A real-time detection method for cable printing quality based on image key point matching

Abstract

The invention relates to the technical field of cable printing detection, and in particular to a real-time detection method for cable printing quality based on image key point matching, wherein a plurality of areas with the largest amount of information are automatically extracted from this part of the video as standard areas through an algorithm; the standard areas are used for matching with subsequent video frames, and areas with more than 2 matched key points are retained; the difference D between the standard areas and the matching areas is calculated, and if D is greater than a specific threshold, it is considered to be a printing abnormality, otherwise it is normal, through key frame extraction, image graying, and key area matching; various types of abnormality detection are realized, and the measures adopted include multi-category feature calculation and multi-category difference measurement; a printing abnormality detection method is adopted to automatically extract and match key areas, thereby avoiding manual input of a to-be-matched area and information interaction with an extrusion device, and realizing low coupling and automatic detection between systems.

Images