一种新的彩色图像边缘检测算法

传统的边缘检测方法大都基于灰度图像,不能充分利用彩色图像的全部信息。针对已有算法中存在的像素点扩散、边缘定位不准确、边缘不连续等问题,提出了一种彩色图像边缘提取算法,基于图像自身梯度方向信息和多通道信息融合技术,将灰度边缘模板算子扩展应用到彩色图像的边缘检测中,在RGB空间中对原彩色图像进行多通道边缘检测;同时采用滤波来抑制噪声,依靠边缘生长保证检出边缘的连续性,并提出了自适应确定边缘提取门限值的方法。该文提出的彩色图像边缘检测算法计算量小,实验结果表明了其能充分利用图像的颜色和梯度信息,有效地消除噪声,提高边缘检测的准确性,保证边缘的连续性。

Color image edge detection algorithm

Traditional edge detection methods are mostly based on gray images,so they cannot efficiently process messages in color image.A color edge detection method was developed to correctly reproduce distinct,continuous edges based on the multichannel data fusion technique and the gradient direction information contained in the image.The algorithm employs a multipass processing approach that removes noise from the RGB components with a pre-filter,with the color edge detector then based on an extended gray scale edge gradient template detector.Edge growing is used to remove edge discontinuities with an adaptive method to select the edge extraction threshold.Simulation results show that the algorithm reproduces accurate, smooth edges due to the use of the gradient and color information.