基于模糊LQR的汽车线束压接缺陷视觉检测法探微

Visual Detection Method of Automobile Wiring Harness Crimping Defects Based on Fuzzy LQR

针对目前汽车线束压接缺陷视觉检测方法存在的检测精度较差、速度较慢问题,基于模糊LQR研究了一种新的汽车线束压接缺陷视觉检测方法。通过LQR技术对汽车线束数据进行预处理,计算相机镜头焦距数值,清洗无效数据,完成线束对比,获取有效数据。对非铜线区域数值进行过滤,划分数值结果,分割数值区域,选取两个区域的共同结果作为实际需要的区域结果,实现汽车线束压接缺陷定位与特征提取,根据提取的线束部分提取区域数值,判断线束端子的区域,对比线束端子的上孔洞和下孔洞大小,识别线束压接缺陷。建立线束压接缺陷检测函数,综合分析多个不同线束区域的多个不同缺陷转化数值以及它们之间的相互联系,利用模糊处理技术LQR数据提高参数整定效果,完成汽车线束压接缺陷视觉检测。实验结果表明,基于模糊LQR的汽车线束压接缺陷视觉检测方法能够有效提高检测精度,确保检测时间在500 ms以内,检测精度为±0.05 mm,能够很好满足汽车运行和生产要求。

Aiming at the problems of poor detection accuracy and slow detection speed in the current visual detection methods of automobile wiring harness crimping defects, a new visual detection method for automotive wiring harness crimping defects was studied based on fuzzy LQR. The LQR technology is used to preprocess the car wiring harness data, calculate the focal length value of the camera lens, clean the invalid data, complete the wiring harness comparison, and obtain valid data. Filter the value of the non-copper wire area, divide the numerical result, divide the numerical area, select the common result of the two areas as the actual required area result, realize the location and feature extraction of the crimping defect of the automobile wiring harness, and extract the area according to the extracted wiring harness part. Value, judge the area of the wire harness terminal, compare the size of the upper hole and the lower hole of the wire harness terminal, and identify the wire harness crimping defect. Establish a wire harness crimp defect detection function, comprehensively analyze multiple different defect conversion values in different wire harness areas and their interrelationships, use the fuzzy processing technology LQR data to improve the parameter setting effect, and complete the visual inspection of automotive wire harness crimp defects. The experimental results show that the visual detection method of automobile wiring harness crimp defects based on fuzzy LQR can effectively improve the detection accuracy, ensure that the detection time is within 500ms, and the detection accuracy is ±0.05mm. The method designed in this paper can well meet the requirements of automobile operation and production.