Hacker News Reader

Comment by eck | Hacker News Reader

1. Alice installs app on smartphone. Alice puts smartphone in cradle on dashboard. Smartphone captures license plates Alice drives by on way to work. Alice drives by Bob.

2. Carol goes to website. Carol enters Bob's license plate number. Website says, "25 hits found. Enter credit card number to see times and locations". Carol enters credit card. Carol finds Bob.

3. Alice gets some of Carol's money, since she took the photos.

Why is that not a thing yet? License plate reading, unlike, say, tapping undersea fiber, is not something you need either a government's resources or a government's ability to skirt the law to do. This is easy stuff.

View on HN

It's already been done. PIs, LE and insurance companies utilizes data of a few companies that have scanners all around US; it's just not a social thing at the moment because when most drive they either text or check email or browse net or listen to the music. Dashcams are still not popular here but the services like you mention could be a huge hit in Russia when dashcams are extremely popular.

Here only few would actually go into lengths to set it up on their dashboard and be capturing licenses with the hope to make a dollar.

Good start to read more: http://consumerist.com/2014/03/05/the-repo-man-might-be-scan...

Exactly right. This isn't a hypothetical. This is already being done by repo companies.

Here's another link: http://www.betaboston.com/news/2014/03/05/a-vast-hidden-surv...

Mistake is #2. Don't make people search for the info, give it to them without friction. "Pay" them by providing more info. Real money would be in paid data mining.

Add ability to comment on other drivers. "Idiot cut me off!!!" turns into automatic notification when any instance of the app sees that plate ("Warning: 8 reports of risky behavior for vehicle"). Include automatic marking of "seen before" ("augmented familiarity").

I think you actually have a fundable idea there.

You've also accidentally found a very interesting side business model for Mapillary.

(except that afaik they've done considerable amount of work on actually detecting and blurring license plates, so I guess Peter would rather sink Mapillary than entering this business ;-) )

I think the actual reading of the plates is a much harder problem than "Smartphone captures license plates" otherwise, I concur, fundable.

There are a number of open source libraries for this. Here is one: http://www.openalpr.com/

Eh, recognizing state versus state has to be difficult. Can't rely on color.

Could work in Russia, massive existing installed base of dashboard cameras.

It would be better though if the dashcam had GPS as well.

It's possible to get around it by selling a device you plug the camera into, and your device has built-in GPS. I think Russia's dashboard camera "culture" might mean the device is a much easier sell - in Russia "Monitor traffic" is a problem people already seek solutions for, while the dashboard cameras is a good enough quality video people don't have to pay for again, whereas in, e.g. USA you will have to convince people they have this problem in the first place, and to pay for the cost of the device including a good video camera and storage.

In Russia people buy dashboard cameras because there is less trust in the "system", and is consistent with the paradigm:

License plate reading, unlike, say, tapping undersea fiber, is not something you need either a government's resources or a government's ability to skirt the law to do.

Maybe tie the the back end into repossession company databases. Have the user turn it on when in parking lots and they get a commission when found. A local repo service is instantly alerted with GPS coordinates whenever there is a hit. The app could even alert them to stop and observe until the repo company gets there.

Certainly not a mass consumer thing, just like most people won't become Uber drivers, but some people would do it and there is potential for making decent money.

>Alice installs app on smartphone. Alice puts smartphone in cradle on dashboard.

I tried this, but camera phones are (currently) too terrible. I would be curious if anybody more skilled than me (I'm just tinkering) could make it work.

Not all cameras in all phones are too terrible, that said having done the experiment, I found it is much easier to do license plate extraction from a video rather than from some sort of stills. Using an off the shelf "GoPro equivalent" on the dash of my car, post processing the video for plates was relatively straight forward with OCV[1] and now that I think of it, would make for an excellent tutorial on the technology. Were I to do it again I would provide a copy of GPS co-ordinates on the screen to make it a bit easier to tie the two together.

For me the interesting bit I was trying to figure out was this technique (car mounted) gave you license plate 'segments' which is to say the car in front of you may have been there for two blocks and then turned right giving you a segment that was two blocks long with a right turn at the end. If you had multiple subscribers, then you could match that segment up with a segment from another vehicle and then interpolate the path of any observed plate through an urban area. Since the general color and shape of the car is trackable, you can actually create a number of segments for different cars if you've captured their plate at any time, and if you manage 'untagged' segments (cars seen but you have yet to see a plate for) then you can add those as well. Needless to say, the taller vehicles are a "win" here for more complete surveillance.

[1] http://opencv.org/

I was also using opencv, but I had a ton of trouble getting clean video from a couple different cellphones and a tablet (lot of blurring from movement) and was putting off buying a gopro to try. This is great info, thank you.

Some commercial units apparently have tail-light detection to narrow down scanning for plates.

See Mappilary work on this area. Although they work on static pictures, I wouldn't be too surprised if their work could be succesfully applied to real-time video.

Another area you can explore are commercially available units. French police use on board license plates readers which seem to be quite fast (at least when you watch the french equivalent to the "cops" tv show).

http://www.teb-online.com/fr/analyse-images/lapi-lecture-aut...

LAPI(Automated Reading of License Plates) is the acronym you want to look for. I think several hardware and software manufacturers equipped different police units so if you're really interested, you'll be able to find several datasheets and derive a reasonable minimum computing power from them.

What if instead of using or in conjunction with using vehicle mounted cameras, you also utilized traffic cameras and street cameras. Many are accessible online.

Were you able to figure out the states?

No, that would have taken a lot more work. The algorithm was really primitive, find rectangles, look for letters and numbers which were > 50% the height of the rectangle.

>Why is that not a thing yet?

Seems sort of like crowd-sourcing data that would be used to infringe on the freedoms and happiness of people without restraint beyond "did the payment clear?"

Count me out, thanks.

Sadly, I do not think your willing participation would a requirement. This seems like the sort of thing that would be imposed upon you by the sheer force of adoption.

That will end up becoming an interesting court case, followed by a set of laws, followed by a court case, until such a thing is almost certainly illegal (in the United States), by pressure of societal desire alone.

One could do something similar with facial recognition software, but if someone tried to make a large business out of it, I'm sure there would be significant friction.

Which is why it should be done in the most outrageous fashion by someone who cares about privacy.

Maybe that's what is needed to get people to realize just what the next decade or two have in store.

You're trying to get the frog to jump out of the pot before it is too late but you're underestimating the degree to which it has already become drowsy.

there may be significant friction and even laws against it, but the government, behind closed doors would pay you a significant sum for your company to continue doing it and I would wager would lobby the government to pass bills to not only make it legal, but protect its legality.

The value proposition for the user is very weak - too weak to get people to spy on each other.

What are the odds that I'm going to get a payout? That someone out there is interested and willing to pay for one of the license plates I've seen? Probably very low. Do I really want to waste my smartphone battery doing this? I'd rather have maps and music on.

That's why you would pay the drivers to run this app. Then via the Long Tail, they could get royalties every time the data they collected is re-sold.

This might work for fleet purchases, but the economics for individual motorists don't look favorable.

You would spend more charging your phone than you would get from the service.

> What are the odds that I'm going to get a payout?

The expected value of slot machines is negative too. Humans have a psychological weakness for "Just one more, I haven't won in a while, I'm due." In this case though it wouldn't even cost money to play.

The expected value of slot machines is negative, but there is that slight chance of making a significant amount of money. That sort of "lottery odds" behavior doesn't really work here, imo.

Then make it a lottery. Payouts for certain licence plates at certain times of the day.

I think the value is when the local police pay for the query as well.

>The value proposition for the user is very weak

The value prop for the average user is also very weak for encrypting data. I am more likely to lose all my data than prevent some theoretical malfeasance, never mind the extra time and effort. I am not a secret agent.

Tie it in to law enforcement, get paid for detecting someone driving a vehicle with a warrant against their number plates.

Or an Amber Alert

Insurance companies too.

Generally, a system like this might work by handling queries via two ways:

* A "point" system - spend "points" (or credits) to find a user by license plate. * You can buy credits with money directly, or you can earn them by contributing information

That way, if you're a data contributor, you have access to the information without having to pay. That's how data.com works.

They don't seem very popular in the US, but always-on dashboard cameras already have obvious value. Maybe a dashboard cam cloud storage company could offer a discount on their services if you allowed license plate data to be gleaned from the footage you upload. Or maybe they would just throw a line in their ToS and do it.

IIRC they are very popular in Russia because the insurance companies reduce your prime. This cause a lot of interesting videos, like the meteorite a few years ago, and a lot ow weird accident videos: https://www.youtube.com/results?search_query=dashboard+camer...

I would be happy to submit free videos to the police if they'd issue citations. When people break the traffic laws they are almost always putting other people's lives at risk. I'm happy to pay some of my time (pushing a button that uploads a 20-30second video of the violation) in exchange for safer roads.

I believe that in many European countries the data protection laws offer a certain degree of protection against this.

This is already being done by companies and police forces using their own vehicles. It's more efficient than outsourcing it through an app because cameras mounted on the roof of a vehicle can scan license plates in 360 degrees.

This is very common with towing, repossession, or "recovery" companies, who can drive around with automated cameras looking for "hits": cars with repo orders. Companies (ex: Digital Recognition Network [DRN]) sell cameras, provide a database of "wanted" vehicles, and collect the incoming scans—which they also turn around and sell.

DRN "claims to collect plate scans of 40 percent of all US vehicles annually."

You don't need to crowdsource because there's already a critical mass of users willing to pay YOU for the privilege of collecting information.

There are a lot of business models in crowdsourced image data. For instance, Google could give away high resolution dashboard cameras in return for the right to use the video for ground truth data. With sufficient intelligence onboard higher-level information like signage, license plates, or even faces could be uploaded on demand. You'd probably even be able to bundle an insurance premium discount in.

I'm sure there are various tracking networks that track browsing across websites, it is sort of a logical extension. UPS or Fedex could easily build out a big tracking network.

This is trivial to do for a number of feedback/analytics/cdn/socialnetwork companies. Basically, the people who have JS living on large swaths of the web. There are some names you sorta expect (Facebook, Google, Twitter) and other names like Mixpanel, Foresee, Opinionlab, Survey Monkey that might not be as obvious.

It would be pretty easy for a Foresee to deliver a customer profile aligned on verticals in near real-time to clients based on data their network captures from major websites around the world.

In the EU, that's illegal under the Data Privacy Directive.

Reply on news.ycombinator.com