In this paper we present a method to grasp workpiece by the industrial robot arm based on the machine vision. Most existing methods use the kind of camera calibration methods to establish the spatial position relationship of the robot arm and the workpiece in the image. Instead, we propose a new location method, which is based on using the Hermite Interpolation and mapping in a big checkerboard to confirm the relationship of workpiece in the image and the robot arm. Firstly, we acquire the position of workpiece center in the image through the image correction, the image binary processing and weighted vote. Then, we confirm the relationship between workpiece center in the image and the checkerboard coordinate through the Hermite Interpolation and mapping. Lastly, we use the least square method to determine the relationship between the checkerboard coordinate and robot arm coordinate. Experiments are conducted on a 5-DoF robot arm. Comparing with other methods, results show that our proposed method can achieve high precision in the visual location of workpiece.
展开▼
