Hi everyone,
I'm working on an ANPR (license plate recognition) project in B4A using:
CamEx2 for the camera
YOLOv8 (TensorFlow Lite) via a custom Java wrapper
ML Kit TextRecognizer (Latin) for OCR
Currently:
License plate detection is working correctly
The red rectangle is aligned with the preview
The license plate is cropped correctly
I've implemented simple tracking to avoid continuous OCR
I've optimized the OCR so that it only runs when the license plate is stable for a few frames, to improve performance and reduce latency.
The link is below.
, , , ,
I'm working on an ANPR (license plate recognition) project in B4A using:
CamEx2 for the camera
YOLOv8 (TensorFlow Lite) via a custom Java wrapper
ML Kit TextRecognizer (Latin) for OCR
Currently:
I've optimized the OCR so that it only runs when the license plate is stable for a few frames, to improve performance and reduce latency.
The link is below.
, , , ,