Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Friday January 31 2020, @02:54PM   Printer-friendly
from the don't-look-at-me! dept.

Anyone with a camera and $5 can now have a license plate reader:

Automatic license plate readers (ALPRs), software that allows computers to separate and analyze license plates from camera footage, could soon become ubiquitous in American neighborhoods thanks to a company called Rekor Systems. On Thursday, the firm started selling a product called Watchman. The $5 per month subscription allows homeowners to add the company's OpenALPR software to almost any home security camera.

[...] there are a couple of limitations to the $5 package. The software won't automatically log every single license that passes your home. As a homeowner, you'll also won't be able to obtain someone's name, address and location history from their license plate. That's a feature only law enforcement can access.

[...] privacy advocates fear the technology could be easily abused by both homeowners and law enforcement agencies to erode the privacy of innocent people further. And advocates have good reason to be skeptical of companies like Rekor. Amazon's Ring security service spent the majority of 2019 defending its partnerships with law enforcement agencies. In one instance, a report from Motherboard showed that the company had coached police on how to convince homeowners to hand over their Ring camera footage without a warrant. Similarly, it's easy to imagine a context in which police agencies could abuse the widespread proliferation of technology like OpenALPR.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2) by takyon on Friday January 31 2020, @03:06PM (3 children)

    by takyon (881) <takyonNO@SPAMsoylentnews.org> on Friday January 31 2020, @03:06PM (#951784) Journal

    Is it though? Doesn't look like it.

    In any case you could see an open version of this querying some crowdsourced database of license plates in the future. Even with names.

    https://www.openalpr.com/img/searchable-database.jpg [openalpr.com]

    --
    [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    Starting Score:    1  point
    Karma-Bonus Modifier   +1  

    Total Score:   2  
  • (Score: 1, Interesting) by Anonymous Coward on Friday January 31 2020, @03:42PM (1 child)

    by Anonymous Coward on Friday January 31 2020, @03:42PM (#951799)

    I'm pretty sure I know the answer to this, but what's the $5/mo for then?

    • (Score: 5, Touché) by Immerman on Friday January 31 2020, @03:53PM

      by Immerman (3985) on Friday January 31 2020, @03:53PM (#951802)

      Profit, what else?

      What, you think you should be allowed to contribute to their massive for-profit surveillance network for free?

  • (Score: 2, Interesting) by Ethanol-fueled on Friday January 31 2020, @05:55PM

    by Ethanol-fueled (2792) on Friday January 31 2020, @05:55PM (#951858) Homepage

    Beat me to it. But with OpenCV-based stuff, even a beginner/intermediate level Python coder could bang out something similar with Google-Fu finding open repos and tutorials and using all the usual other useful Python stuff: OpenCV, SciPy/NumPy and all other supporting packages associated with those.

    One could have a basic example working in literally a few minutes, then the rest comes with either getting more training sets (images if applicable) and then figuring how to add more functionality to it to increase plate reading accuracy or associate the plate to the vehicle color and type ("white minivan," etc.).

    A good approach for an interested noob would be to start with basic facial recognition-style example. Download sample code that can detect features such as eyes, nose, mouth (a common basic example is the use of Haar cascade) and then download some example code that can properly detect spectacles and not be confused by it. Look at the two codebases and figure out how to splice the glasses-detecting code in with the eye-detecting code. Then modify the code to prevent facial moles from falsely being identified as eyes (I have personal experience with this particular problem when I was getting used to OpenCV, hahah.) Etc, etc.