Phantom 4 pro and gravel pile mapping

[jasonjlee]

Just a side question, but has anyone used Drone Deploy to map a large area of land and make some money from doing it? I downloaded the app on my phone, but I don't see how I can use it like I do with the DroneBase app to fly and make some profit.

 

[Dave Pitman]

I have never used, the truth is that since PPK corrections are post-flight, the images are geotagged again with post-flight corrections?

That is correct. The dji drone flies and captures images like normal. But the on-board gnss receiver also logs the position of the drone when the image is captured, only much more precisely.

Back at the office, the log file from the gnss receiver is processed with CORS logs to correct the positions for atmospheric anomalies. This is Post Process Kinematic (PPK) as opposed to Real Time Kinematic (RTK) which corrects the logs of a rover in real time. PPK does not require a live connection to a base or a CORS network making it an ideal tool for UAS work.

Then the corrected log is used to re-write the geo-tags in the images exif data. Then they are processed like normal only this time with centimeter accuracy, in theory anyway.

 

[ArrUnTuS]

That is correct. The dji drone flies and captures images like normal. But the on-board gnss receiver also logs the position of the drone when the image is captured, only much more precisely.

Back at the office, the log file from the gnss receiver is processed with CORS logs to correct the positions for atmospheric anomalies. This is Post Process Kinematic (PPK) as opposed to Real Time Kinematic (RTK) which corrects the logs of a rover in real time. PPK does not require a live connection to a base or a CORS network making it an ideal tool for UAS work.

Then the corrected log is used to re-write the geo-tags in the images exif data. Then they are processed like normal only this time with centimeter accuracy, in theory anyway.

Thank you

Let me see if I've understood how the two systems work. When I investigated it, I understood it as follows. Please correct me if I say anything wrong.

RTK: Do you have a GPS base, next to the ground station and which connects to a data center, CORS network? and get the position of the GPS base. The GPS base then communicates with the Rober GPS that is on the aircraft and gives it real time corrections and proceeds to save the coordinates in the images.

PPK: GPS base that takes data from its position. GPS rober that stores the flight data in the aircraft itself. There is no communication between the two. Once the flight is complete, the information of both GPS is combined and corrections are made, and the coordinates are stored in the images.

Is it also possible to combine both methods? RTK on the flight and then in the office PPK to finish leaving it as accurate as possible?

 

[Dave Pitman]

You are close.

With RTK, you have a few options.

1. This is the most common for pro surveyors when they are within a network. You have a base unit that is stationary and receiving CORS network corrections. It then sends corrections to a rover unit which doesn't have to spend much time at a spot to provide a very accurate position.
2. You set up a base unit on a Known Point, basically a marker that you or some other entity has established as very accurate. You specify the location of the point into the base unit. The base unit compares the position it is getting from the satellites with the position that you specified. It calculates the difference and that is the correction it sends to the rover. The longer the base sits static, the more accurate the correction calculation that is sent to the rover becomes because it has more correction values to average.
3. You can also have a rover that connects directly to a CORS network for corrections without the base unit.

The pros of RTK are that the surveyor can quickly (relatively) set out some points on the ground that are very accurate. Either a CORS network or a Known Point from which to start on are required.

With PPK, there are also different options.

1. If you have a Known Point, you can set up the base unit on the known point, specify the position, and the base will begin logging in order to establish a correction. But unlike RTK, this correction is not sent to a rover, it is just logged. The rover unit logs it's position (at the time if image capture in the case of a UAV) for the session. At the end of the session, you use software to take the corrections recorded by the base and use them to correct the positions recorded by the rover using the exact time from each.
2. If you have CORS logs for a station within range available, you don't need the base unit at all. The rover logs in the same manner as #1, but there is no base on site. After the session, you upload you rover's logs to the CORS provider. You select the station that is closest to your project and the time frame of your logs and they will adjust the positions collected by the rover with the corrections from the CORS station you selected.

It's all pretty complicated. But once you get your head around it, it make perfect sense.

 

[ArrUnTuS]

You are close.

With RTK, you have a few options.

1. This is the most common for pro surveyors when they are within a network. You have a base unit that is stationary and receiving CORS network corrections. It then sends corrections to a rover unit which doesn't have to spend much time at a spot to provide a very accurate position.
2. You set up a base unit on a Known Point, basically a marker that you or some other entity has established as very accurate. You specify the location of the point into the base unit. The base unit compares the position it is getting from the satellites with the position that you specified. It calculates the difference and that is the correction it sends to the rover. The longer the base sits static, the more accurate the correction calculation that is sent to the rover becomes because it has more correction values to average.
3. You can also have a rover that connects directly to a CORS network for corrections without the base unit.

The pros of RTK are that the surveyor can quickly (relatively) set out some points on the ground that are very accurate. Either a CORS network or a Known Point from which to start on are required.

With PPK, there are also different options.

1. If you have a Known Point, you can set up the base unit on the known point, specify the position, and the base will begin logging in order to establish a correction. But unlike RTK, this correction is not sent to a rover, it is just logged. The rover unit logs it's position (at the time if image capture in the case of a UAV) for the session. At the end of the session, you use software to take the corrections recorded by the base and use them to correct the positions recorded by the rover using the exact time from each.
2. If you have CORS logs for a station within range available, you don't need the base unit at all. The rover logs in the same manner as #1, but there is no base on site. After the session, you upload you rover's logs to the CORS provider. You select the station that is closest to your project and the time frame of your logs and they will adjust the positions collected by the rover with the corrections from the CORS station you selected.

It's all pretty complicated. But once you get your head around it, it make perfect sense.

Okay, now I remember why I chose the RTK system. To not need a total station to establish any points. Although what you explain in PPK doesn't have to be necessary either. That second point had escaped me.

In principle to make topographic surveys are not usually, except on roads, too large to lose the signal between the base and the rober, of course always in VLOS.

What I will have to look at now is according to models, what is the signal range of the different models between the rober and the base to determine which would be the best option.

Thank you so much

 

[Dave Pitman]

I'm happy it was helpful. I am pretty new to this technology but I'm learning.

 

[ArrUnTuS]

In this case I have neither total station nor knowledge of use, although I have just been told that the equipment is rented, eliminating the need to have a topographer that takes at least the coordinate of the base is an advantage and an added value for the customer. With these new systems you get between 2 and 3 cm that are valid in most civil works projects.

When researching I found out that I have good coverage of CORS networks in the whole area where I am going to work. In case any mountain shades the signal or something like that, I'll have to choose a high place that also serves as a takeoff base for the drone.

All with the intention of not complicating the transaction for the client. Although it is true that in this sector, the customer works with total stations and I don't think there was any problem

 

[Jesse G]

You are close.

With RTK, you have a few options.

1. This is the most common for pro surveyors when they are within a network. You have a base unit that is stationary and receiving CORS network corrections. It then sends corrections to a rover unit which doesn't have to spend much time at a spot to provide a very accurate position.
2. You set up a base unit on a Known Point, basically a marker that you or some other entity has established as very accurate. You specify the location of the point into the base unit. The base unit compares the position it is getting from the satellites with the position that you specified. It calculates the difference and that is the correction it sends to the rover. The longer the base sits static, the more accurate the correction calculation that is sent to the rover becomes because it has more correction values to average.
3. You can also have a rover that connects directly to a CORS network for corrections without the base unit.

The pros of RTK are that the surveyor can quickly (relatively) set out some points on the ground that are very accurate. Either a CORS network or a Known Point from which to start on are required.

With PPK, there are also different options.

1. If you have a Known Point, you can set up the base unit on the known point, specify the position, and the base will begin logging in order to establish a correction. But unlike RTK, this correction is not sent to a rover, it is just logged. The rover unit logs it's position (at the time if image capture in the case of a UAV) for the session. At the end of the session, you use software to take the corrections recorded by the base and use them to correct the positions recorded by the rover using the exact time from each.
2. If you have CORS logs for a station within range available, you don't need the base unit at all. The rover logs in the same manner as #1, but there is no base on site. After the session, you upload you rover's logs to the CORS provider. You select the station that is closest to your project and the time frame of your logs and they will adjust the positions collected by the rover with the corrections from the CORS station you selected.

It's all pretty complicated. But once you get your head around it, it make perfect sense.

Great explanation of RTK vs PPK here, Dave! I'll add a couple things to the discussion. Having 15 years experience using professional RTK systems for surveying I have come to find most of its strengths and weaknesses. The biggest advantage of having a traditional RTK system is that no known points are actually required. You can set a point and then perform a "static observation"(set youre base station on the point and start recording GPS/GLONASS data) on it and create a known point by processing it against CORS stations after the fact. This is the same way that Aeropoints work, by observing GPS corrections for a given amount of time and then processing it against CORS stations this is also known in my industry as "post processing". Basically the same thing as PPK except only for ground points and not the entire flight. Now I can't say that is the case with the RTK systems being marketed for the drone market as there may not be an option to collect and access the static data on the base station of a drone system.
The biggest downfall of RTK drone systems will be maintaining a "fixed" solution while performing a flight. RTK systems rely on a constant stream of data between the base and rover/drone over radio frequencies, if this stream is broken due to radio interference or some other factor then the coordinates logged for the photos taken are going to be way off until a fixed solution is achieved. Distance also come into play here since most RTK systems have a transmission signal of around a mile unless a high powered radio is used which can push it out another couple miles depending on terrain. This could cause an operator to set up their RTK base in multiple locations to cover a whole site. Some RTK systems utilize mobile network cards so you can have a connection anywhere cell service exists which may or may not make things better. Another factor when using RTK systems is the speed at which the corrections take place. Even when using my high dollar RTK system when collecting coordinates at a constant speed the data tends to lag behind unless you are traveling around 5mph. Meaning that if I am driving with my rover attached to the roof of my truck and I'm traveling 15mph and I store a coordinate, the coordinate stored will actually be from a few feet behind where I told the system to store it and not at the exact place I wanted it. This could obviously cause some problems since flying so slow impacts how long a flight will take and how much ground you can cover. This small issue seems to be unaddressed in the drone sector but perhaps they have found a way around it. PPK systems don't suffer from this since a connection is not required so they seem to be the better choice for consistency with drone mapping. Also when using a PPK system without a base station you take out any chance of human error during that step. Its absolutely critical that the base station be set up level and plum over the known point on a tripod or range pole and then an accurate vertical height of the base station must be taken if on a tripod. I know that seems like an easy thing to do but I have seen it happen many times where something is done incorrectly and the days data is off.
Now one thing I would like to put out there is that no mater if you use RTK or PPK during a flight, without some sort of GCPs on the ground you'll never be able to say or check, without a doubt, how accurate a given map or 3d model is so the need for a traditional RTK system or total station doesn't go away with the purchase said system. The need for GCPs will always be there but with a PPK or RTK system you could go from needing 10 or more to just needing 2 or 3 which on larger sites could save a considerable amount of time.
All that said while there is definitely a place for RTK enabled drones like the m210 for precise flights, I find myself doubting the validity of RTK drone mapping systems and have been looking into PPK options.

 

[ArrUnTuS]

Verification points are usually advisable, at least when you start using the equipment and until you are sure that the data obtained is accurate. That's the way I see it at least, I don't know.

I had not considered the issue of signal range between base and rober in an RTK system. But of course, they are usually based on the drone's telemetry link, so if you lose on the telemetry link you also lose the base link to the rober in the drone. If it is 433 or 915 Mhz is indifferent, the important thing is the range that equipment has.

@Jesse G what I'm worried about is what you say about the speed of the rober. 5mph is a good speed for a multirotor but you can hardly go that slow with a fixed wing. The system I've looked at has an 8hz upgrade. The idea I had was a base and 2 rober. One for the multirotor and one for the fixed wing. Now I have more doubts

 

[fberna]

Hi guys, I've been using my P4P for mapping sites that have gravel piles. Using Drone Deploy we are able to accurately map the sites and calculate the size of each pile. Accuracy is extremely good and is saving our client a lot of expense!!

what post processing are you using? What you get from P4P, or any similar program, is by far not enough for an accurate calculation.

First you absolutely need gcp, or you end up with a sloping end out of scale model.

In most situation the terrain is not just flat but has some sort of slope, so after building the 3D you use other programs in order to accurately build the plane under the pile.
Often you find multiple piles beside each other or on top of each other. Here again the key to a decent job is to build the proper plane under the pile.

It is possible to produce very accurate volume measurements, but it requires quite a bit of work after getting the 3D file from the drone.

 

[Sojourner60]

Y'all are leaving me scratching my head. I'm wanting to learn how to use my P4 to measure stock piles. Drone Deploy and Maps Made Easy both say it can be done. I must be wrong in thinking how it works. I thought you had the drone circle the pile taking pictures. Then the software stiches it together and measures the piles volume measurements in tons or cubic yards.

What an I missing? Is there better software? I'm trying to convince the company I work for to start using drones to measure their stock piles. They have hundreds if them, sand, rock in various sizes, and cinders. Probably can't fly the salt since it's stored in a building.

Thanks in advance from a new by to measuring stock piles.

I installed DD on my phone but it won't run. Sits there loading for a LONG time. Gotta get a better mobile device.

 

[ArrUnTuS]

Y'all are leaving me scratching my head. I'm wanting to learn how to use my P4 to measure stock piles. Drone Deploy and Maps Made Easy both say it can be done. I must be wrong in thinking how it works. I thought you had the drone circle the pile taking pictures. Then the software stiches it together and measures the piles volume measurements in tons or cubic yards.

What an I missing? Is there better software? I'm trying to convince the company I work for to start using drones to measure their stock piles. They have hundreds if them, sand, rock in various sizes, and cinders. Probably can't fly the salt since it's stored in a building.

Thanks in advance from a new by to measuring stock piles.

I installed DD on my phone but it won't run. Sits there loading for a LONG time. Gotta get a better mobile device.

You're not misguided but there's something that makes the data you give to the customer as accurate as possible. You have to give to the program accuracy data to be the result good. It can be done in two ways, using GCP's or scales. Otherwise you will only provide data with the accuracy of the GPS of the drone you use and the P4 has a normal GPS that has a hdop of between 1.5 and 2 meters. The vdop is usually 1.5 times the hdop.

If the data given is based on data from such a large error, it is difficult to give quality data. Using the above mentioned methods of a meter error, an error of a few centimeters is passed to an error of a few centimeters which is considered acceptable today.

 

[Jesse G]

Y'all are leaving me scratching my head. I'm wanting to learn how to use my P4 to measure stock piles. Drone Deploy and Maps Made Easy both say it can be done. I must be wrong in thinking how it works. I thought you had the drone circle the pile taking pictures. Then the software stiches it together and measures the piles volume measurements in tons or cubic yards.

What an I missing? Is there better software? I'm trying to convince the company I work for to start using drones to measure their stock piles. They have hundreds if them, sand, rock in various sizes, and cinders. Probably can't fly the salt since it's stored in a building.

Thanks in advance from a new by to measuring stock piles.

I installed DD on my phone but it won't run. Sits there loading for a LONG time. Gotta get a better mobile device.

As ArrUnTus mentioned the first problem lies in the GPS accuracy of the drones. When you dont use Ground Control Points(GCPs) that are shot in with a RTK GPS system or a total station than the only coordinates you have are the ones that the pictures are tagged with from the drones GPS which as mentioned can fluctuate from 3-6 feet. This wont give you good, consistent results and it makes repeat flights introduce even more error since each new map can be a few feet offset from the previous so thing don't line up.

When you add GCPs it ties everything down to established coordinates so that elevations and horizontal scale is correct as it is in the field and keeps all future flights lined up if GCPs are placed in the same coordinate system.

If your company currently utilizes traditional surveyors to measure stock piles than perhaps you could convince management to get them to set you some control points around the site in areas that won't be disturbed the next time they measure piles and you can start running some tests and compare the results. You should read up on flight planning as its not as simple to get accurate 3d data as dronedepoly makes it seem. Double grid missions of oblique images(images not shot straight down) are needed to really nail down 3d features and the DD app doesn't even have that feature. I believe MME's app does but I'm not familiar with it.

One thing to remember is that no software will be able to give you accurate tonnage since the weight/volume ratio is different for each material. The best you'll get is a cubic yard volume with which you can calculate tonnage if you have a fixed wheight your company uses for each material.

If you need some help with selling points to try to get management on board feel free to reach out as I've helped change more than a few minds. Just be prepared to jump into a big learning curve to get the most accurate data.

 

[Sojourner60]

Y'all are leaving me scratching my head. I'm trying to figure out how to use a drone to measure stock pile. I figured out what CORS is and you explained PPK and RTK. It sounds like it is a lot more complicated than just flying your drone around taking photos and let the software stick it together. Are there any resources I can read up on to get a better understanding of how it works and what is required. Can you use JUST Dronedeply or Maps Made Easy to calculate a stock pile?

 

[HawkView]

Double grid missions of oblique images(images not shot straight down) are needed to really nail down 3d features and the DD app doesn't even have that feature. I believe MME's app does but I'm not familiar with it.

It is hard to keep up with this fast evolving technology. DroneDeploy now has Crosshatch (Double Grid) mission with Oblique support. I have flown them a couple of times and it works well.

 

[Sojourner60]

Hawkview,
Thanks! I will have to go to Dad's page and learn more about it.

 

[TAZimmer]

We routinely map aggregate piles and compute volumes, for both our provincial government as well as local municipalities. We use a P4 Pro, but we've found we definitely need a set of GCPs to meet the accuracy requirements of most of our clients. The GCPs are tied in using RTK with the correction signal coming in via cellular -- works great and very quick/easy to use. Flight planning is via either DroneDeploy or Pix4D Capture. I like Capture better, but to each his own.

The problem we are running in to now is that everyone with a $199 drone from Wal-Mart thinks they can now map gravel piles and compute volumes. Talked to an RM Manager last month; his dad (retired farmer) was using an RC drone to map the RM's gravel piles. He uploads the photos to an online service, and they stitch the ortho and calculated the volumes automatically. Even generates a report that was sent to the RM automatically. The old guys charges like $100 to do a gravel pit -- we can't even get in our truck and drive to the site for that little.

And the RM Manager didn't seem to care that much about the accuracy issue. We're obsessing over +/- 1 cm accuracy, and he's thinking volumes +/- 50 cu.yards is good enough.

 

[ArrUnTuS]

We routinely map aggregate piles and compute volumes, for both our provincial government as well as local municipalities. We use a P4 Pro, but we've found we definitely need a set of GCPs to meet the accuracy requirements of most of our clients. The GCPs are tied in using RTK with the correction signal coming in via cellular -- works great and very quick/easy to use. Flight planning is via either DroneDeploy or Pix4D Capture. I like Capture better, but to each his own.

The problem we are running in to now is that everyone with a $199 drone from Wal-Mart thinks they can now map gravel piles and compute volumes. Talked to an RM Manager last month; his dad (retired farmer) was using an RC drone to map the RM's gravel piles. He uploads the photos to an online service, and they stitch the ortho and calculated the volumes automatically. Even generates a report that was sent to the RM automatically. The old guys charges like $100 to do a gravel pit -- we can't even get in our truck and drive to the site for that little.

And the RM Manager didn't seem to care that much about the accuracy issue. We're obsessing over +/- 1 cm accuracy, and he's thinking volumes +/- 50 cu.yards is good enough.

You've touched on two important points. Online services that are becoming increasingly popular and for very little money process the images and provide you with enough data. It can't be retouched much later, you have to request it but for many applications are enough.

The other point, which we have already mentioned, is precision. Depending on the size of stockpiles or piles of material, the error can vary greatly. It is not the same to have an error of 5 cubic meters in a 50 cubic meter pile than to have an error of 50 cubic meters in a 500 cubic meter pile. We're talking about a lot of trucks of difference. Depending on the situation, the client has to assess it. Anyway throwing away the market price, which occurs in all sectors, is never a good idea.

 

[Dave Pitman]

@TAZimmer Are you setting at least 1 or more gcps on the top of any piles? I've read that for GCPs to be able to dial in z axis on varied terrain (like stock piles) they must be placed dispersed both horizontally and vertically in order to calibrate the difference in elevation in addition to horizontal scale.

I can tell you that I have done some volumes for owners that even after explaining greatly reduced accuracy by not using GCPs have elected to not pay to use them. They said that what they use the data for, "it's close enough". I cannot tell you for what purpose "close enough" and to "what" is though as I have no way to quantify without truthing obviously.

 

[ArrUnTuS]

@TAZimmer Are you setting at least 1 or more gcps on the top of any piles? I've read that for GCPs to be able to dial in z axis on varied terrain (like stock piles) they must be placed dispersed both horizontally and vertically in order to calibrate the difference in elevation in addition to horizontal scale.

I can tell you that I have done some volumes for owners that even after explaining greatly reduced accuracy by not using GCPs have elected to not pay to use them. They said that what they use the data for, "it's close enough". I cannot tell you for what purpose "close enough" and to "what" is though as I have no way to quantify without truthing obviously.

I had never heard of putting GCPs at different heights to try to give more precision to the Z axis.

Can you please indicate any links to that? I find something interesting.