Automated Human Identification and Obstacle Avoidance for Visually Impaired
This paper provides a method for human identification and obstacle avoidance for visually impaired. Visually impaired people faces lots of difficulty in accomplishing their day to day activities. Among one such difficulty is to recognize the person, this paper comes up with a technique which will helps blind people to identify person approaching them. Here DWT technique is used for face recognition. In this technique the entire image is decomposed into discrete wavelet bands. From this bands required features of image is obtained. This features when subjected to classifier gives proper output by identifying the person. Another part of paper deals with obstacle avoidance by using a blind stick. Blind stick uses an ultrasonic sensor and PIR sensor that detect obstacle at a distance of 100 cm. This stick can be used as alert signal for blind people.