Visual impairment and Blindness Assistance
It is difficult, nearly impossible, for blind people to walk without an aid .Even with an aid such as a white cane or a guide dog can be sometimes inconvenient, uncomfortable, and perhaps inaccurate in avoiding obstacles.According to World Hea
2025-06-28 16:29:58 - Adil Khan
Visual impairment and Blindness Assistance
Project Area of Specialization Artificial IntelligenceProject SummaryIt is difficult, nearly impossible, for blind people to walk without an aid .Even with an aid such as a white cane or a guide dog can be sometimes inconvenient, uncomfortable, and perhaps inaccurate in avoiding obstacles.According to World Health Organization(WHO) report there are 285 million visually impaired people among which 39 million are totally blind.The estimated number of blind people in Pakistan is 1.14 million.
In order to overcome this problem the idea we came up with is an Integrated Machine Learning System which allows the Blind Victims to identify and classify Real Time Based Common day-to-day Objects and generate voice feedbacks and calculates distance which produces warnings whether he/she is very close or far away from the object. The same system can be used for Obstacle Detection Mechanism.
Project ObjectivesThe project consists of the following objectives:
- To design and implement the AI based project for efficient and assisstant for visually impaired and blind people in the general masses using open-source technologies to save economic resources.
- To achieve object detection and distance measurement using image processing
- To develope an economic and efficient substitutiion for the nedical industry.
- Sensor Selection :As this all project is based on detection so its first part of the design to find a correct sensor that fulfills our requirement and can easily available in the market. For this, our efforts are split into looking at verity possible type of sensors.
Ultrasonic Sensor (HC-SR04) will discharge/emit waves and use the distance formula to calculate the distance of reflection of wave which stricken back from object.
Laser distance sensor GY-530 The long range laser sensors are optoelectronic sensors designed for non-contact measurement of distance, displacement and speed. They feature a large measuring range from 0.5m to 3000m.
- Raspberry Pi Camera Module: The camera will keep capturing pictures in the form of frames in real time and store in the memory of raspberry pi.
- Headset for audio feedback with distanced object
- Microcontroller Selection :There are several microcontrollers available to operate ultrasonic sensor HC-SR04 like Arduino UNO, TI MSP430,Raspberry Pi Model B+, Arduino Mega ETC. Based on our experience, requirement and research Arduino UNO R3 is the best.
-
ARDUINO UNO: This will be used to synchronize the distance calculated from sensor and integrate it with raspberry pi.
-
RASPBERRY PI-4(4GB/8GB): This is the main hub of our project will have data trained in AI, which will process the image taken, identify the objects with classes we trained(4 classes we will have trained due the inefficient processor we have available currently), take delayed input from Ardiuno Uno at certain intervals to calculate the distance and integrate it with the object detected, output the results using the gTTS (google Text To Speech Library in Python) to the headset provided to the individual.
Following are the benefits of our project:
- The main goal of our project is to reach out the blind and assist them in every possible way.
- The project we are working on will be portable, as most of the project requires laptop or Pc to get along with but it our case we are making it feasible for the blind person to go around without any supervision(i.e for us it's not a big thing but they also want to feel independent and self-sufficient as most of us do).
- This technology, which benefits from lightweight design, encompasses a wide range of virtual perceptions.
- Their innovative wearable, empowers the visually impaired with navigational independence through touch-based directions and haptic feedback. This assistive technology allows users to be guided without having to look at their phones and enables the blind to "see" the world.
- Previously projects with similar aim has been made but contains high delay rate. Our project will overcome that flaw and produce the best results through its working.
- Less expensive than the approaches made in past.
The proposed system comprises of the following final deliverable:
- Raspberry Pi 4 (4GB/8GB) – Python, OpenCV for computer vision and Image Processing
- Tensorflow for collection of workflows to develop and train models.
- Raspberry Pi Camera Module for capturing images of objects
- Headset for audio feedback
- Ultrasonic Sensor/Laser Sensor to calculate the distance between the user and the next upcoming object
- Cane(Optional) to which sensor will be attached in order to calculate distance along with a vibrator/buzzer to alert the person.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 73700 | |||
| Raspberry pi-4 | Equipment | 1 | 41500 | 41500 |
| Raspberry Pi Camera Module V2 | Equipment | 1 | 9000 | 9000 |
| Power Adapter 5V | Equipment | 1 | 600 | 600 |
| Arduino UNO R3 | Equipment | 1 | 2500 | 2500 |
| Ultrasonic sensor | Equipment | 1 | 300 | 300 |
| Laser Sensor GY-530 VL53L0X | Equipment | 1 | 800 | 800 |
| Handfree | Equipment | 1 | 2000 | 2000 |
| Structure Material | Equipment | 1 | 3000 | 3000 |
| Wiring and tools | Equipment | 1 | 4000 | 4000 |
| Poster and Printing | Miscellaneous | 1 | 8000 | 8000 |
| Model Decoration | Miscellaneous | 1 | 2000 | 2000 |