Detection And Recognition Based On Service Robot

Object recognition has been one of the hot and difficult research field of machine vision,pattern recognition and so on.lt has a wide application prospect and important theoretical research value.In the filed of intelligent human-computer interaction has achieved remarkable results,and can be used for a key technology in object recognition is the SIFT algorithm,therefore this algorithm has caused extensive concern of the researchers.In this paper, the study on the theory and development in recent years compared to SIFT algorithm, and put theory into practice, found that the number of feature points extracted by the algorithm of registration and the problem of low efficiency, and put forward the method of combined the SIFT algorithm and cascade classifier combination.Firstly using Adaboost method for object detection based on Haar features;secondly, the scale invariant feature transform (SIFT) feature extraction methods of the object, to improve the real-time and accuracy of recognition; finally, through the K-means algorithm and the coordinate transformation respectively to find stable points of the target object,and cnvert the camera coordinate into the image points needed for the point of service robot.Finally, this paper achieve the above algorithm based on object recognition, will send out the coordinates of the points to the chassis module and arm module so as to realize the moving objects and grasping movements. And the accuracy and validity of the method is verified on the robot platform.