Implementation of a device adjustable to the white cane for the detection of objects and obstacles in the motion of people with visual disabilities

Abstract

This paper presents the implementation of a device adaptable to the white cane for the assistance of people with visual disabilities in the recognition of objects and obstacles and the distance at which they are. The device emits alerts through audio messages in which it indicates which object is in front and at what distance it is in a range of 1 m to 4 m. Also, it sends vibrating alerts to the person using it. The device uses an image classification algorithm using artificial intelligence techniques such as deep learning. The algorithm detects a large number of objects, this is conditioned and limited through Arduino so th at the device works based on the determined specifications and the needs of the visually impaired. The device detects the objects and obstacles determined, in addition to correctly measuring the distance at which they are. The alerts that it issues when communicating with the user are adequate and consistent with the operation of the device.