Files
avm/docs/SAMPLE_DATA.md
T
2025-06-29 16:59:41 +08:00

116 lines
3.7 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# Sample Data Description
## Input Images Overview
The AVM system comes with sample fisheye camera images that demonstrate the complete processing pipeline.
### Image Specifications
- **Resolution**: 1280×960 pixels
- **Format**: PNG
- **Camera Type**: Fisheye lens with wide field of view
- **Calibration Pattern**: 2×4 rectangular grid visible in each image
### Camera Positions
#### Front Camera (`assets/images/front.png`)
- Position: Front bumper center
- Field of view: Covers front area and left/right sides
- Calibration board: Positioned on ground in front of vehicle
#### Back Camera (`assets/images/back.png`)
- Position: Rear bumper center
- Field of view: Covers rear area and left/right sides
- Calibration board: Positioned on ground behind vehicle
#### Left Camera (`assets/images/left.png`)
- Position: Left side mirror or door
- Field of view: Covers left side area
- Calibration board: Positioned on ground to the left of vehicle
#### Right Camera (`assets/images/right.png`)
- Position: Right side mirror or door
- Field of view: Covers right side area
- Calibration board: Positioned on ground to the right of vehicle
### Vehicle Model
#### Vehicle Overlay (`assets/images/su7.png`)
- **Type**: Top-down vehicle silhouette
- **Format**: PNG with alpha channel for transparency
- **Usage**: Overlaid on final stitched result to show vehicle position
- **Positioning**: Automatically centered in panoramic view
### Mask Images
The system uses blending masks to create smooth transitions between camera views:
#### Mask Specifications
- **Location**: `assets/masks/`
- **Files**:
- `maskFront.jpg` - Front camera blending mask
- `maskBack.jpg` - Back camera blending mask
- `maskLeft.jpg` - Left camera blending mask
- `maskRight.jpg` - Right camera blending mask
- **Purpose**: Define blending regions for seamless image stitching
- **Format**: Grayscale images where brightness indicates blending weight
### Calibration Board Details
#### Pattern Configuration
- **Grid Size**: 2 rows × 4 columns
- **Corner Count**: 8 total corners
- **Board Size**: Approximately 50cm × 100cm (physical dimensions)
- **Visibility**: Clearly visible in all four camera views
#### Detection Requirements
- **Contrast**: High contrast between board and ground
- **Lighting**: Even illumination without shadows
- **Position**: Fully within camera field of view
- **Orientation**: Rectangular pattern aligned with vehicle axes
## Using Your Own Images
### Camera Setup Requirements
To use your own images with the AVM system:
1. **Camera Positioning**:
- Install fisheye cameras at four positions around vehicle
- Ensure overlapping fields of view between adjacent cameras
- Maintain consistent height and angle
2. **Calibration Board**:
- Create a 2×4 rectangular calibration pattern
- Position board in overlapping areas between camera views
- Ensure board is flat and clearly visible
3. **Image Capture**:
- Capture images simultaneously from all four cameras
- Maintain consistent lighting conditions
- Ensure calibration board is visible in each image
4. **File Preparation**:
- Save images as PNG or JPG format
- Name files: `front.png`, `back.png`, `left.png`, `right.png`
- Place in `assets/images/` directory
### Camera Parameter Adjustment
If using different cameras, you may need to adjust parameters in the source code:
- **Focal Length**: Modify based on your camera specifications
- **Distortion Coefficients**: Calibrate using OpenCV calibration tools
- **Image Resolution**: Update if using different image sizes
- **Fisheye Scale Factor**: Adjust based on lens characteristics
For detailed parameter tuning, refer to the [Technical Specifications](TECHNICAL_SPECS.md).