Smart Navigation System for the Visually Impaired Using Tensorflow

Electronics and Communication Engineering

The project aids the visually impaired to navigate independently using real time object detection and identification. The proposed system consists of a Raspberry Pi-3 processor enabled with cloud features and loaded with a pre-trained Convolutional Neural Network model(CNN) developed using tensorflow. The processor was connected to a camera and an ultrasonic sensor. The processor was coded in python, a high level programming language, needed to process images in real time. The camera captures an image in real time which was processed by the Raspberry Pi-3 processor. The python code uses the CNN model to identify the obstacle with boxes and category index. The recognized image category was extracted and stored in a text file. The ultrasonic sensor finds the distance at which the obstacle is present. This distance was also appended into the text file. The contents of the text file are converted to voice using the Google Text To Speech module. This system is extremely flexible and can be used in any environment, without any priming.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.