algertc
[Feature Request] Support for direct camera ALPR data without CodeProject.AI dependency
Problem: I have a Hikvision ANPR camera that already does plate recognition and sends data to Home Assistant. The current API requires CodeProject.AI format, but I already have the processed data.
Requested Feature: Add a simple REST endpoint for direct ALPR data submission:
POST /api/plate-reads/direct
Headers: x-api-key: YOUR_API_KEY
{
"plate_number": "ABC1234",
"timestamp": "2025-01-15T10:30:00Z",
"camera": "entrance_gate",
"direction": "FORWARD", // or "REVERSE"
"image_base64": "data:image/jpeg;base64,..."
}
Benefits:
- Works with any ALPR system
- Better Home Assistant integration
- No redundant AI processing
- Maintains all existing ALPR-Database features
Would this be possible to implement?
I had been wondering whether those NVRs could send data out. Delighted to see this actually.
Yes, it's possible and should be pretty quick. New release is coming out in the next few days after I add a couple extra MQTT features. I'll squeeze this in and get it in the release.
Could you please copy and paste this into the roadmap? That's what I usually work off of and track progress with. https://alprdatabase.featurebase.app/roadmap
Thanks for writing such a good request and including all the info I'll need. Will make it easy to implement.
Thank you so much for the quick response and for considering this implementation! I really appreciate how receptive you are to community feedback. I've already added the feature request to the roadmap as you requested. Just to provide some additional context about my setup: I have a Hikvision iDS-2CD7A46G0/P-IZHS camera that performs all the ALPR detection internally (no NVR involved). The camera can send metadata via ISAPI (Hikvision's protocol) or by sending images to an FTP server with metadata embedded in the filenames. Currently, I'm using the FTP approach with Home Assistant's FTP addon to receive images in the media folder, combined with the Folder Watcher integration to monitor for new files. When a new image arrives, I extract the metadata through an automation and store it in a sensor - this is where I could forward the data to your ALPR-DB. For each detection event, my camera sends 3 images via FTP:
20250806171539530_PLATENUMBER_FORWARD_COUNTRYCODE_VEHICLE.jpg (cropped vehicle image) 20250806171539530_PLATENUMBER_FORWARD_COUNTRYCODE_PLATE.jpg (cropped plate image) 20250806171539530_PLATENUMBER_FORWARD_COUNTRYCODE_BACKGROUND.jpg (full camera image with metadata overlay)
I noticed that when data comes from CodeProject.AI, it includes the plate crop "coordinates" within the full image. Since my camera provides pre-cropped plate images, would it be possible to send the cropped plate image directly in this new endpoint instead of coordinates? Thanks again for developing this excellent tool.