Research and application of spacecraft equipment electrical connectors appearance defects detection system based on machine vision

The detection of appearance defects in electrical connectors of spacecraft equipment primarily relies on manual visual inspection with low efficiency and poor accuracy and consistency. This paper proposes a machine vision-based appearance defect detection technology for electrical connectors and develops an automated and rapid detection system. Through a monocular camera, high-resolution overall images of various surfaces of the spacecraft single-unit products are captured and image recognition is conducted. By utilizing color space conversion, ROI extraction, and coordinate transformation techniques, the positions of each electrical connector on the spacecraft equipment are obtained. A binocular camera then moves above each electrical connector to capture its stereo images, followed by ROI extraction and pin positioning. Based on the coordinates of the pin tips, 3D reconstruction and point set registration are performed, enabling rapid quantitative detection of defects such as misaligned pins, retracted pins, and missing pins in electrical connectors. Experimental validation shows that the detection uncertainty of this system is better than 0.1 mm. This research provides a reference for the appearance defect detection of electrical connectors.