包装物料形状特征提取和识别方法

目的为了提高包装过程物料抓取成功率,采用机器视觉设计一种物料识别和定位方法。方法以串联机械手臂为载体搭建一种基于机器视觉的物料识别、定位、抓取平台,包括物料传送模块、图像采集模块、视觉分拣模块、机器人控制模块和抓取模块等。重点论述相关图像处理算法,包括基于双边滤波的图像预处理方法,基于Canny算子的图像边缘检测,图像特征提取和质心定位等。最后进行实验研究。结果实验结果表明,码垛机器人的物料形状正确识别率可以达到99.25%,抓取成功率能够达到99.5%。结论所述物料形状识别和抓取定位方法可有效解决图像特征提取、定位等问题,具有识别率高、抓取准确等特点,能够满足包装搬运要求。

Method of Shape Feature Extraction and Recognition of Packaging Materials

The work aims to design a method of material recognition and positioning based on machine vision, in order to improve the success rate of material grasping in the packaging process. A material recognition, positioning and grasping platform based on machine vision was built with the series manipulator arm as a carrier, including material transfer module, image acquisition module, visual sorting module, robot control module and grasping module. The related image processing algorithms were emphasized, including image prepossessing method based on bilateral filtering, Canny operator based image edge detection, image feature extraction and centroid location. Finally, an experimental study was carried out. The experimental results showed that, the correct recognition rate of material shape of the palletizing robot could reach 99.25% and the success rate of grasping could reach 99.5%. The method of shape recognition and grasping and positioning of materials can effectively solve the problems of image feature extraction and positioning, and has the characteristics of high recognition rate and accurate grasping, which can meet the requirements of packaging and handling.