Show HN: Finding and plotting solar panels in drone videos using computer vision | Hacker News Reader

Show HN: Finding and plotting solar panels in drone videos using computer vision

github.com

29 pointsyeldarb20 comments

Heads up, if you haven't already, make sure you have your commercial drone license!

Anything past flying only for recreation requires a commercial license. The FAA says "if there's any doubt, get the commercial license".

Having gone through it myself for mapping my forest land, it's not too difficult. You'll want to study up beforehand, but it's not too challenging.

Candid question but you’ll have to trust that I’m asking it in good faith -

What considerations are you making about your neighbors’ privacy, both about flying drones overhead, and then also ID’ing their use of solar?

That video looks like 30+ homes now exposed as owning Solar and open to whatever marketing comes their way now.

I can never get an answer out of drone operators about this. Since home owners can’t control who flies over head, the onus is on the pilot for thinking about it. I can’t tell if pilots ever do, but if they do, what conclusions do they come to?

Edit - I get the laws haha, to the point of my comment, what do pilots feel is right to do vs what’s legal? This is what I mean by rarely getting a straight answer.

As I understand it, the current state of the law (in the US at least) is that airspace is considered a public space governed by the FAA and you have no expectation of privacy in your front/back yard. It's basically the same as satellite imagery being made publicly available on eg Google Earth.

Whether this is or is not "right" in the age of ubiquitous drone availability is up to us (via our government) to decide/change.

Disclaimer: IANAL and before someone starts a company based on this, they should definitely consult one.

The pilot must maintain visual line of sight with the drone at all times, sufficient to determine heading of the drone by eye, unassisted by binoculars.

It must not fly over people or vehicles with people in them.

State privacy laws may provide restrictions above FAA airspace guidelines.

Drone operators could get in trouble for taking off or landing on private property.

Source: I'm a commercial license holder and have done some personal research into what's ok for the purposes of drone mapping.

Edit: That appears to be in Class B airspace, starting at 680 feet above mean sea level. Ground level there appears to be 250 feet above sea level, so flying a drone means you could go 430 feet above ground level before you do something the FAA gets real angry about.

As a drone operator, you can fly 400 feet above the ground OR 400 feet above any structure you are surveying. So it's something this pilot may want to be very, very careful about.

Yep, the Airmap app is really useful for getting those local restrictions & any applicable TFRs.

Right, so what do you feel as right as the pilot doing it (vs the laws)?

I can’t get an answer like this. Usually, it’s pivoted to discussing the laws like this vs the accountability of the pilot for the changes in privacy of those they observe (vs someone flying around their own land).

How do you think re: I should/should not do a project like this?

To put a finer point on it, Google Earth and the related ingestion and use is controlled and guided by laws vs a drone pilot and their personal decisions beyond a license requirement. So not apples to apples. Again, candid not critical, I know pilots get a lot of nastiness about this stuff.

> what do you feel as right as the pilot doing it (vs the laws)?

Personally, I think it's fine, and don't have a problem with people flying over my backyard & collecting images just like I don't have a problem with satellites adding images of my yard to Google Earth. I think it's pretty cool that I can zoom in and see my car in my driveway as seen by a satellite flying around the Earth at 17,000 mph.

Fair enough.

If there’s a differentiation between the data governance and privacy reqs the satellite is controlled by vs me and my drone, is it relevant to you?

That the large difference in the analogy you’re making.

I'd certainly be interested in any legal differences if there are any!

My opinion is that if the outcome is the same (a view from above is made publicly available) the how isn't a "large difference" from a moral perspective (which is what you were asking about).

Had a lot of fun building this; did an in-depth writeup here as well: https://blog.roboflow.com/georeferencing-drone-videos/

I think you could productize this and sell it to solar companies. I imagine this data would be helpful to them for things like selling add-ons, upgrades, etc.

I could see it be productized for anyone needing to do drone imagery!

Love that idea -- and I think there are probably several startups that could be spun up on top of this (off the top of my head: search & rescue, monitoring swimmers in triathlons, helping governments identify zoning violations like unregistered pools, monitoring core infrastructure like pipelines for potential hazards like excavators nearby, finding people with bonfires in high risk fire areas, tracking wildlife, finding boats fishing in restricted areas, counting cars in parking lots, monitoring construction progress, etc)

I would be thrilled to have someone pick this repo up and run with it. We're really focused on the developer tools and core infrastructure behind the computer vision part. We don't have the bandwidth to capitalize on all these opportunities (and this is only in the aerial space which is a tiny portion of the use-cases we've already seen for the technology; engineers at over 2/3 of the Fortune 100 have used Roboflow for things in every industry from retail to broadcasting to manufacturing to agriculture to healthcare).

Having meaningful businesses built on top of our tools is a great sign that we're building something useful (and we'd be ecstatic to help them succeed if they want to become a customer)!

This is really cool. I have friends running solar panel deployment companies and this would be extremely helpful.

Nice! I’d love an intro if they’re interested in using drones/computer vision to streamline their work

Very cool project!

Reminds me of https://www.dronedeploy.com - I think some solar companies use this to help map roofs and generate panel layouts.

How does it distinguish between solar collectors and solar panels?

It doesn't currently, this[1] is the dataset of ~50 images I collected that I trained the model on.

I'm not really an expert so I just labeled the things I thought looked like solar panels. It's mostly meant to be an example -- if you want to use something like this in a production environment you should definitely spend more than the ~2 hours on image collection, dataset labeling, and training that I did for this!

[1] https://universe.roboflow.com/brad-dwyer/aerial-solar-panels...

Yeah I was just wondering, it's probably non-trivial to distinguish between them. Cool project though!

Yea, I spotted a few panel-like objects that weren't captured in the example on the repo. Not sure if they were panels, but now I understand most of this is a framework for doing your own analysis. I suppose using your code, someone could also recognize pickup trucks, people, pools, etc. I think it's really neat that you can do stuff like this with drones these days, and am sad I live too close to a military airbase to use one for anything other than backyard piloting.

Reply on news.ycombinator.com