Recommended Posts

I was asked by a possible client if I was able to fly over about 2 square miles with a infrared camera that would be able to discern the temperature gradients of the ground.  I told him I would have to do research as I currently don't own any infrared or thermal imaging equipment. 

My question is there a camera out there that can depict that level of detail?  I don't think the XT has that ability to be as finite as this individual wants.  Anyone have personal experience with the XT or is there a "better" system out there?

Thanks!

Link to post
Share on other sites

Hey Juggernaut and gang, 

My name Is Brendan Stewart and I taught the Intro to Aerial Thermography course. I've got a fair bit of time on the Inspire and XT, and Alan sent me a note about you're question. 

Short Answer: Yes, the XT can easily determine soil surface temperature. 

2 miwould be a haul for the little inspire though. I punched the mission into Drone Deploy for giggles, and at 400' you'd end up needing about 103 minutes of airtime to complete the mission. That's 7 TB48s through the Inspire just to run the path once. 

Long(er) answer:

have a conversation with your client to determine mission goals and outputs. My spidey-sense tells me this is Ag related, or at least a solar study. understand the information they're looking to acquire, and how they intend to use it. I've had a lot of questions from students in the thermography course about Infrared and Ag. Its important to remember infrared is a broad spectrum just like radio waves, and light. The XT and other thermal imagers only look for heat, or long wave IR. Most of the Ag applications use Short wave IR, or NDVI, or both. 

The Bottom Line:

Long wave Infrared (LWIR) imagers like the XT are fantastic tools for making inferences about how heat is transferring through materials. Make sure you clarify what types of measurements your client is requesting ti qualify the opportunity. If you're looking for a simple "hot soil here" reading, the XT will do just fine!

 

Hope this helps! 

  • Like 2
Link to post
Share on other sites

The client own's a bunch of different oil wells/ fields and uses the thermography as one of his many tools to help decide where to place wells.  So it is just a soil/rock temp he looking for.  

Just not sure how easy or well thermo images stitch together.  Probably need to fly RGB first then attempt to lay the thermal stills over the RGB?  

 

Link to post
Share on other sites
57 minutes ago, Uaviator53 said:

Just my 2 cents here. Two square miles is 1280 acres. That is really a job for a fixed wing UAV. Consumer grade quadcopters were not really intended for this size job. You'd need something like this.  $6k for the bird and $5k for the FLIR. 

I know this is more for a fixed wing, but at the moment all I have is my m100 which does well for me.  I can't buy a fixed wing for one project, but if I could get a contract with him for the rest of the fields then it's a no brainer.

Link to post
Share on other sites

Got it!

Thermal will stitch, but be ready to use serious overlap. I use 90% frontlap and 80% sidelap flying roof missions, and that tends to create enough pixel density to generate a useful point cloud. If you want to be absolutely sure, fly a high overlap grid with a departure heading of 360 (so, north to south) then fly a second grid with a departure heading of either 270 or 090 (straight east or west) then stitch both passes. Make sure your departure headings are upwind if possible, any headings seperated by 90 degrees will work You really need the overlap to provide enough data for the stitching software to work with. The more "pixels on target" you provide the software, the better the Ortho and point cloud will be. Also, radiometrics help immensely, and pick high accuracy GCPs  with an RTK GPS if you can. 

Now, onto the practical. A 2mi2  mapping mission with 90% frontlap and 80% sidelap at 400' would take about 473 minutes to complete. That's 7.8 flight hours, just for one pass. If I had a bunch of room to take off and land, and a bunch of money I'd fly this with a Precision Hawk. I'll assume you don't have $40k to work with, so I'd set up the mission like this. 

Day 1: RGB Ortho/DSM

X3 or X5, approximately 9 batteries. (factoring in +2 for weather related efficiency penalty) Estimate 4 hours onsite. 

Day 2: N-S Thermal 1st pass:

 Zenmuse XT, 12 batteries, generate 50% of the North-South scan. Estimate 6 hours onsite

Day 3: N-S Thermal 2nd pass:

 Zenmuse XT, 12 batteries, generate 100% of the North-South scan. Estimate 6 hours onsite.

Day 4: E-W Thermal 3rd pass:

 Zenmuse XT, 12 batteries, generate 50% of the East-West scan. Estimate 6 hours onsite.

Day 5: E-W Thermal 4th pass:

 Zenmuse XT, 12 batteries, generate 100% of the East-West scan. Estimate 6 hours onsite.

Yeah, I can see it easily becoming a week of work before processing, so make sure to budget accordingly. Are you planning on using Pix4D locally? or on the cloud? that'll definitely add some time to the process after all the imagery is gathered to make it intelligible.

Bottom line, You've got this! It'll just take a little sweat. Estimate accordingly. 

Edited by BrendanS_AVI
  • Like 2
Link to post
Share on other sites

I was going to build my own render server, but the building for my home grown shop isn't complete yet so I will take a look at the amazon aws...because that will take some monster CPU crushing to get all of that rendered.  A Huge shoutout and thanks to @BrendanS_AVI for your incredible response! 

Once I get some decent weather I will give it a shot.  I quoted 3.50/ acre for regular ortho and if thermal or 3d then its 5.00/acre.  With the time for flights/recharge and processing it may be a bit high...dunno first job at this size and level of difficulty. 

Link to post
Share on other sites
On ‎1‎/‎26‎/‎2017 at 1:52 PM, BrendanS_AVI said:

Yeah, I can see it easily becoming a week of work before processing, so make sure to budget accordingly. Are you planning on using Pix4D locally? or on the cloud? that'll definitely add some time to the process after all the imagery is gathered to make it intelligible.

 

I actually will be using a SlantRange and their software called Slantview, haven't used their software as of yet, but this will provide a good test for it ;)  I am going to demo Pix4d with my zenmuse z3 to see how it behaves with this large of a project.

Link to post
Share on other sites
2 hours ago, Juggernaught said:

I was going to build my own render server, but the building for my home grown shop isn't complete yet so I will take a look at the amazon aws...because that will take some monster CPU crushing to get all of that rendered.  A Huge shoutout and thanks to @BrendanS_AVI for your incredible response! 

Once I get some decent weather I will give it a shot.  I quoted 3.50/ acre for regular ortho and if thermal or 3d then its 5.00/acre.  With the time for flights/recharge and processing it may be a bit high...dunno first job at this size and level of difficulty. 

My pleasure! I love seeing an ambitious project come to life :) 

When we looked into AWS, it was incredible how much processing power you could get your hands on. Not only that, but you only pay for the hours it runs! A p2.16xlarge instance we used had 64 virtualized CPU cores, and 732GBs of ram for something like $11/hr. It ran a 1,500 image map in minutes.

Maybe not as fun as building your own rig, But that's more horsepower than I could fit on my desk anyway!

I think that's a solid pricing model. Ends up at $6,400 total, minus the fixed costs for aircraft (and XT!) depreciation, insurance, cost of processing and so-on. If you're just providing the Ortho, you'll end up billing out 4,480 for the same amount of work, same insurance cost, same software processing cost, ETC. 

With the technical expertise that's involved, I think your pricing model is quite fair! 

Link to post
Share on other sites
3 hours ago, Juggernaught said:

I actually will be using a SlantRange and their software called Slantview, haven't used their software as of yet, but this will provide a good test for it ;)  I am going to demo Pix4d with my zenmuse z3 to see how it behaves with this large of a project.

Slantrange eh? Looks really cool. Haven't had much time to play with that software yet, but it looks really well suited to your app. Would love to see your outputs! Good luck!

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.