Abstract:Microholes, as important structures in fields such as aerospace, precision molds, medical devices, and electronic manufacturing, their taper parameters directly affect assembly accuracy, fluid transmission performance, and product lifespan. Traditional microhole taper detection mainly relies on equipment such as optical microscopes and coordinate measuring machines, which not only have low detection efficiency but also have certain subjective errors in manual interpretation, making it difficult to meet the online quality inspection requirements in the context of intelligent manufacturing. To address these issues, a microhole taper automatic assessment method integrating computer vision and machine learning is proposed. Firstly, an industrial vision system is used to obtain high-resolution images of microholes, and image preprocessing, edge enhancement, and contour extraction are employed to automatically acquire the geometric information of microholes. Secondly, a multi-scale geometric feature and texture feature fusion model is constructed to extract key features such as hole diameter, edge gradient, contour inclination angle, and gray distribution. Finally, a machine learning algorithm is utilized to establish a microhole taper prediction model, enabling intelligent assessment of taper parameters.