Monday, August 22, 2016

Automatic Plate Readers Using Dash Cams and Backup Cameras

Project Status: App Idea


Ever hear an Amber Alert and think, wow, the chances of me remembering that license plate number and seeing the vehicle are slim to none?  Have you ever had your vehicle stolen? Each of these scenarios makes me feel bad and helpless.

Stealing cars, abducting children and basically committing crime in general should be made as difficult as possible in an effort to make it not worthwhile for someone to do so.


Currently plate readers are in active use in places such as toll booths, traffic cameras and even on most police vehicles. What I am proposing is casting a wider net to blanket less traveled areas and those people that would attempt to hide from commonly known plate readers.

How effective has Amber Alert been?

according to
As of December 23, 2015 there have been 800 children rescued and returned specifically because of AMBER Alert.


I would think it would be great to develop an Android Auto app that uses any of the car's camera feeds as a source to search for license plates. It is a perfect time for someone to start working on an app like this seeing as Android Auto is starting to show up in more and more cars as well as many cars now coming with their own WiFi packages and GPS.

Reading a license plate from a camera feed is relatively easy, many people have already been working on solving this problem and open-sourcing solutions for general use by others. It is getting so good that you can actually also log the color, make and model of a vehicle.

I think that reading plates and logging them could run in the background, as a low priority service, only when there are enough free resources to do so. In other words if my device is already bogged down with me using navigation or other features then as a user I would rather suspend this service than to have a laggy experience with the device.

The end user should not have knowledge or access to any of the specifics gathered by the app. I think the data should be sent directly to the police or agencies that are using data from plate readers already in use.

The end user of the app should only receive aggregate and generic reports sometime after the fact in a way that they are unable to pinpoint which vehicles were notable. Giving users instant feedback with alerts like "Stolen car in front of you" would likely cause some users to put themselves and others in danger as they personally become vigilantes trying to chase down stolen automobiles or abducted children.

Contrary to how most plate readers are used today, the device reading plates should not have any database available to it of suspicious plate numbers. The data should instead be sent off to more secure systems so the proper authorities who already have these databases may compare and react to results as deemed appropriate.

Reports could show things like the number of stolen vehicles reported or the number of Amber Alerts which had been reported in cases where the children were safely returned home. I know I would love to see data like that and be motivated by it rather than the feeling of defeat I currently experience as I read that passing highway sign showing "Elderly missing" followed by a vehicle description.


All those things we intend for good could be used for bad. I think it's important that we do our best not to blindly supply this information to the powers that be but that we also diligently keep them in check and track their usage of it.

I think we should always assume that this information can be hacked, as of course anything can be hacked. Viewing the data from that mindset we should do our best such as by keeping from storing sensitive information for too long of a time period.

The data should likely be treated as a starting point for investigators to begin their research rather than as definitive court worthy proof. People could likely find ways to report false data and for many reasons such as this the inflow of data should be treated as an anonymous hotline for tips rather than an evidence stream.

I truly think that there should be a strong process set into place to approve access to the data in a timely manner so that it is still current enough to be useful, but at the same time protects it so that people can't view the data for any personal or unnecessary reasons. I believe that to help with that there should be an audit log keeping track of who looked at what data, when they did and the case involved which initiated the search.   

If you are one of the many people who are totally against any big-brother type logging such as this I would love to hear from you.