Smart Blind-Spot Detection System
by aishwarynarayan in Circuits > Microcontrollers
15 Views, 0 Favorites, 0 Comments
Smart Blind-Spot Detection System
The Smart Blind-Spot Detection System is an AI-powered safety system designed to detect objects or vehicles present in the blind spots of a vehicle.
Using Quarky Intellio and PictoBlox, the system uses a camera to monitor the left and right sides of a vehicle. When an object is detected in a blind spot, the system provides a visual and audio warning to the driver.
The project can be demonstrated using a robotic car model and toy vehicles to simulate real-world traffic conditions.
Supplies
Pictoblox
Quarky Intellio
Description
The Smart Blind-Spot Detection System is an AI-powered safety system designed to detect objects or vehicles present in the blind spots of a vehicle.
Using Quarky Intellio and PictoBlox, the system uses a camera to monitor the left and right sides of a vehicle. When an object is detected in a blind spot, the system provides a visual and audio warning to the driver.
The project can be demonstrated using a robotic car model and toy vehicles to simulate real-world traffic conditions.
How the Project Works
The project works using the following process:
Intellio Camera
↓
Captures the Dummy Vehicle Area
↓
PictoBlox AI Model
↓
Identifies the Object Position
↓
Left / Right / Both / Safe
↓
Displays the Warning on the Computer
The entire system can be demonstrated on a table using a cardboard vehicle model.
The Dummy Vehicle Model
I am using Intellio Rover Kit
Creating the Blind-Spot Zones
Create two areas around the dummy vehicle.
Left Blind Spot
Mark an area on the left side of the vehicle.
Label it:
LEFT BLIND SPOT
Right Blind Spot
Mark an area on the right side.
Label it:
RIGHT BLIND SPOT
Setting Up Quarky Intellio
Place the Intellio camera in front of the dummy vehicle model.
Adjust its position so that the camera can see:
- The dummy vehicle
- Left blind-spot area
- Right blind-spot area
- Objects placed around the vehicle
The Intellio should remain stationary during the demonstration.
Creating the AI Model
Open PictoBlox and set up the AI/ML model.
Create four classes:
Class 1 – Safe
No object is present in either blind spot.
Class 2 – Object Left
An object is placed in the left blind spot.
Class 3 – Object Right
An object is placed in the right blind spot.
Class 4 – Objects Both Sides
Objects are placed in both blind spots.
Collecting Training Data
For each class, capture multiple images using the Intellio camera.
Safe
Keep both blind spots empty and capture several images.
Object Left
Place a small toy vehicle in the left blind spot and capture several images.
Object Right
Move the toy vehicle to the right blind spot and capture several images.
Objects Both Sides
Place toy vehicles on both sides and capture several images.
Try to capture the objects from slightly different positions so that the AI model learns different situations.
Training the AI Model
After collecting the images:
- Check the images in each class.
- Make sure the classes contain different examples.
- Train the machine-learning model.
- Test the model using the live Intellio camera.
- Check whether PictoBlox correctly identifies each situation.
Testing the System
Once the AI model is trained, start the live camera.
Test 1 – Safe
Keep both blind spots empty.
Expected result:
SAFE – NO OBJECT DETECTED
Test 2 – Left Blind Spot
Place a toy vehicle in the left blind spot.
Expected result:
⚠ OBJECT DETECTED – LEFT
Test 3 – Right Blind Spot
Move the toy vehicle to the right blind spot.
Expected result:
⚠ OBJECT DETECTED – RIGHT
Test 4 – Both Blind Spots
Place objects on both sides.
Expected result:
⚠ WARNING – OBJECTS ON BOTH SIDES
PictoBlox Program Logic
The program continuously reads the prediction from the AI model.
When Green Flag Clicked
↓
Start Intellio Camera
↓
Start AI Model
↓
Forever
↓
Read AI Prediction
↓
If "Safe"
↓
Display "SAFE"
↓
If "Object Left"
↓
Display "⚠ OBJECT LEFT"
↓
If "Object Right"
↓
Display "⚠ OBJECT RIGHT"
↓
If "Objects Both Sides"
↓
Display "⚠ BOTH SIDES"
Making the Demonstration More Interactive
To make the project more impressive, create a driver dashboard in PictoBlox.
The dashboard can display:
┌───────────────────────────────┐
│ BLIND SPOT MONITOR │
│ │
│ VEHICLE STATUS │
│ │
│ LEFT : SAFE │
│ RIGHT : WARNING ⚠ │
│ │
│ OBJECT DETECTED │
│ RIGHT SIDE │
└───────────────────────────────┘
This makes the project look like an actual vehicle safety interface.