一种自适应增强图像纹理的边缘检测方法

An adaptive image enhancement algorithm for edge detection

边缘提取是图像预处理的重要组成部分,针对传统方法在图像边缘提取中含有噪声的不足,根据图像边缘在纹理细节上的差异特征,提出基于自适应算子的方法。通过图像梯度起伏,判断图像平坦区域与锐利的边角区域,基于梯度约束增强图像纹理,并联合数值约束抑制噪声。然后采用均值迭代的策略计算阈值,选取边缘点。实验结果表明,文中算法在检测图像边缘时能够有效防止边缘细节的丢失和断裂,并具备较好的抗噪声性能。

Edge extraction is an important part of image preprocessing. In this paper, we discussed an edge de-tection method based on adaptive texture enhancement operator. First, we distinguished the image flat region and sharp corner region through the image gradient fluctuation. Second, we enhanced the image texture based on gra-dient constraint and reduced image noise based on data constraint. Finally, the edge points were selected through an iterative algorithm to calculate the optimal threshold value. Preliminary numerical results show that the pro-posed method can effectively remove image noise and have a better effect of image edge preservation.