Aerospace Plug Pin Detection Method Based on YOLOv8 and Edge Detection

This paper proposes a fault detection method based on YOLOv8 and edge detection for handheld aerospace plugs photographed by operators at the launch site. The method identifies the socket and pin areas in the aerospace plug images using the YOLOv8-based object detection method. The pin images are processed using the local adaptive binarization method based on the Sauvola algorithm to obtain binarized pin images. The slopes of the four sides of the socket are obtained using edge detection algorithms, and the parameters involved in position correction are calculated. Finally, the transformed template pin coordinates are compared with the pin coordinates in the binarized image to determine if there are any faults in the plug pins. This method can effectively handle aerospace plug images with significant shadows and avoids pixel errors caused by transformations of the original image, proving its feasibility in real on-site detection.