The first run of mapping with the Mavic.

[bakerboy]

Just hope they start selling light planes with gimbals (actually never searched for such models, maybe they already exist).

I'm a Mavic and Phantom owner, but my subject of interest (mountains) is too large for them.

A simple servo driven gimbal attached to an APM based flight controller will provide ample correction compared to what is currently used, when I get off my *** I will deck a bixler out with a setup like that with a gopro and 5.4mm lens for a trial, until then quads are the go for small areas

 

[bakerboy]

I did a quick model of a site where they required the fill volume and contours. I GPS surveyed it as well as flew it, it's a side project I'll get around to finishing but results so far have been very good
17 3015 Quarry In-Fill Model by West Coast Aerial Imagery - 3D model

 

[Sencorp]

You can data gather and output at a survey precision level with the use of GCP, a local surveying company does most of their surveying like this, and I have done the odd job like that too. TBH I don't see what you're bringing to the table if it's not accurate or can be used to get results from, do you do any ground sampling for the VARI plant health data? Whats the point of the DEM if you have no height checks? Contours will not be accurate if you're trying to show surface water flow etc

At the very least you can place scale bars for a greater 2D precision, just food for thought

When I do mapping its to take the area and turn it into a manageable edit file for construction and area experiment use. Meaning once I have done my mapping, I think get to work taking the mesh file and breaking it up into element that then can be placed in the VMS. Terrain can be changed, within terraforming simulation. You can add in a building, or remove it or move it. For simulations I do not require any of what you suggest. MY mapping, and my clients are project specific. And TBH in slip management there are no need for height checks. Just saying.

 

[bakerboy]

I guess I'm used to people wanting to be able to accurately design something from a feature survey, your clients but not something I would suggest to others to go out and promote without understanding the errors involved

I mean, whats the point of adjusting the terrain if it's not accurate to begin with, nor do you have any checks on the results DD and maps made easy have given you?

That's just my view, as a surveyor, when I give people data I know it's correct

 

[Sencorp]

I guess I'm used to people wanting to be able to accurately design something from a feature survey, your clients but not something I would suggest to others to go out and promote without understanding the errors involved

I mean, whats the point of adjusting the terrain if it's not accurate to begin with, nor do you have any checks on the results DD and maps made easy have given you?

That's just my view, as a surveyor, when I give people data I know it's correct

So to be honest I have been trying to steer this conversation between us away from the direction your taking as the post was to show the first run with using a MAVIC to map with. I feel now your taking a slight piss at me because somehow by me posting anything to do with mapping it somehow has offended you. If you would like to "Discuss" mapping techniques, I would be more then happy to do so. I employ several very talented senior survivors that work in various aspects of construction and slip management. I of course would never critique what it is you do, or question how you do it just from a few exchanges on a forum post. I would like that same courteousy. I can go through our H&S checklists before flights, LBS comparison charts ect ect ect, but that is not what this post is about. And frankly I am not fond of pissing matches on form boards that accomplish nothing.
If I am misreading your intentions then please forgive me.
If not, lets just leave it here then and get back to the post.

 

[bakerboy]

Simply pointing out that from your posts it would seem like you're promoting your method as a reliable way to map/collect data, I'm just concerned people will see it and think they can create decent maps without any control or checks, and market themselves only to get caught short. This isn't a pissing contest, I don't care how you do your work because it doesn't affect me, just not wanting to see people get the wrong idea that throwing photos into the cloud processing websites will net reliable results with meaning if they are unaware of what else is required

Can they create an updated ortho, and a fun 3D model to look at? Absolutely. Would I be using it to design from? Not me personally.

You wanted some feedback, I'm giving you my view This post is still very much on track. What is it your company is trying to achieve?

 

[Logger]

I like what you are doing @Sencorp
Would be interested to know how you do it. For example could I use this methodology to create a plant health map of of 20 acre paddock beside my house? (or is it too large an area) I am not interested at this point in ground sampling. More interested in identifying subtle changes in pasture that may not be immediately visible to the naked eye. I have DJI GS Pro. Can I use it to get the images then perhaps process them through something.

 

[bakerboy]

Logger the problem is that you will want a base point/ground sample if you're wanting to compare between flights, the change in angle of the sun can be enough to throw your results even on the same day, you will need some sort of control. This is my point I was trying to get across, you can't just throw the data in and expect a result you can use to compare between flights without control. Check check check, that's what data collection is about IMO

You can use your RGB images and run it through Drone Deploy's VARI algorithm to get some results but you wouldn't be able to compare results from flight to flight because you have no control

Real-time Drone Mapping with Fieldscanner | DroneDeploy

 

[Sencorp]

Simply pointing out that from your posts it would seem like you're promoting your method as a reliable way to map/collect data, I'm just concerned people will see it and think they can create decent maps without any control or checks, and market themselves only to get caught short. This isn't a pissing contest, I don't care how you do your work because it doesn't affect me, just not wanting to see people get the wrong idea that throwing photos into the cloud processing websites will net reliable results with meaning if they are unaware of what else is required

Can they create an updated ortho, and a fun 3D model to look at? Absolutely. Would I be using it to design from? Not me personally.

You wanted some feedback, I'm giving you my view This post is still very much on track. What is it your company is trying to achieve?

The notion that because of this post, people will suddenly flock to the streets as mappers is a bit hard to imagine. Second, I am not promoting anything other then my first run with a mavic. Nuff said there.
You have pointed out several times in this post that its drone deploy that was used for this mapping. Im positive its something you are more then familiar with. And other then loading it there because I can, there is no claim, support or anything of the sort that would cause anyone to rise up and protect the masses from "Mis-information."

As for your questions of "What is my company trying to achieve?" Been to Italy seven times working with major GIS software producers, have designed UAS system for the past six years, and am the lead tech for BVLOS for CAA northern proving grounds in New Zealand.

That brings us back to the original part of my post that if you have been doing this type of tech, would love to discuss that with you, and to send me a message. The testing proving grounds and the BVLOS project we are working on we have kicked around the idea of hosting some mappers from other countries for a week here to work in the testing grounds for some projects. We have three local UAV manufactures that are local that are coming on board and open to more diversity from other counties.

 

[Sencorp]

I like what you are doing @Sencorp
Would be interested to know how you do it. For example could I use this methodology to create a plant health map of of 20 acre paddock beside my house? (or is it too large an area) I am not interested at this point in ground sampling. More interested in identifying subtle changes in pasture that may not be immediately visible to the naked eye. I have DJI GS Pro. Can I use it to get the images then perhaps process them through something.

Thanks mate. Greatly appreciated. As you can see Bakerboy is very "passionate" about this subject. There are things you can do. As Bakerboy has pointed out that depending on the level of accurate of data you want to collect you will need to do several steps to insure that data is correct. If you want to start to dabble into this field and doing this type of work there are several places to start. Drone Deploy is one place to start. It makes it simple in a sense.
If you wanted to start getting serious about things, moving up to something like menic would be the end goal. Extremely comprehensive platform and used world wide. But I would start off with Drone Deploy.

 

[Logger]

No worries. Thanks for the suggestion. Have installed Drone Deploy on my S7edge and iPad just now. Signed up for a trial account and created a plan over the paddock I am interested in. Actually it is 50 acres not 20! I new this but understated it because I thought it would be too large to consider doing the whole lot. DD says it can scan the whole thing in 12 minutes. Wow! Should be well within a single battery even at 30mph. Cant wait to try it out now. Will :dabble" with this for while and come back an annoy you if I need more advise. Thanks

 

[Sencorp]

No worries. Thanks for the suggestion. Have installed Drone Deploy on my S7edge and iPad just now. Signed up for a trial account and created a plan over the paddock I am interested in. Actually it is 50 acres not 20! I new this but understated it because I thought it would be too large to consider doing the whole lot. DD says it can scan the whole thing in 12 minutes. Wow! Should be well within a single battery even at 30mph. Cant wait to try it out now. Will :dabble" with this for while and come back an annoy you if I need more advise. Thanks

Not a worry.
Attached is the 3D scan. Something I neglected to point out is the automation program that would have flown this mapping was being buggy at the time. I had to manually fly this mapping using camera viewing and manual photo triggering. This was the result. Not the greatest but for manually flying not bad. Been flying less stable aircraft for some time so was easy.

 

[wrassman]

Greetings all.
I wanted to place a few images of the first scans using the Mavic in 3D mapping. I did a mapping of a mates new home over the weekend, and collected the following data;

• Plant health
• Land elevation
• And 3D modeling

I have tons more data other then the three top categories and am in the process of 3D printing the 3D scan of the property. I would love to hear from anyone else out there that is active in mapping now, that would like to be added to the coming collaboratory that my company is developing to be part of a think tank program around this type of tech. If so, please feel free to get in touch with me.

**Just adding an edit to the post with the 3D scan ** Please note this was done with only 36 images and manually flown.

What sofware do you use??

Looking @ drone deploy. But want a free program

 

[dmanisgnarly]

I suggest looking into a ground based RTK to get some good control points. If you seriously want to get into 3D mapping, look into eBee. We are setting our mavic up for NDVI and saving the mapping for the big boys.

 

[dmanisgnarly]

What sofware do you use??

Looking @ drone deploy. But want a free program

Free and good results are not to things you will find together in the GIS community.

 

[Aerial Perceptions]

1496172906_DCLODIOOPENPIPELINE

I did this 3D model of an old abandoned hospital in Guthrie, Ok. After I did the vertical imagery of the property I did two orbits around the building at different heights for the oblique imagery. Probably could have made another orbital pass in-between my two orbits to make the model more precise.

 

[Arthur H H]

Now this has got my attention and from a fellow NZ'der well done. As to the debate on accuracy of measurement, even without high accuracy there is huge scope here.
We have used a company to survey large (20,000 ton) loose product piles for volume and it was within 0.1% of the physical survey results done at the same time. Also at about %1 of the physical survey price.

 

[eltonloh]

For those interested. Here are the low rez 3D prints of the scan.
This can give you a fair good idea of the quality of the scan. This was done using only 36 photos. If I had spent a fair amount of time getting better shots of the trees and fencing it would have been a much more sharp print.
I do find it funny though that it even printed the sheep in the field. Just two,


View attachment 14471 View attachment 14472 View attachment 14473 View attachment 14474

How long does it takes to capture the 3D model. Can the mavic pro battery last for the entire capture in one single run?

 

[Scarab69]

[
This is just too cool. Digging the sheep. Lol

 

[AgroCam]

We have a simple solution for mapping (including NDVI) with Mavic Pro / Mavic 2:
DJI NDVI upgrade | AgroCam