Research On Polymorphic Defect Detection System Of Wire Harness Connector

Harness connector belongs to the core technology industry of electronics and plays an irreplaceable role in the production and operation of electronic and electrical equipment.The production quality of wire harness connectors has also received more and more attention.At present,the detection of wire harness connectors is highly dependent on people,especially the detection of terminal insertion and related sub parts.Most of them rely on manual detection.Because people are prone to fatigue after working for a long time,it is difficult to ensure the stability of product quality.The comprehensive inspection system of wire harness connector based on machine vision,the last quality assurance after the production and assembly of wire harness connector,is the main content of this research.To sum up,this paper mainly studies and realizes the following aspects:Firstly,according to the requirements of the comprehensive inspection of the wire harness,the hardware structure of the comprehensive inspection system of the wire harness is designed,and each component is selected and compared according to the requirements.In order to solve the problem that the metal terminal insertion characteristics of the harness connector are not obvious and cannot be detected normally and effectively,a method of using line laser as the background and structural light to assist feature extraction was proposed,and the scheme was verified,and the system software was designed and verified.Secondly,for the line sequence detection of wire harness detection,the method of bilateral filtering was used to process the detection area,and a relatively ideal image was obtained.According to the requirements of extracting single wire harness,a variety of gray-scale methods were compared,and finally the weighted average method was selected to gray-scale the image.In order to effectively extract the boundary of a single wire,a dynamic binarization method based on sampling values was proposed,and the boundary was extracted by Canny operator.Finally,an adaptive color quantization feature extraction algorithm based on lightness(V)channel was proposed,which effectively solves the color detection problem in the case of reflection and handwriting interference on the wire surface.Thirdly,using the line laser as the structural light,the two different morphological features of the terminal insertion area of the harness connector were projected into the two-dimensional plane features to realize the simultaneous detection of different morphological features.At the same time,the contour tracking extraction algorithm was used to extract the position deformation features of the rubber shell,and the least square fitting method was used to obtain the relative position feature values of the metal terminal and the rubber shell.The feature extraction and calculation were simplified,and the proposed method was verified by the actual collected images.Finally,a variety of harness connectors were used to test the whole system.The results show that within the required detection range,the accuracy of the system detection can reach 100%,and the false alarm rate is less than 20%.