How muhammad-zeerak-khan/automatic-license-plate-recognition-using-yolov8 works
• Video Input Component: Represents the source video (as provided in the README for demo/testing) that will feed into the system.
• Preprocessing & Detection Component (main.py):
• This module is responsible for extracting frames from the video and feeding them into the detection pipeline.
• It utilizes the pre-trained YOLOv8 model, demonstrated by the “license_plate_detector.pt” file (along with the YOLOv8n model for vehicle detection).
• That this component interacts with the external “sort” module, which is used for tracking.
• Data Output Component: The detection results are output (for example, saved as CSV in “test.csv”), which then becomes the input for post-processing.
• Interpolation Module (add_missing_data.py):
• This module processes the CSV data to add missing frames or smooth the detection data.
• Uses the interpolated CSV output to generate visualizations for demonstration.
• Utility Functions (util.py):
• As supporting/helper functions that might be shared among the main modules.
Relationships and Data Flow
• Main detection (main.py) interacts with the YOLO models and external dependencies (e.g., the SORT module, if used) to output initial detection results.
• Next, connect the CSV output to the Interpolation Module (add_missing_data.py), which processes and smooths the detection data.
Architectural Patterns and Design Principles
• The system follows a pipeline pattern: data flows from input (video) → processing (detection) → post-processing (interpolation) → output (visualization).
• Separation of concerns is evident since each major functionality is decoupled into its own script.
• External dependencies, like the pre-trained YOLO model and SORT module, should be clearly tagged as they are integral to the detection pipeline.
Technologies and Frameworks
• Python is the primary technology, with dependencies outlined in requirements.txt.
• YOLOv8 (used for object detection) and a specific license plate detection model (“license_plate_detector.pt”) are central.
• The external “sort” module from Abewley’s repository is required for tracking.