VisionGuide is a smart assistive system designed for visually impaired individuals using Raspberry Pi and AI-based computer vision. It detects obstacles, people, vehicles, and important objects in real time and provides voice guidance to help users navigate safely and independently.

Components Used
Raspberry Pi 5 4GB Single Board Computer 2.4GHz 4 Core 4GB RAM Broadcom BCM2712 Arm Cortex-A76	X 1
DFRobot MAX98357 I2S Amplifier Module MAX98357 - Output Class D Audio Amplifier Evaluation Board	X 1
MicroSD Card Adafruit Accessories 16GB Card with NOOBS 3.1 for Raspberry Pi Computers including 4	X 1
Wire Hook-up Wire 788133 RED 100 FT	X 1
Raspberry Pi Camera Module V2 CSI-based camera module used to capture real-time video for object and obstacle detection	X 2
4Ω 3W Mini Speaker Low-power 4-ohm speaker connected to the MAX98357A I2S amplifier module to provide clear audio output for voice instructions and obstacle alerts in the smart vision system	X 1
PD Power Bank High-output USB-C Power Delivery (PD) power bank used to supply stable power to the Raspberry Pi 5. It should support at least 5V/5A (25W–27W) output for smooth operation of the camera, AI processing, and audio modules.	X 1

Description

Video

Step 1: Hardware Setup

First, all hardware components are connected together.
The Raspberry Pi 5 acts as the main processing unit. The Pi Camera Module V2 is connected through the CSI ribbon cable to capture real-time video. The MAX98357A I2S audio amplifier module is connected to the GPIO pins of the Raspberry Pi, and the 4Ω speaker is connected to the amplifier for audio output. A power bank supplies portable power to the system.

Main Connections

Raspberry Pi 5 ↔ Pi Camera Module V2
Raspberry Pi 5 GPIO ↔ MAX98357A Amplifier
MAX98357A ↔ 4Ω Speaker
Power Bank ↔ Raspberry Pi

Step 2: Operating System Installation

Raspberry Pi OS is installed on the microSD card using Raspberry Pi Imager. After booting the Raspberry Pi, Python libraries such as OpenCV, Ultralytics YOLOv8, NumPy, and pyttsx3 are installed for AI processing and voice feedback.

Software Used

Raspberry Pi OS
Python 3
OpenCV
YOLOv8
NumPy
pyttsx3

Step 3: Camera Captures Real-Time Environment

The Pi Camera continuously captures live video frames from the surrounding environment. These frames are sent to the AI model running on the Raspberry Pi for processing.

Step 4: AI-Based Object Detection

The YOLOv8 AI model processes each video frame and detects important objects such as:

People
Vehicles
Chairs
Poles
Trees
Potholes
Obstacles
Animals
Classroom objects

The system also estimates the approximate distance of nearby obstacles for safer navigation.

Step 5: Audio Guidance System

After detecting objects, the system converts the detection results into voice messages. The audio signal is sent through the MAX98357A amplifier module to the 4Ω speaker, providing clear voice guidance to the user.

Example Voice Outputs

“Person detected ahead”
“Obstacle on the left”
“Vehicle approaching”
“Path is clear”

Step 6: Wearable Smart Assistance

The complete system can be mounted on wearable glasses or a head strap. This allows visually impaired users to receive real-time environmental information while walking independently.

Codes

Downloads
	IMG	Download

Credit

PathFinder

Components Used