Monitoring image-based obstacle information extraction for telerobot system

An obstacle information extraction method for telerobot system based on monitoring image is presented. The information of static and moving obstacles is obtained in 3D environment without any other sensor. The position and size of obstacles are detected with remote monitoring image in real time. The key for the proposed method is selecting effective 3D information from 2D images acquired via different angles. The measuring procedure in XYZ directions is demonstrated in detail, and the relationship between real position in workplace and pixel position in image is also discussed. The information of the obstacles in the manipulator´s coordinates system is obtained with the formulas fitted by Matlab with the calibrating data. Experiments are carried out to verify the proposed method. The results show that the error between measured and calculated values of obstacle information is in a reasonable range, which can be well applied in some collision avoidance (CA) algorithms for telerobot system.