Object Detection and Recognition System for Pick and Place Robot

Object recognition system plays a vital role in controlling the robotic arm for applications such as picking and placing of objects. This paper is directed towards the development of the image processing algorithm which is the main process of pick and place robotic arm control system. In this paper, soft drink can objects such as "Shark", "Burn", "Sprite" and "100 Plus" are recognized. When the user specifies a soft drink can object, the system tries to recognize the object automatically. In the system, the target object region and the motion of the object are firstly detected using Template Matching (Normalized Cross Correlation) based on YCbCr color space. The detected image is segmented into five parts horizontally to extract color features. In feature extraction step, mean color and Hue values are extracted from each segmented image. And then, Adaptive Neural Fuzzy Inference System (ANFIS) is employed to recognize the target object based on the color features. After recognizing the user specified object, the robotic arm pick and place it in the target region. Experimental results show that the proposed method is efficiently able to identify and recognize soft drink can objects.