一种采用改进Canny算子的机械臂识别与定位算法设计  

Design of robotic arm recognition and positioning algorithm using improved Canny operator

作　　者：罗小青 胡荣[2] LUO Xiaoqing;HU Rong(Electromechanical Laboratory,Jiangxi Institute of Mechanical Science,Nanchang 330000,China;Nanchang University College of Science and Technology,Jiujiang 332020,China)

机构地区：[1]江西省机械科学研究所机电研究室,江西南昌330000 [2]南昌大学科学技术学院,江西九江332020

出　　处：《现代电子技术》2024年第13期27-31,共5页Modern Electronics Technique

基　　金：江西省教育厅科学技术研究项目(GJJ2207505)。

摘　　要：为进一步提高机械臂图像处理的精确度,文中在改进Canny算子的基础上,提出一种适用于机械臂视觉伺服的图像识别与定位算法。该算法利用张正友标定法、灰度化、滤波及阈值分割等方法,完成了精确的相机标定和图像预处理,并在此基础上通过改进传统的Canny算子,降低了图像处理中边缘检测算法的检测误差,从而有效提高了机械臂的目标识别与定位精度。仿真实验结果表明,与传统算法相比,基于改进Canny算子的识别及定位算法具有更高的处理精度和更快的执行速度。An image recognition and positioning algorithm suitable for manipulator visual servo is proposed based on the improved Canny operator to further improve the accuracy of manipulator image processing.In the algorithm,Zhang′s calibration method,grayscale,filtering and threshold segmentation are used to fulfill accurate camera calibration and image preprocessing.On this basis,by improving the traditional Canny operator,the detection error of edge detection algorithm in image processing is reduced,so as to effectively improve the object recognition and positioning accuracy of manipulator.The simulation results show that the recognition and positioning algorithm based on the improved Canny operator has higher processing accuracy and faster execution speed in comparison with the traditional algorithm.

关 键 词：机械臂 视觉伺服 相机标定 灰度化 图像滤波 边缘检测 轮廓提取 CANNY算子 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP274.51[电子电信—信息与通信工程]