Feature Detection Method of Assembly Position Based on Machine Vision

The detection of assembly station scene features is a hot research issue in the field of computer vision. Its purpose is to superimpose assembly-related information on the corresponding assembly station scene and provide intuitive online guidance to assembly operators. To this end, this paper proposes a method based on machine vision to detect assembly station scene features, and uses dimensionality reduction processing to simplify the detected high-dimensional features, and uses a bag of words model to construct a Code Book representation method of assembly station scene feature vectors. The dimensionality reduction method is used to solve the problem of the large amount of calculations directly using natural features to match the scene.