基于Canny理论的自适应阈值彩色图像边缘检测

Adaptive Threshold of Color Image Edge Detection Based on Canny Theory

传统的Canny边缘检测算法基于灰度图像,不能充分利用彩色图像的全部信息,且阈值需要人为设定,自动化程度不高。论文提出一种新的算法,基于彩色图像多通道融合技术,根据图像梯度直方图信息,对图像进行自适应阈值处理。将Canny灰度边缘检测算子扩展到彩色边缘检测,利用彩色图像各个通道自身的梯度直方图和梯度方差作为局部阈值,有效解决彩色图像各个通道之间的差别。实验结果表明,其能充分利用图像的颜色和梯度信息,提高边缘检测的准确性。

The traditional Canny edge detection algorithm is based on gray image, and all the information of color image cannot be fitly utilized,and the threshold should be set artificially, the degree of automation is not high. A new algorithm is proposed for image adaptive threshold processing, based on multi-channel color image fusion technology, according to the image gradient histogram. The canny edge detection operator is extended to color edge detection, using the histogram of gradient and gradient variance of each channel image of color image and as the local threshold, and this can effectively solve the difference between each channel of. the color image. The experimental results show that, making full use of the color image and gradient information, this algorithm can improve the？accuracy of edge detection,