Evaluation of the effectiveness of computer vision methods based on machine learning and classical image processing for automatic control tasks

This article presents a comparative analysis of two fundamentally different automated optical inspection methodologies in the context of electronics manufacturing processes: a classical algorithmic approach based on the OpenCV library and a machine learning-based approach using the Edge Impulse platform. The research focuses on the task of inspecting the installation of electronic components on printed circuit boards. The purpose of the work is to conduct an objective assessment of the effectiveness of each of the methods according to key criteria for industrial implementation, such as: detection accuracy, verification speed, as well as the complexity of development, configuration, and requirements for source data. Based on the analysis, practical recommendations are formed for choosing the optimal approach, depending on the specific conditions of the production process and the tasks set.