Research On Mobile Photogrammetry Methods Based On Industrial Robots

In the manufacturing industry, measurement equipment can find and analyze the problems in the process of production in time, which is very important for quality control and production efficiency. Moving photogrammetry has been widely used in production measurement with advantages of high precision, high expansibility and so on. However, the application scenarios are limited by the shortcomings of traditional technology. First, markers sticking on the surface of the object are needed so the automation and flexibility is low. The tedious work of pasting and clearing for encoding markers may damage the surface of the object. Second, camera orientation depends on encoding markers, which may be invalid in face of occasion of rapid measurement.In view of those disadvantages for the traditional moving photogrammetry technology, a new measurement method is proposed and a measurement system is designed by combining the moving photogrammetry technology and robot. The camera is fixed on the end effector of the robot (i.e. robot gripper) and the markers are generated by the laser projector. In place of manual labor, robot is used to move the camera so the degree of automation is greatly improved. It can realize non-contact measurement by avoiding the limitations caused by sticking markers artificially. A new method for camera orientation is proposed based on kinematics parameters of robot, which improves the orientation speed. Finally, a matching method combining epipolar constraint and eigenvector is proposed to eliminate the matching ambiguity caused by epipolar constraint, which improves the matching rate and accuracy.The experiment result shows that the measurement error of the system is less than 0.1mm in the range of 0.5mx0.5m, which is of great value for on-line measurement of free curved-surface.