Smart Eye for Visually Impaired Community is a Computer Vision and Artificial Intelligence incorporated project. It is based on real-time data analysis using the camera module OV2640. What makes it stand out is its Object detection, distance estimation between the Visually Impaired and the object, and Text to Speech conversion to intimate a voice command to the user using Bluetooth transmission.
Objectives
1- Our main objective is to make a device that is able to capture real-time data by means of live streaming.
2- The device should be able to perform object detection on the video stream.
3- Estimate the distance between the Visually impaired and the object in front of him.
4- The device should intimate a voice command to the Visually Impaired using Bluetooth transmission.
5- The device must have minimal hardware so that it is easy to carry, and also cost-efficient.
Socio-Economic Benefit
Increased Independence: People who have low vision or are completely blind Smart Eye for Visually Impaired Community enables them to navigate their surroundings with independence.
Improved Access to Information: The feature that is beyond the traditional assistive technologies text-to-speech conversion allows the visually impaired to receive real-time auditory cues, enabling them to understand their surrounding without just relying on visual perception.
Technological Advancements: The project drives advancement in the field of Computer vision and Artificial Intelligence.
Methodologies
We have successfully developed Smart Eye For the Visually Impaired Community using the ESP32-Cam Microcontroller and Computer Vision Algorithms. The initial step was to research the problems and the requirements of these special people and decide how we can help them. Then we analyzed the existing technologies in this field, so we can come up with a more effective solution that stands out for the Visually Impaired. In the next step, we choose ESP32-Cam Microcontroller as our main component of the project. After that, we selected the programming language Python and performed object detection, distance estimation, and text-to-speech conversion. The last Phase was the testing phase where we tested the final product in real-time so we have an idea about its working and accuracy.
Outcome
The project has yielded the following outcomes:
1- Real-time video live streaming using camera module OV2640.
2- Object detection using the video live stream using the Tensor Flow lite object detection and Single Shot detection model.
3- Distance Estimation using the triangular similarity method for the distance estimation between the visually impaired and the object.
4- using Google text-to-speech voice command is intimated to the visually impaired using Bluetooth transmission.
