基于机器视觉的布匹色差在线检测系统

On-line detection system of cloth color difference based on machine vision

针对目前人工检测劳动强度大、准确率低,在线检测图像尺寸大、处理速度慢等问题提出布匹色差在线检测算法。将采集到的布匹图像进行边缘处理,去除图像的黑边;利用图像金字塔技术降低图像尺寸;结合色差计算公式和直方图交叉法计算布匹的左、中、右色差波动图和合格率。为使用方便,设计基于机器视觉的布匹色差在线检测系统,利用布匹色差在线检测算法替代人工检测系统。经过试验结果对比,证实布匹色差在线检测系统与人工检测系统在色差检测过程中具有一致性,并且在线检测系统的验布速度由0.2 m/s提高至0.8 m/s。

Aiming at the problems of high labor intensity,low accuracy,large size of online detection images,and slow processing speed,an on-line detection algorithm for cloth color difference was proposed.The algorithm processed the edge of the collected cloth image to remove the black edge of the image.Image pyramid technology was used to reduce image size.Combining the calculation formula of color difference and histogram crossing method,the left,middle and right color difference fluctuation chart and qualified rate of cloth were calculated.In order to use conveniently,an on-line detection system of fabric color difference based on machine vision was designed,which used the algorithm of on-line detection of fabric color difference to replace the manual detection system.The comparison of the test results proves that the cloth color difference on-line detection system and the manual detection system are consistent in the color difference detection process,and the cloth inspection speed increases from 0.2 m/s to 0.8 m/s.