Research on Identification Technology of Aviation Plug Solder Cup

Abstract

Aviation plugs are characterized by safety, reliability, waterproof and dustproof which have been widely used in various industries. Aviation plugs are currently mainly welded manually, a method with low welding efficiency and a high error rate. The identification and positioning of the welding cups of aviation plugs is a key point in the automatic technology of welding aviation plugs. This paper proposes a combined algorithm based on YOLOv5 and image processing algorithms to realize the solder cup recognition and localization of aviation plugs. In order to improve the recognition and positioning accuracy of the solder cup of the aviation plug, this paper carries out improvement and optimization on the basis of the original YOLOv5 model: One is to embed the ECA attention mechanism in the network; Second, the Wise-IOU (WIOU) loss function is used to replace the loss function of the original model. Improved YOLOv5 algorithm combined with image processing algorithms. Using algorithms to solder cup images noise reduction, gray scale binarization, Hough transform contour recognition. Finally, the feature extraction match function is used for template matching. This enables the identification and positioning of solder cups for aviation plugs. Experimental validation on a home-made dataset shows that the improved algorithm has improved accuracy in the identification of solder cups for aviation plugs, which meets the needs of aviation plugs for soldering operations.