SISTEM PERINGATAN DINI ANTISIPASI KEJAHATAN DENGAN POLA GERAKAN TANGAN

Abstract

This research aimed to create an early warning system to anticipate crime with the patterns of hand motion. This system consisted of Webcamp, Personal Computer, GSM Modem, GSM shield, Arduino uno, and alarm speaker. The webcamp detected input of a hand gesture that had been determined, if the pattern was recognized then the modem will automatically sent a short message to Shield GSM installed in the nearest police station and Arduino Uno would turned on the speaker as an alarm. The algorithms used to identify finger and hand gesture were image segmentation, finger and hand pose determination, and translation of finger position. Based on the trial results, the distance from the camera had an effect on the level accuracy of object detection. The maximum distance from the camera to the object was 2 m (meters).