Paper Details: Downloads: 2571
Serial Number: P1151712554
Title: Precise Obstacles Avoidance System for Visually Impaired People using Xbox 360 Kinect
Authors: D. Barrinuevo and G. Jacosalem and J. N. Lagahit and M. J. Lobedica and M. R. Salar and R. Tolentino
Abstract: This study proposes a system for obstacle avoidance for visually impaired people that uses Canny Edge Detector to eliminate the reliance of detection in the segmentation of the floor plane. The system acquires the images through Microsoft Kinect of Xbox 360 which is the primary input device. The process occurs in the depth image that will undergo depth thresholding to limit the distance. The resulting image from depth thresholding will undergo noise reduction to fill the broken areas. Then the edge detection through Canny will be performed. The extracted edges that appear as a contour will be used to determine the presence of an obstacle and will be enclosed in a bounding box drawn to the color image. Hereafter, decision making will be made which is responsible for determining if the user must go forward, left, right or stop through a sound feedback. The system was tested based on the rate of its detection out of 385 samples and on a real-time trial navigation on a structured environment by a blindfolded and blind user.
Keywords: Canny Edge Detector, visually impaired, obstacle detection, Kinect
Journal/Conference: International Journal of Graphics, Vision and Image Processing
Volume: 17
Issue: 1
Submission Date: 3/18/2017 12:00:00 AM
Review Date: 4/20/2017 12:00:00 AM
Publishing Date: 5/6/2017 12:00:00 AM
Article Downloads: 2571
Download:

Facebook