Research On The Visual Recognition Technology In The Stud Automatic Sorting System

Sorting is a key link in the process of manufacture studs. The accuracy and speed ofsorting have a direct impact on the production line. Sorting studs is often in two ways: manualsorting and sorting machine. Manual sorting has low efficiency and poor reliability. Sortingmachine is not popularized and does not reach a high intelligent degree. In this paper, in orderto achieve automatic sorting of studs, applying visual recognition technology in studautomatic sorting system, and focusing on developing the technology and algorithms ofrecognizing and localizing the studs based on machine vision. The main contents of the paperinclude:(1) A stud visual recognition method based on feature matching is proposed. Therecognition and location process is divided into two stages: training stage and recognitionstage. In the training stage, a database of features of the stud samples is established. In therecognition stage, determining the type of studs in the image to be recognized by using themethod of feature matching, and calculating the position and orientation information of studs,and it would be used to sort out the studs.(2) Different algorithms are designed to extract the features of stud, fulfill the featurematching, recognize and localize the studs, such as the stud sample region extractionalgorithm based on line scanning, stud image segmentation algorithm and stud edgepositioning algorithm, and so on.(3) A stud recognition and location experimental system based on machine vision isdesigned and implemented to test the methods above. The hardware and software of theexperimental system are developed. The system realizes the function of recognizing andlocalizing the studs.Experimental results show that the proposed stud automatic recognition and localizationmethod is feasible and effective. The experimental system is able to recognize and localize thestuds that are non-occluded and in arbitrary pose in the field of view.