How dyfcalid/cameracalibration works
Main Components
a) Core Modules (represented as large rectangular blocks):
• Intrinsic Calibration Module
• Extrinsic Calibration Module
• Surround Bird's Eye View Module
• Tools Module
• Camera/Video/Image Input Sources
• Calibration Parameters Storage (.npy files)
• Output Images/Video
Specific Module Details
For each core module, create sub-components showing:
IntrinsicCalibration:
• Input Handler (camera/video/image)
• Calibrator Engine (InCalibrator class)
• Calibration Modes Manager (CalibMode class)
• Parameter Storage (K, D matrices)
ExtrinsicCalibration:
• Image Pair Processor
• Homography Calculator (ExCalibrator class)
• Transformation Engine
• Parameter Storage (H matrices)
SurroundBirdEyeView:
• Multi-Camera Input Handler
• BEV Generator Engine
• Image Blending Module
• Balance Processor (brightness/white balance)
• Parameter Loader (K, D, H matrices)
Tools:
• Data Collection Utility
• Undistortion Processor
• H Matrix Decomposer
• Time Alignment Tool
• Image-to-Video Converter
Connection and Data Flow
• Raw image flow from input sources
• Parameter flow between modules
• Processing pipeline direction
• Output generation paths
Additional Elements
Include:
• Version information
• Required dependencies (OpenCV, NumPy)
• Input/Output format specifications
• Processing mode indicators (online/offline)
• Modular architecture allowing independent use of components
• Clear data flow between calibration stages
• Parameter sharing between modules
• Tool integration points
• API accessibility for external use
Keep the abstraction level focused on the main processing components while maintaining enough detail to show the relationship between calibration stages and the final bird's eye view generation.