SISTEM PENGENALAN GESTURE TANGAN UNTUK KENDALI GERAK MOBILE ROBOT BERBASIS PENGOLAHAN CITRA

Abstract

Abstract-Implementation of hand gesture recognition system for navigation of mobile robots using Bluetooth communication media apply image processing techniques is presented. The preprocessing image of the camera capture is resized, and color space conversion from RGB (red, green, blue) to HSV (hue, saturation, value), then segmentation process is performed to remove the background and get a hand gesture object. Feature extraction using methods Convexity Hull Defect to get the point finger tips, calculate the moment and the Centre of Gravity (CoG), then look for the angle parameter finger tips and the Euclidean distance, it’s using the process of pattern recognition hand gestures. The test results of the system obtained on hand gesture recognition accuracy rate of 80.4%, or there is an error amounting to 19.6 %. In implementation systems show mobile robot navigation system is good enough to control the distance of more than 15 meters. Keywords: Gesture, control, HSV, Convexity Hull Defect, Euclidean Distance