Obstacle Detection for Assisting Navigation of Visually Impaired People Based on Segmentation Process

Abstract

Blindness is a condition when a person's sense of sight experiences a disturbance. Hence generally, it requires a tool for help him. One of the assistive devices for people with blind disabilities in carrying out their mobility is a stick or known as "The White Cane. In general, the blind stick has a permanent shape or cannot be folded. This stick works by tapping or sliding in all directions around the blind person standing, so it is very possible for the surrounding environment to feel disturbed. Based on these problems, we propose a device that can help the blind in walking or doing mobility based on computer vision analysis without having to disturb the surrounding environment. Connected Component Labeling is used to get a blob from an image. The blob that has been detected, then analyzed using segmentation based on the threshold process. The experimental results show that our proposed method using Min-max threshold is able to detect obstacles with an accuracy rate of 82.89 %.