的最大值,得出了新的高阈值计算方法,即每个像素的高阈值都按照各自的灰度特征进行计算,最终完成改进的边缘检测。实验结果表明,改进后的方法即使在阈值不恰当的情况下仍能获取较好的边缘提取结果。改进后的方法可以为摄影测量内业处理提供参考。
In order to solve the weakness of the traditional canny algorithm in edge detection using uniform threshold,this paper proposes an improved edge detection method.First,the convolution template is expanded to 3×3,and the gray gradient of each pixel is obtained.Then,the gradient is replaced by the average gradient according to the gradient in the 3×3neighborhood.Combined with maximum image intensity gradient,we obtained a new high threshold calculation method.The high threshold of each pixel is calculated in accordance with its respective gray features.Finally,the improved edge detection is finished.Experimental results show that the proposed method can not only improve the accuracy of edge location,but also obtain better edge extraction results in the case of inappropriate threshold.The improved method can provide a reference for internal processing of photogrammetry.
引文
[1]何晓宁,莫毅君,童俊涛.快鸟影像用于泥石流冲毁区自动提取[J].遥感信息,2015,30(5):77-82.
[2]秦家鑫,万幼川,何培培,等.一种建筑物点云轮廓线的自动提取方法[J].遥感信息,2015,30(4):3-8.
[3]石立坚,施英妮,王其茂,等.水平集方法在卫星遥感图像海冰边缘检测中的应用[J].遥感信息,2015,30(3):89-93.
[4]雷丽珍.数字图像边缘检测方法的探讨[J].测绘通报,2006,12(6):56-60.
[5]李武.Matlab中图像边缘检测算法研究[J].北京测绘,2014,116(3):65-68.
[6]刘洪鸣.图像边缘检测技术的研究与比较[J].科技信息,2010(1):875-876.
[7]李小红.基于LOG滤波器的图像边缘检测算法的研究[J].计算机应用与软件,2005,22(5):107-108.
[8]CANNY J.A computational approach to edge detection[J].Pattern Analysis and Machine Intelligence,1986,8(6):679-698.
[9]李俊山,马颖,赵方舟,等.改进的Canny图像边缘检测算法[J].光子学报,2011,40(1):55-54.
[10]陈宏希.基于边缘保持平滑滤波的canny算子边缘检测[J].兰州交通大学学报,2006,25(1):86-90.
[11]王蓉,高立群,柴玉华,等.综合Canny法与小波变换的边缘检测方法[J].东北大学学报,2005,26(12):1131-1133.