Visually Impaired Aid using Computer Vision to read the obstacles

Abstract

The world for normal human being is far different than visually impaired, due to either lack of vision or no vision. The difficulties in their daily routines can be minimized with help of technological support which is usually aids that can be used for travelling. Computer vision a field of artificial intelligence provides the assistance for helping impaired.  The assembly of the device is made as handy as of user is using mobile and the architecture of YOLO (You Only Look Once) is used for accurate object detection. The feature detection of YOLO is more appropriate in real time. The object or the obstacle from which user can collide or the pedestrian localization is indicated to user with help of speaker which make system valuable. The compact size, Raspberry Pi 4 B 8 GB is used or processing, which has proved accuracy of 98% in real time scene.

Background:

The structure of designing a system for visually impaired   to classify the aspect of the visual scene which represent the most important features for navigation and object identification (presence of the objects and their position in space). The auditory system, which is capable of combining information by classes of clues, plays a crucial role for the navigation. The thought of Computer vision is to acknowledge and interpret pictures an equivalent approach humans do, distinctive them, classify them, and type them supported their characteristic traits, like size, color, then forth. Images play a big role in human perception. However, in contrast to humans, computers or machines rework a picture into digital kind and perform some method thereon to urge some substantive data out of it.

Objectives:

Objectives identified were

1. Designing the compact size device that works on real time for visually impaired.

 2.Cost effective system for understanding scene.