System Architecture
End-to-end processing pipeline for real-time license plate recognition, authentication, and visitor management.
IP Camera Stream
RTSP / HTTP
RTSP / HTTP
Stage 01
Image Acquisition
Continuous frame capture from surveillance hardware.
OpenCV StreamRaw Frame Data
1080p @ 30fps
1080p @ 30fps
Inference Engine
Stage 02
Plate Detection
Localized object detection to identify vehicle plates.
YOLOv8 NanoBounding Box
Confidence Score
Confidence Score
Normalization
Stage 03
ROI Cropping
Extraction and perspective correction of the plate region.
Image ProcessingSegmented Plate
Grayscale Optimized
Grayscale Optimized
Deep Learning OCR
Stage 04
Character Recognition
Translating visual characters into machine-readable text strings.
EasyOCR / PyTorchASCII String
e.g. "ABC-1234"
e.g. "ABC-1234"
Local Storage
Stage 05
Authentication
SQLite 3
Whitelist Query
Match Verification
Match Verification
Outcome A
Dashboard Update
Successful authentication triggers real-time UI logging and gate command.
Authorized Entry
Outcome B
Visitor Redirection
Unrecognized plates are flagged and redirected to the help desk terminal.
Manual Review