ROBOTIC ARM CONTROL BY USING COMPUTER VISION ALGORITHMS WITH CONVOLUTIONAL NEURAL NETWORK

Abstract

In this paper, we present a robotic arm control system by recognizing hand gestures from the operator. The system is based on three main steps: locating hand gestures on received images from webcam, determining the contours of hand gestures and recognizing these gestures using artificial neural networks and Deep Learning technology. The use of area ripping and contour detection reduces the computational weight, thereby speeding up the hand gesture recognition process, enabling the robotic arm to perform real-time operations. Experimental results show the positive effect of the proposed method.