Thinking about installing a LPR camera on this pole. Here in my state we don't have front plates. So, what is everyone's take on best option? Just zoom in on the peak of the road and make the trigger catch snapshots of incoming vehicles and then another as it leaves with the second one capturing the plate? Let me know what you guys think.
Yea what I was thinking. A bullet zoomed in on peak of road at plate height should work.. one question.. if camera is set to take several pics for the lpr will the software still recognize the plate even if it's days the 4th or 5th pic taken ?if that makes any sense
If you are using Plate Recognizer with Blue Iris, then it takes some trial and error to get the image right, or you run it with DeepStack and a custom model to only take a pic if there is a plate in the camera view. On a cul-de-sac, you could probably stay under the 2500 images per month.
If you are using OpenALPR, they use the video stream. Plate Recognizer has video stream option as well, but it is expensive.
Or you use the custom model that a member here wrote to read it for free:
I took a break from my DeepStack Custom Model project and started work on a DeepStackALPR solution. I made a lot of progress on the DeepStackALPR. The solution is API that crops and rotates (if needed) the license plate. Then the API reads all of the characters in the plate and logs the license...
If you are using Plate Recognizer with Blue Iris, then it takes some trial and error to get the image right, or you run it with DeepStack and a custom model to only take a pic if there is a plate in the camera view. On a cul-de-sac, you could probably stay under the 2500 images per month.
If you are using OpenALPR, they use the video stream. Plate Recognizer has video stream option as well, but it is expensive.
Or you use the custom model that a member here wrote to read it for free:
I took a break from my DeepStack Custom Model project and started work on a DeepStackALPR solution. I made a lot of progress on the DeepStackALPR. The solution is API that crops and rotates (if needed) the license plate. Then the API reads all of the characters in the plate and logs the license...
Record continuous. Then if the recognition software does not get you a good plate, you can always fall back to looking at the video with your own eyes.