Edge Based Obstacle Detection Model Focused on Indoor Floor-Based Obstacles

Abstract

There are several obstacle detection and avoidance technologies developed to assist the visually impaired community in the current market. They are categorised into different types such as non-vision or vision-based technologies. However, a majority of these technologies are not ready for distribution as it may require additional steps due to mass production and distribution. This paper introduces a simple obstacle detection model built for mobile device. Utilising the Canny Edge detection algorithm, the edge information is used to compare whether the user’s walking path is obstructed by obstacles. The proposed model utilises a mobile device’s proximity sensor and accelerometer to adapt itself accordingly in various environment. The model is tested in various scenarios resulted to an average obstacle detection rate of 85%. The model may be used as a supplementary assistance to the white cane, for people with visual impairment.