Detecting retracted pins in electrical connectors based on three-dimensional vision

The reliability of electrical connectors directly influences the success of aerospace launches, and retracted pin detection is a critical inspection step for quality assurance. This study presents an improved method for detecting retracted pins using binocular vision technology. First, based on the analysis of the stereo-matching process, a feature pixel extraction method utilizing the grassfire algorithm and a feature pixel matching method employing global perspective transformation is proposed. Second, after obtaining the three-dimensional coordinates of the feature points, the Random Sample Consensus algorithm is employed for retracted pin detection. Experimental results demonstrate that the proposed algorithm achieves a precision rate and recall rate exceeding 90% for detecting electrical connectors with retracted pins; especially, with a recall rate of 80%, the precision rate reaches 96%, and with a precision rate of 74.4%, the recall rate reaches 96.7%.