基于小波变换的电磁线绕包图像边缘检测方法

Method of edge detection in wrapped magnet wire image based on wavelet transform

基于机器视觉的电磁线绕包质量检测的关键技术是提取绕包图像的边缘,传统的边缘检测算子存在着定位精度不高或对噪声敏感的问题。为满足不断提高的工业要求,提出一种基于小波变换的电磁线绕包图像的边缘检测方法。首先利用小波函数求出图像的梯度模极大值,然后经过自适应阈值和非极大值抑制去除伪边缘点,归一化处理后得到电磁线绕包边缘图像。与Canny算法比较,该方法能够有效去除噪声影响且边缘提取连续准确,具有较好的适应性和鲁棒性。

To extract image edge is the key technology in the quality inspection of wrapped magnet wire based on machine vision,the traditional image edge detectors have the question of low positioning accuracy or very sensitive to noise. To meet the rising requirements of industry,an edge detection algorithm of wrapped magnet wire image based on wavelet transform was proposed. First it computes the gradient modulus maxima using the wavelet function,then removes false edge points through adaptive threshold and the non-maximum inhibition algorithm,the edge detection of electromagnetic wire image would be realized after normalization processing. Compared with the Canny algorithm,this method removes the noise influence more effectively and gets continuous and accurate edge,it has better adaptability and robustness.