基于颜色聚合向量的线序检测方法

A Line Order Detected Method Based on Color Coherence Vector

摘要:

颜色聚合向量方法被采用于检测各种连接线的线序。首先对采集到的线束图像进行中值滤波，消除部分噪声；然后截取检测线束的检测区域图像并将其二值化；接下来分割各根导线，得到每根导线的轮廓，再根据各轮廓位置获取原图像中对应各导线的图像块，并计算各根导线图像块的颜色聚合向量（color coherence vector，CCV）特征；最后利用欧氏距离将各导线的归一化的CCV颜色特征和模板各导线的颜色特征进行比较。实验结果表明，该方法在线序检测中能获得较高的准确率（97.8%）和速度（15fps），满足实时检测要求。

Abstract:

The color coherence vector method is used to detect the line order of various connection lines. Firstly, the collected image of the wire harness is filtered by median filter to eliminate the noise. Then, the image of the detection area of the wire harness is cropped and binarized. After that, each wire is segmented to get their contours. According to the position of each wire, the corresponding image blocks of each wire in the original image are extracted, and whose color coherence vector (CCV) features are calculated. Finally, the normalized CCV features of each detected wire and the features of each wire in the template are compared by using Euclidean distance. The experimental results show that the proposed method can achieve high accuracy (97.8%) and speed (15fps) in the line order detection, and meets the real-time detection requirements.