Development of a Visual Aid Tool for Blind People Based on Faster R-CNN

Abstract

The technology used to help blind people has been developed using various technologies such as an ultrasonic sensor in the form of sonar system in the form of belts and watches. The sonar vision works based on the principle of reflecting a wave with a frequency of 20,000 Hz so that it can be used to determine the distance of an object in a range of up to 3 meters. However, the sonar system cannot identify the object. Computer Vision emerges as one of the most promising solutions. This article focused on developing blind visual aid using a Faster R-CNN identify an object. The Faster R-CNN is instilled on the raspberry pi equipped with a camera. The output is an audio signal about the object and its position refer to the users. The results of this study are able to recognize objects automatically well and accurately.