Application of Image Recognition Technology in Automatic Teaching of Humanoid Educational Robots

Abstract

In recent years, more and more robots are being applied to school and home widely. Teaching and playback are a universal method to develop robot functions. For purpose of simplifying the process of teaching humanoid educational robots, a method of human motion and gesture recognition, based on digital-image processing technology, is proposed for automatic teaching of robots. In this method, raw images taken by camera are converted to human body gesture binary images by using the technology of image subtraction and image preprocessing. Then, the body gesture images are mapped to body mathematical models by the skeletonization and recognition algorithms. Finally, the control commands of robots are calculated and drive robots to do the similar actions. Compared with other methods of robot teaching, the teaching method by using digital-image processing and recognition technology has an easier usage and lower cost. By series of experiments, it is seen that this method has a good performance on teaching robots.