| In view of the current problem of automatic pearl sorting and low level of intelligence on the market,based on the research of pearl shape sorting system in the early stage,the pearl color intelligent sorting system based on visual servo is further studied.The main contents include the design and implementation of the basic components,the algorithm implementation of the pearl color classification,and the robot target recognition.Firstly,in order to reduce the reflection of the pearl surface and ensure the consistency of the pearl color,this paper builds a sealed image acquisition environment,and designs a reflective light illumination system.For the lack of single control of the pearl feeding mechanism,the outer groove of the eye is designed.The wheel-type pearl single control feeding mechanism;at the same time,the air-sealed sealed pearl multi-surface image acquisition device and the three-dimensional classification container capable of realizing pearl size classification are improved.Secondly,in the classification of pearl color,this paper studies the corresponding image preprocessing method according to the characteristics of the image,and uses the area of area deletion and threshold segmentation to achieve accurate extraction of pearl target.At the same time,in the L*a*b color space,The three-channel color average is designed by the GA-SVM classifier for pearl color classification.Finally,this paper applies the industrial robot vision servo system to the pearl sorting industry,and deeply studies the camera calibration,coordinate system conversion and human eye calibration,etc.,which realizes the robot’s position recognition of the sorting target.According to the national standard,the color of the pearl is divided into five color systems.Through the comparison test of various pearl color recognition methods,the accuracy of pearl color classification is improved from 74.21% to 98.42%,and the high-speed movement of the arm is used to achieve high pearl efficiency.Grab the sorting. |
PDF Full Text Request