Automatic Analysis of Calibration Board Image Orientation for Online Hand-Eye Calibration

Visual robot hand-eye calibration is the most basic step in robots researching, and the accuracy of hand-eye calibration has a great impact on the working process. In order to perform hand-eye calibration accurately, it is necessary to collect a large amount of robot position data and calibration plate image data effectively. Different data will bring different calibration results. Collecting reasonable data will improve the accuracy of calibration results. This paper proposes an analytical method for the orientation of the calibration board, which can achieve the automatic adjustment of the shooting position of the robot arm. The method combines the image ratio analysis with the block image histogram to complete the calibration board orientation analysis. The image ratio analysis first blocks the image, then binarizes the sub-image, and obtains the proportional parameter of the desired pixel in the sub-image. The block image histogram uses the second-order Gaussian function to fit the histogram to obtain the peak, peak position and half-width information, and obtains the orientation information of the calibration board by solving the peak position and the area of the peak. Experiments show, this method can accurately determine the orientation of the calibration board, then improve the data acquisition efficiency of the calibration board and hand-eye calibration accuracy.