Data Acquisition and Processing Strategies for Linear Drone Mapping Projects
We are from you, hey everyone and welcome to another awesome edition of ask drone you. My name is paul and my name is rob and yeah, certainly hoping that you think its awesome. We certainly think its awesome to be sitting here hanging with you, and we definitely appreciate that. You are spending a few minutes of your day with us and for those of you that have taken some time to send in your question to askdroneyu.com. We are very appreciative for those of you that have become members. We are very appreciative for those of you who have left reviews, even those that we arent very excited to hear about. We still like to learn from you and know how we can do better. So we appreciate you too, so thank you guys for being you and uh lets hop in yeah. Weve got a great question today, um regarding very large data sets, which actually kind of coincides with an article uh that i wrote this week about um the best mapping drones for small areas and for large areas, as we had some students rob who are still buying The wrong drones because everyone thinks its more convenient to get that phantom 4 pro plus, but you cant download any software on that little screen, so uh its actually useless um. I feel like that is a nice word to use um and so that for mapping yeah for mapping – and that said, uh also talked about you know the wingtra one and how powerful that is for these larger sites and uh.
You know. I think that this is an interesting question that we have here today, just kind of regarding deliverables, because um a lot of people run into problems when they get in these larger sites and they and they realize the limitations of their computing power. And so i think this is going to be a good one, and i just want to say if any of you are really looking to. You know find a more sequential chronological training on drone mapping. You want that comprehensive drone mapping training. You know we have been talking for a while about the props program and how its really built for teams, but it can also work and is built for singular pilots as well. And if you thrive in a more structured environment – and you want that more sequential training. More frequent quizzes tests, you, like the scenario based uh, you know, certificates and whatnot. I really recommend you check out props.thedronyou.com or propsflightschool.com check it out. I think youre really gon na, like it also uh in the hopes that uh we dont get another shutdown. We are planning and in texas, obviously uh. We are planning an experience. Training in texas, which is going to be seven days of training and uh, were hoping. It all goes through uh because its going to be a real world uh mission for pilots and its gon na cover creative and technical aspects. Think of it really like a mini fly in and then were ending it with karas brand new class about helping drone pilots scale all of their business systems right from the start, just to make it a whole lot easier to run your business over time.
This is not, you know a um whimsical here are some, you know: uh, templated logos and and marketing material. This is like actually autonomy for your company at every given level to really allow for guys like myself that sometimes struggle with maintaining the hustle over time. It just makes it so much easier, so the experience training will be open to eight people. Only and the sales class will be open to 20 people. Only kara will be teaching that its really about business systems, not so much sales, but it does include autonomous upsells, which is really nice um. But that said, im hoping that we get to do that. We are going to require the prerequisite of our mapping class to attend that training, at least the experience part of the training, and i i brought all this up, because props is a great way to go through that. Otherwise, you can become a member and check out the comprehensive mapping class and kind of go at your own pace, but that was actually a nice little segue into our sponsor for this question, which is props.thedroneyou.com or check out propsflightschool.com, hey guys, this is gabe from maryland. Had a quick question regarding deliverables for linear mapping projects, i just got asked to map 30 miles of transmission right of ways, and accuracy is not very important, so gcps are not required and with it being 30 miles, im gon na have to break this up into You know several sections to comply with part, uh, 107 rules and regulations, and you know stay compliant with that.
So my question is: what would your deliverables be? If you were me um and you, you know, flew these 30 miles and had to obviously break them up into sections or through or flew. You know 50 miles and had to break them up into numerous sections and gcps were required or even if they were not required, would you provide them an overall large orthomosaic and digital service model, and if so, how would you do that and how would you also Provide you know the other deliverables that you may offer im curious to know what you guys think, because i keep kind of running into an issue of getting you know into qgis and merging these tips in there. However, the tips are too large to do so. So the merge fails. You know looking forward to any solutions that you guys might know about and uh. I look forward to hearing your response. Thank you so much. I appreciate all you guys do and have a great day thanks. Gabe um appreciate the uh detailed question and sounds like you got a lot going on. There sounds like youre, also in the middle of a pretty steep learning curve, which i feel you but youll get there and i think paul might have some ideas, for you would be my guess. Well, i mean, i think you know, rob uh often times we kind of follow the clients lead uh in regards to deliverables. Typically, they they have a specific problem that they want to solve, albeit you and i both know.
That is not always the case. As often times uh, we have seen even public utilities purchasing aircraft without having any idea how to use them or who will fly them or what to get out of them. So while uh its easy to sit here on a high horse and be like well hey sure, no water lemon rolls, i dont think we, i dont, think its so easy to uh to assume and in an effort to not be an rob im, not gon na Assume im going to ask some questions in an effort tonight thats a good effort. I think thats a solid effort to pursue uh. Maybe peter will try it as well. I dont know i doubt it. I doubt it new yorkers are a whole different animal. Sorry guys, love you, sorry, not sorry peter uh anyway, um that said, uh deliverables, first off um, without getting into specific deliverables as far as an ortho, which is probably what theyre looking for at the end of the day and ignoring uh the digital surface model. But at least he said dsm, i mean often times a lot of people uh, you know say that they are going to get a dem, an elevation model and not a surface model, and there is a difference between those. So we got to give them props there, get it props, props.the, droneyou.com, um, but uh anyway. Long story short, is uh shameless uh. Oh absolutely, i got no shame um, but that brings us to a good point which is hes having difficulty merging those tiffs together and he uh.
He foresees issues in being able to do that. So i want to bring up a couple of potential solutions. Um under you know, he mentioned that its 30 miles, if we kind of take the vague definition of visual line of sight from the faa. This sounds like its going to be about 60 sections of flights and in an effort to get that entire 30 mile segment of data to auto merge together, um youre, going to have to have like 10 to 20 percent overlap from one mission set to the next Mission set, so i mean like imagine the computer that you, you know, you mapped a say lets say halfway. You would want to start, you know, say 10, 20 in and then complete the rest of that section. When you do that, then you can actually merge all the photos together in one project in pix4d mapper. But so a couple butts here. One 30 miles is a lot. Is it possible on mapper? Yes, ive seen much larger than that in mapper, but he might want to look at pix40. Matic matic is made for those larger data sets um. That said as well, you know we dont know what kind of computer he has rob, but lets assume that he is not like most people who completely overestimate the power of their computer uh and then come to mapping class and have a harsh reality when they get Here but um that said lets imagine hes got a super powerful computer um.
I honestly, if georeferencing is not imperative, then what i would do is i would use pix4d react and the reason that i would do that is twofold. Uh number one youre gon na get the ortho the fastest way possible. Number two with pix4d react. You only fly a single grid mission, so overall, the time for acquisition is going to be cut relatively in half, which is going to be very valuable for him. In addition to that, hes going to have the overall output much faster than matic or mapper lets give them another option. Uh. The other option is: is that if again lets imagine that hes got a computer fast enough to run all the data, or maybe he doesnt lets go with he doesnt he can import all of those photos into pix4d mapper on his desktop now key point: youll have To have all those photos saved in a folder on your desktop, which this is something that still really irks me about pix4d using the cloud system. You always kind of have all that crap on your desktop and it slows your computer down astronomically its like a double negative. You know its like burning the wick at both ends here um, so what he could do is he could set up his project. He could set up step one. He could set up the processing options for step two if its non geo reference. Obviously, he can just run step one and two together uh, but then what i would do is i would push it to the cloud and i would let the cloud servers run it if his computer is not fast enough.
In order to do that, you will have to have all the photos on your desktop to do it so just recapping option, one pix4d react single grid. Nader lets say he needs more 3d data to really understand where those tree lines are well, then you know, maybe he wants to do. A double grid run pix4d mapper. He can run it in the cloud if he doesnt have a fast computer option. Three is pix4d. Matic, which is made for those larger data sets, but i think you know, as i analyze the options that weve given him on the show. It sounds like a single grid, nader, camera orientation, utilizing pix4d react – might be the best overall solution, as it will reduce his processing time. In addition, it will reduce the acquisition time, as the acquisition strategy will be a simple single grid rather than a double grid. Wow. Well done because 30 miles i mean its going to take some time. I mean 30 miles that particular job. If, if hes working as an employee for the the client, i think it would be worth getting that you know they call them beyond visual line of sight. Waivers but when you read through the fine print, you realize theyre more, like extended visual line of sight, and so i mean that might be in his best interest frankly to really look up applying for one of those, particularly if hes going to be doing more of These, yes – and we do have a template for that.
We have not released it publicly, but now we have one um and, if youre interested in in that, we would be happy to help you acquire that waiver. That said, uh overall i mean, i think, its a very interesting mission, and i would say that you know often times even us experienced mappers can run into missions that the problem thats trying to be solved, isnt quite clear um, as communicated from the client, which can Sometimes you know create a haze when determining what deliverables might really be adequate and on that point its a because he did not allude because hes asking you us whatever what deliverables he should be. Giving right. Yeah tells me he may want to go back to the client and have a couple conversations and make sure its understood. What theyre looking for a hundred percent agreed, because you dont, want to do a whole bunch of work because its going to be a fair amount of work. Even if you use something like react, its going to be a lot of work worked work and to find out that maybe a section they did want geo referenced or who knows i mean you got to kind of talk them through it 100, and you know if Hes also having issues in qgis or arcgis, i cant remember which one he said he was using um. I would then try to merge all the data in the processing, not in the post processing.
In addition, if you had processed those projects under different parameters or processing templates, you could run into stitching errors through arc and queue, which would be why its not working. So i could see a couple of potential issues coming up frankly, sure that makes sense yeah its gon na have to be very uniform to work. That way, i would imagine yeah totally cool, but i think that answers the question i i do want to remind everyone. You know we have a extensive guide on drone mapping. If you go to thedronu.com forward slash what dash is dash drone dash mapping dash software forward, slash um, you can check out a guide there. We do have a virtual mapping class coming up its not on the site yet, but we actually have in person mapping classes as well and were going to have them anyway, no matter what so they are going to be here at headquarters, and you know rob this Brings up something that really excites me, which is thanks to you. What really and your familia we have a nice. We have a really nice cool setup for mapping classes now out in the bosque, and i mean we did a mapping class out there a couple weeks ago and it was an absolute blast. It was so much fun, perfect subject for 3d and 2d assets. I mean uh im just im just yeah. I had a lot of fun im so glad to hear that im glad it worked out.
It is perfect, theres, nobody out there, its peaceful um airspace is completely clear. I mean it just checks all the boxes. It really does its a really beautiful house and a big house and its its nice. It is – and i mean its its just the perfect size yeah – to get mapping exercises in im not going to give the dirty details, but i mean um im going to post that 3d model here shortly, in fact reminds me i i want to go run another Project combo, and i think that that particular exercise will be very good for um our new mapping classes. In addition, you know ive been telling you that i want to do a complex 3d, modeling class and kind of you know showcase other softwares other than pix and its time to write the curriculum and get it done so anyway. Im excited for that fantastic um for all of you out there. I really do appreciate your support uh for those of you who are droney members, greatly appreciate it um and for all of those of you who are thinking about you know, is there a way? I can maybe level up my knowledge in one particular way or another. You might want to consider becoming a member of drone. U i promise its worth it. That said its gon na do it for us today my name is paul. My name is rob.