Research And Implementation Of Air Plug Identification And Detection Technology Based On Machine Vision

Aviation plugs are the most basic components in the aerospace industry and are widely used in the aerospace industry.At present,aviation plugs are mainly welded by manpower,but this method is inefficient and error-prone.With the development of artificial intelligence technology,the automatic welding technology of aviation plugs has become an important research direction in the field of aerospace welding.The key to automatic welding is aviation plugs.Accurate positioning of the welding cup.This article uses machine vision widely used in the industry to realize the identification and positioning of aviation plugs,and conduct qualified inspections on the aviation plugs after welding.This paper mainly studies two parts of aviation plug recognition and detection based on machine vision.The first part is the identification technology of aviation plugs.First,the collected images are grayscaled,noise filtered,background removed,and image enhancements are preprocessed to remove interference information.,To improve the image quality,and then segment the aviation plug welding cups by threshold segmentation.Finally,according to the segmented welding cup information,the outline information of all the welding cups of the aviation plug is obtained through the contour search,and the coordinates of the aviation plug welding cups are obtained by the gray-scale center of gravity method.So as to realize the identification of aviation plug.The second part is the inspection technology of aviation plugs.It mainly detects the tin lining and welding conditions of aviation plugs to ensure that the tested aviation plugs are qualified and can be put into use.This article is based on the difference in the contour area of the aviation plug before and after tinning.And after the welding cup is finished welding,its contour information disappears and other features are set up algorithms to realize the detection of aviation plugs.Finally,this paper designs a human-computer interaction interface based on Py Qt5,which enables the staff to operate the recognition and detection system and understand the recognition and detection conditions in real time.Experiments show that the maximum error of the aviation plug recognition algorithm designed in this paper does not exceed 1 pixel value,the actual distance is less than 0.04 mm,and the pass rate of tin enamel inspection is 97.6%,and the detection rate of wrong welding and welding deviation is higher than 90%,To meet the design requirements.