Angle detection method for surface mount components based on weighted clustering

Abstract

To ensure that electronic components are accurately placed on PCBs, the precision and robustness of the SMT machine's vision system are critical. In light of this, we focus on the challenges associated with detecting the position and orientation of rectangular components, proposing an innovative solution aimed at enhancing detection accuracy and stability. Firstly, to address the impact of varying illumination on edge recognition, we introduce an entropy-based adaptive thresholding method. This approach maintains stable performance under different lighting conditions, ensuring accurate extraction of component edges. Secondly, to tackle the challenge of extracting edge segments against complex backgrounds, we propose a clustering algorithm based on weighted fusion, combined with RANSAC linear fitting techniques. This effectively extracts complete edge segments of the components. Finally, by analyzing these edge segments, we achieve high-precision measurement of the deflection angles of rectangular components. To validate the effectiveness of the proposed method, we conducted experimental tests on eight different types of chip packages and compared the results with those from two state-of-the-art line segment detection algorithms currently in use. The experimental results demonstrate that our method exhibits significant advantages in terms of both precision and robustness. Overall, the proposed scheme not only improves detection accuracy but also enhances the reliability of the system in practical applications, showcasing its potential value in the field of industrial automation. Through comprehensive evaluations of various package types, this study confirms the superior performance of the proposed method in handling complex environments. It provides new insights and technical support for the further optimization of vision systems in SMT machines.