tomsriddle/alpr-engine repository overview

⁠Automated License Plate Recognition

⁠Overview

This image provides a script to read video streams using ffmpeg and detect license plates

⁠Usage

⁠Pull Image

To pull an image use the following

docker pull tomsriddle/alpr-engine:1.0

After pulling the image check that it is present using following

docker image ls

⁠Run Image

To run the image use following

docker run -p <host port>:<container port>/udp -v <host directory>:/workspace/shared-data -e PYTHONUNBUFFERED=1 -e video-stream-url=udp://host.docker.internal:<container port> -e video-stream-base-folder=/workspace/shared-data/license-plates/ -e predictions-run-folder=prediction2/ -e original-image-folder=original/ -e cropped-image-folder=cropped/ -e models-folder=models/ -e prediction-threshold=2 "tomsriddle/alpr-engine:1.0"

Note: See the volume mapping - this is needed for the images sub folder where the images captured from the video stream will be kept. Also notice the port mapping has the "/udp" suffix to ensure that you are reading via UDP

⁠FFMPEG

The image assumes you are running a video streaming server using FFMPEG. For this first install FFMPEG https://evermeet.cx/ffmpeg/⁠

After installing FFMPEG you can use a command like following to start a video stream

To run the image use following

ffmpeg -i <MP4 Video> -qscale:v 2 -vcodec mpeg4 -f mpegts udp://<host>:<port>

An example below

ffmpeg -i LicensePlateReaderSample_4K.mov -qscale:v 2 -vcodec mpeg4 -f mpegts udp://127.0.0.1:23002