Modelo de apoio à navegação em calçadas para pessoas com deficiência visual

Abstract

There are more than 35.7 million people with visual impairments in Brazil, of these 6.5 million are totally blind or have great difficulty to see, even with lenses. The white cane and the guide dog are the most used resources to support the mobility of those individuals. However, the white cane offers little information and the guide dog is a resource difficult to access. Considering this context, there are several studies related to navigation support in urban environments for individuals with visual impairments, but there are still many gaps to be researched. The main goal of this work is to present a model that, based on computer vision techniques, can help people with visual impairments to walk on sidewalks. This model offers an integrated solution for location and identification of tactile paving, detection of aerial obstacles, detection of obstacles on the ground and location of crosswalks. The environment information is obtained in several ways, such as camera, ultrasonic sensor and satellite images, and the feedback to the user is sonorous. Experimental results are presented and demonstrate the viability of this approach.