SENSOR VISUAL UNTUK DETEKSI POSISI JARI DAN TANGAN

Application of human hand gesture recognition technology has been widely used in various fields, one of them is robotics field. In this research a visual sensor system is designed to detect human fingers and hand positions. Finger and hand position are the main parameter that measured in this system. Fingertips and wrist hand are marked with different color to simplify the process to detect finger and hand position. Fingertips markers in the image are segmented with thresholding method. Then, centers of gravity of them are identified to form a finger hand feature. The length of each finger vector and angle between them during the open hand gesture become a reference for determining value of vector finger length ratio and ratio of angle between finger vectors. Those values become the main variable that used to represent the current position of fingers movement. System testing was done by using some data that have been varied based on lighting levels, distance between hand and camera, and shape of hand as well. From the tests result, effect to system performance by lighting level variations is greater than the distance variations, with distance between hand and camera not more than 1 m. The system success rate in detecting finger and hand position is 94.5%, while the system success rate in representing finger and hand position is 86.74%. Keywords: telerobotic, robotic hand, visual sensor, image processing, finger and hand detection, thresholdin