Research on fast detecting system based on image recognition

摘要

This paper presents one designing and developing method for the fast detecting system of the piece goods surface quality based on image recognition. In this system, operations are handled in equidistant sampling via the fast-camera on the moving piece goods firstly, and in necessary preprocessing and characteristic pick-up on the collected image. Then from matching the image to the defined module, the system can get the MatchByte (that is also the detected flat amount) to determine the final quality rank of the piece goods; furthermore, the embedded research has been doing to those algorithms (including Sequence Similarity Detection Algorithm, Adaptive Immune Genetic Algorithm, etc.) in the aspect of the detail applications during the developing course.