基于机器视觉的航空电连接器插针检测技术研究

发布时间：2019-01-11 13:55

【摘要】：航空电连接器在各类仪器仪表和航天系统工程中都有广泛的应用,其在系统中起到信号传送和电能传输的作用,因此航空电连接器的可靠性和安全性对整个系统工程至关重要。目前国内对插针接触件的检测主要依赖人工目视检测的方式,这种方法检测精度低,效率低,检测结果不具有重复性和可追溯性,因此本文提出基于机器视觉的插针检测方法,进行系统设计,实现插针的快速高精度检测。根据测量的要求和技术指标进行了航空电连接器插针检测技术的研究,确定了测量方案,基于机器视觉的检测方法进行了系统的各功能模块设计和选型,完成了计算机软件的编写。对图像中的插针识别算法进行了研究。首先结合被测的电连接器尺寸特征进行图像的背景分割,保留图像中有效信息区域;接着结合图像质量进行图像去噪算法的研究和选择;最后进行插针识别算法的研究,确定了使用基于灰度特征的模板匹配算法进行插针识别的方案,选用相关系数匹配法来衡量插针模板图像与待搜索图像的相似程度,最后进行插针像素坐标的粗定位。对图像中的插针定位算法进行了研究。首先进行系统的像素当量标定,确定了基于最小二乘法的刻线尺标定方法,进行了系统标定实验;接着研究插针像素坐标的提取方法,分别对灰度质心定位法和加权灰度质心定位法进行实验;最后对插针的坐标系转换算法进行讨论,并进行相关实验。对系统的标准不确定度合成进行了讨论研究。分析了系统的误差来源,用统计方法和非统计方法计算系统中各不确定度分量,最终合成系统的总体标准不确定度,并通过测量已知距离的插针来检测系统的测量精度。本文设计的航空电连接器插针检测系统可以满足圆形连接器插针的检测需求,系统的检测精度优于0.06mm,可以满足工业现场的测量需求。
[Abstract]:Avionics connectors are widely used in all kinds of instruments and spaceflight system engineering. They play the role of signal transmission and electric energy transmission in the system. Therefore, the reliability and safety of avionics connectors are very important to the whole system engineering. At present, the detection of pin contacts mainly depends on manual visual detection. This method has low precision, low efficiency and no repeatability and traceability. Therefore, this paper proposes a method of pin detection based on machine vision. Carry on the system design, realize the fast high precision inspection of the pin. According to the requirements and technical specifications of the measurement, the paper studies the detection technology of the aviation electrical connector pin, determines the measurement scheme, and designs and selects the functional modules of the system based on the testing method of machine vision. The computer software is written. The algorithm of pin recognition in image is studied. Firstly, the image background is segmented with the measured electrical connector size feature to preserve the effective information area in the image, and then the image denoising algorithm is studied and selected in combination with the image quality. Finally, the algorithm of pin recognition is studied, and the scheme of pin recognition using template matching algorithm based on gray level feature is determined. The correlation coefficient matching method is used to measure the similarity between the pin template image and the image to be searched. Finally, the coarse positioning of the pin pixel coordinates is carried out. The algorithm of pin location in image is studied. Firstly, the pixel equivalent calibration of the system is carried out, and the calibration method based on the least square method is determined, and the system calibration experiment is carried out. Secondly, the extraction method of pixel coordinates is studied, and the gray centroid localization method and weighted gray centroid positioning method are experimented respectively. Finally, the coordinate transformation algorithm of the pin is discussed, and related experiments are carried out. The standard uncertainty synthesis of the system is discussed. The error source of the system is analyzed. The uncertainty components of the system are calculated by statistical method and non-statistical method, and the overall standard uncertainty of the system is synthesized, and the measuring accuracy of the system is detected by measuring the pin of known distance. The test system designed in this paper can meet the inspection requirement of circular connector pin, and the detection precision of the system is better than 0.06mm, which can meet the requirement of industrial field measurement.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V242;TP391.41

【参考文献】