Abstract
For automation using various industrial robots, recognition of object attitude and position is very important. Two-dimensional image processing with the CCD camera has entered use for simple shape objects or objects which are arranged in a plane or regularly. To extend the application area, recognition of three-dimensional shapes is necessary. Therefore, we have developed a bin picking system which recognizes the position and attitude of the individual object from disorderly piled objects, through three-dimensional object recognition technology approaches using the geometric matching method between the measurement result of the range sensor and the 3D-CAD model. This paper describes the above development.
