Minimum Overlap

[aerialimagery]

I'm going to be photographing some landfills ranging in size from about 50 to 100 acres as a test for a potential long-term client, and I billed out the job very cheap so that I can demonstrate what's possible. I plan to fly them at 400' to keep the imagery and processing costs down. What would be the minimum overlap that I need? (I will be processing through Maps Made Easy).

I have done a few test maps with MME and they all came out a little bit too small, and they said I might need more overlap. I was hoping I could do these at 60%. Is that sufficient?

 

[Jesse G]

If you're doing 60% front/side you're going to have problems. 60% side overlap and 80% front overlap is the minimum for 3d mapping if you're only doing a 2d map you could get away with 75% front but I would recommend upping the side to at least if 70% to get a decent looking orthomosaic...Personally I use 80% side and 85% front overlap minimum when I need a 3d model with survey grade accuracy. What drone platform are you using and what resolution are you trying to achieve?

 

[aerialimagery]

I'm using a P4P and Maps Made Easy. What I'm most concerned with is making reasonably accurate measurements including volumes. The previous maps I made at 60% looked ok, but when overlaid over google maps they looked a little smaller and at MME they said I need more overlap to correct that. Since I don't have a lot to budget for processing costs and am not concerned about a high level of resolution, what would be the minimum overlap needed just to keep the map from shrinking?

 

[Daniel]

With our P4 pro for 3d modelling we use 90% front and about 85% side overlap and fly at about 60 metres above ground.
We always fly an area at least 30 metres greater perimeter than the actual job site.

Your shrinking issue may be that you have not flown the perimeter area far enough outside your work site.
If resolution is not critical then the overlap doesn't matter that much but you need to overlap all of your work site in order to model all of your work site.

 

[Jesse G]

Daniel is right on point here for the most accurate 3D results but it doesn't sound like you want to go all out here. One thing you didn't mention is if you will be using GCPs on these projects. If you're not using GCPs than scaling can be off on your maps which could add to the "shrinking" you describe. as far as accuracy, If you fly a P4 pro at 400' AGL your resolution will be around 1.3" per pixel which is kind of large to get good 3d data, add that to the small overlap and you're results will be very poor. I would try to stay around the 1" per pixel area which is about 300' AGL. I would up your overlap to at least 80/85. This would put your image count around 800-900 for 50 acres(if you upped it to 85/90 that number would at least double). That should get you down to within a few inches if you use well placed GCPs which I would consider "reasonably accurate". However, having worked directly with landfills in my area they required all 3d models to be within their own local coordinate system, or a state plane coordinate system and for that you would absolutely need to use GCPs. Keep in mind that without GCPs your data wont be able to get anywhere near a few inches, I've heard lots of pilots talking about maps flow without GCPs are within "5% to 15%" accurate to maps with GCPs which I have found to be incorrect. the GPS in your drone is only accurate to within a few feet horizontally and worse vertically so your map can only be as accurate as that without GCPs.

 

[Daniel]

Jesse, you make good points.

However if the shrinkage being described is noticeable on a map then the "shrinkage" is not due to scaling...

Both online mapping and a drone originated model map are on GNSS grid co-ordinates and have aligned mapping scales.
Furthermore the combined scale factors for sea level and grid to ground corrections would not be noticeable visually as they are in the order of 10 mm per 100 metres (3/8" per 100 feet using your terminology).

As you state, the inherent errors with using a standalone GNSS system for positioning without GCP are far more dangerous but that also will not cause any shrinking effect.

I suspect that the shrinkage being described is simply a smaller overall map area compared to the apparent area planned to be flown due to not extending the flown area far enough in the first place. As with any surveying technique, drone mapping requires working from outside in. Surveying, as with other mathematical issue must include redundancies.

Map a larger area and then concentrate on the area of interest within that area.

 

[ArrUnTuS]

At a minimum, use a scale on the ground to reduce the error of those GPS. The Hdop between 1.5 and 2 meters and the Vdop is 1.5 times the Hdop........

Anyway, don't worry so much about getting an error when you overlay it to Google maps, almost certainly your models will have more precision. Please note that these maps are obtained with satellite images. It is more important that you can give accurate data to the customer depending on what he asks.

75% frontal and 60% side overlap in general cases is the minimum recommended.

 

[aerialimagery]

Thanks for all the input and advice.

Jesse, what has been your experience regarding the frequency of landfills needing to be mapped? Also, would you mind sharing a ballpark figure of what you charge or what you suggest should be charged, considering the job is done thoroughly as you suggested?

How do you suggest obtaining ground control points and how many are needed?

 

[Jesse G]

Hey sorry this thread has taken me so long to get back to...I'll try to give you as much info as possible. The landfills that I have been involved in monitoring perform their inspections either once a year, or quarterly depending on the type. Full disclosure I am a land surveyor and have been involved with monitoring landfills long before I was a certified remote pilot so most of my work with them has been outside the UAS world. Which makes it hars to give you any help in ths pricing department, however mapping about 50 acres for full 3d surface generation and volume calculations if done conventionally with a survey crew and a manned flight(which is how they've always done it) runs them between 10-15k depending on the type of monitoring they are doing. A single manned flight starts at around 5k in my area so coming in around 2500 and up isn't outlandish if you can generate results that rival conventional methods. Which I may add is harder than just doing a single grid flight with a few GCPs.
For GCP placement...I use a topcon GR3 RTK system for placing GCPs since I am a surveyor and already own the equipment. There are some modestly priced options out there loje the emlid reach systems that I have heard produce great results, as well as the propeller aeropoints which are stand alone targets with gps receivers built in so all you do is push a button to turn them on and let them stay on the same place for 90 minutes and process the point locations later. As for the locations that they should be placed...this changes at every location depending on the topography, you should have a good grid of GCPs and then additional GCPs at edges of steep inclines or atvthe bottom of pits as well as try to have GCPs at your highest and lowest points of the project. At minimum for a 50 acres site I would set about 15-20 GCPs so I have at least 10 to use for control and at least 5 for verification.

 

[Florida Drone Supply]

Hey sorry this thread has taken me so long to get back to...I'll try to give you as much info as possible. The landfills that I have been involved in monitoring perform their inspections either once a year, or quarterly depending on the type. Full disclosure I am a land surveyor and have been involved with monitoring landfills long before I was a certified remote pilot so most of my work with them has been outside the UAS world. Which makes it hars to give you any help in ths pricing department, however mapping about 50 acres for full 3d surface generation and volume calculations if done conventionally with a survey crew and a manned flight(which is how they've always done it) runs them between 10-15k depending on the type of monitoring they are doing. A single manned flight starts at around 5k in my area so coming in around 2500 and up isn't outlandish if you can generate results that rival conventional methods. Which I may add is harder than just doing a single grid flight with a few GCPs.
For GCP placement...I use a topcon GR3 RTK system for placing GCPs since I am a surveyor and already own the equipment. There are some modestly priced options out there loje the emlid reach systems that I have heard produce great results, as well as the propeller aeropoints which are stand alone targets with gps receivers built in so all you do is push a button to turn them on and let them stay on the same place for 90 minutes and process the point locations later. As for the locations that they should be placed...this changes at every location depending on the topography, you should have a good grid of GCPs and then additional GCPs at edges of steep inclines or atvthe bottom of pits as well as try to have GCPs at your highest and lowest points of the project. At minimum for a 50 acres site I would set about 15-20 GCPs so I have at least 10 to use for control and at least 5 for verification.

Great information @Jesse G. We use the Aeropoints which have been producing pretty good results for us and our clients. We have tested them for several months and feel comfortable enough that we are considering stocking them. They are sold in sets of 10 so not quite as many as you are using - but do a decent job anyway if they are placed in locations where they are seen and photographed many times. Ground control points - whatever brand a pilot chooses - add a significant increase to accuracy and are worth the expense for someone looking to do a lot of this type of work.

 

[Jesse G]

Also I forgot to mention that another way to place GCPs is to build a relationship with a local surveyor...many surveyors have no idea the amount of time drones can save them and often a simple proving mission done for free or very cheap to show them proof of concept could be worth dividends in the future.

 

[Florida Drone Supply]

Also I forgot to mention that another way to place GCPs is to build a relationship with a local surveyor...many surveyors have no idea the amount of time drones can save them and often a simple proving mission done for free or very cheap to show them proof of concept could be worth dividends in the future.

Agreed - lots of opportunities to work together on projects.

 

[ArrUnTuS]

Hey sorry this thread has taken me so long to get back to...I'll try to give you as much info as possible. The landfills that I have been involved in monitoring perform their inspections either once a year, or quarterly depending on the type. Full disclosure I am a land surveyor and have been involved with monitoring landfills long before I was a certified remote pilot so most of my work with them has been outside the UAS world. Which makes it hars to give you any help in ths pricing department, however mapping about 50 acres for full 3d surface generation and volume calculations if done conventionally with a survey crew and a manned flight(which is how they've always done it) runs them between 10-15k depending on the type of monitoring they are doing. A single manned flight starts at around 5k in my area so coming in around 2500 and up isn't outlandish if you can generate results that rival conventional methods. Which I may add is harder than just doing a single grid flight with a few GCPs.
For GCP placement...I use a topcon GR3 RTK system for placing GCPs since I am a surveyor and already own the equipment. There are some modestly priced options out there loje the emlid reach systems that I have heard produce great results, as well as the propeller aeropoints which are stand alone targets with gps receivers built in so all you do is push a button to turn them on and let them stay on the same place for 90 minutes and process the point locations later. As for the locations that they should be placed...this changes at every location depending on the topography, you should have a good grid of GCPs and then additional GCPs at edges of steep inclines or atvthe bottom of pits as well as try to have GCPs at your highest and lowest points of the project. At minimum for a 50 acres site I would set about 15-20 GCPs so I have at least 10 to use for control and at least 5 for verification.

Topographers are the most likely to benefit from drones. One week's work taking points can be done in one hour. Of course, GPC's have to be taken.

Yesterday I was watching a video of how the Aeropoints work and I was disappointed that it took them up to 2 hours to calculate their coordinates accurately. Many jobs are relatively small spaces, 2 to 5 hectares and the flight is done in a matter of minutes. Having to wait two hours, I don't know, it doesn't catch my eye. The price doesn't call me either, $6000 for 10 Aeropoints and if you buy 20 it's $10,000. I see them expensive.

On the other hand, the use of RTK or PPK systems in drones is a better solution. Emlid and Drotek offer relatively inexpensive RTK L1 solutions that give you instant information with 2 to 3cm accuracy and we're talking about $1,000.

If you are thinking of taking the DJI M210 RTK, you should take a good look at it. It should not be a real RTK system and should not have the same accuracy. That's at least I've read over it.

We have different solutions at our disposal, it's a matter of studying them all well

 

[Dave Pitman]

Also I forgot to mention that another way to place GCPs is to build a relationship with a local surveyor...many surveyors have no idea the amount of time drones can save them and often a simple proving mission done for free or very cheap to show them proof of concept could be worth dividends in the future.

When i started getting experience with photogrammetry a couple of years ago, I reached out to several surveyors in my rural location. Only one even returned my call and invited me in for a meeting. Turns out he was not that interested in the technology but more interested in lecturing me about not offering surveying services without a license. He was pleasant enough, but my attempts to do exactly what you mention were fruitless. I suggested that he have me come to a site they were already surveying and I would set air targets on their markers and provide him with the resulting orthoimage. I suggested he give me the surveyed location of 3/4 of the targets and reserve the rest to prove the ortho. I offered this for free for him to evaluate the technology and workflow.

He wasn't interested and I was disappointed. I hold out hope that some day, one of these guys will become interested.

 

[ArrUnTuS]

To be threatened by a technology that makes her job so much easier........ ........ Doesn't make much sense. He could still get the same paycheck and work a lot less and give you job. I hope that person has rethought over time.

 

[rvrrat14]

Jesse, you make good points.

However if the shrinkage being described is noticeable on a map then the "shrinkage" is not due to scaling...

Both online mapping and a drone originated model map are on GNSS grid co-ordinates and have aligned mapping scales.
Furthermore the combined scale factors for sea level and grid to ground corrections would not be noticeable visually as they are in the order of 10 mm per 100 metres (3/8" per 100 feet using your terminology).

As you state, the inherent errors with using a standalone GNSS system for positioning without GCP are far more dangerous but that also will not cause any shrinking effect.

I suspect that the shrinkage being described is simply a smaller overall map area compared to the apparent area planned to be flown due to not extending the flown area far enough in the first place. As with any surveying technique, drone mapping requires working from outside in. Surveying, as with other mathematical issue must include redundancies.

Map a larger area and then concentrate on the area of interest within that area.

I would think the shrinkage, which I’ve encountered too, may be due to the images processed not being orthorectified. Meaning they are not ‘draped’ to fit the true features on the ground to arrive at a flat surface in the photo. They are simply georeferenced. I know when I used to do this in engineering work, we had known scale factors in a given area to use to arrive at grid coordinates for mapping/design vs true survey coordinates. Design applications assume a ‘flat piece of paper’.

 

[Mike Nevins]

I'm going to be photographing some landfills ranging in size from about 50 to 100 acres as a test for a potential long-term client, and I billed out the job very cheap so that I can demonstrate what's possible. I plan to fly them at 400' to keep the imagery and processing costs down. What would be the minimum overlap that I need? (I will be processing through Maps Made Easy).

I have done a few test maps with MME and they all came out a little bit too small, and they said I might need more overlap. I was hoping I could do these at 60%. Is that sufficient?

I am curious, just what are the deliverables for this landfill job. After reading all the posts, you seem to already be connected (as a surveyer) to them. This is a new area for me, so I would like to know why they map these areas? Do they need to be 3D? Sorry for my ignorance.

 

[UAV_Mapper]

it took them up to 2 hours

We demoed a set and got ~0.05' accuracies within 45 minutes, so the 2 hour mark is very conservative. It's also going to depend on your baseline to the nearest CORS stations, if you're using that method. By the time we laid them out, got setup to fly, and got ready to pick them up, there wasn't much waiting involved. Overall, we were very pleased with them and they can be extremely helpful for those without a full blown survey setup and software to process the data.

 

[ArrUnTuS]

We demoed a set and got ~0.05' accuracies within 45 minutes, so the 2 hour mark is very conservative. It's also going to depend on your baseline to the nearest CORS stations, if you're using that method. By the time we laid them out, got setup to fly, and got ready to pick them up, there wasn't much waiting involved. Overall, we were very pleased with them and they can be extremely helpful for those without a full blown survey setup and software to process the data.

I took it from a video demonstration of the AeroPoints people, it's not my own experience. 45 minutes is at the limit of what I consider an acceptable time. Many times, carrying the scheduled mission from the office a flight can take less time. In my opinion, I think there are other better options.

 

[UAV_Mapper]

I think there are other better options

What are you using for your GCPs and checkshots that are that quick? There's no post-processing beyond uploading the data and opening a website. If you've got a workflow that quick than I think you're going to put us all out of business.