Hello Fellow Yuneec Pilot!
Join our free Yuneec community and remove this annoying banner!
Sign up

520 Vertical Surface Mapping

Joined
May 4, 2017
Messages
21
Reaction score
3
Age
47
Hi
Does anyone have experience using H520 for mapping vertical surface?
We have an upcoming project to survey 400ft high rise building (1 face only).
Q1: would H520 be capable of this mission in CBD environment with regards to possible interference and GPS?
Q2: E50 or E90?
Q3: what is the best software to use for vertical surface ortho map? or would it be more effective to fly manually?
I appreciate your input in advance. Thank you
 
Hi

The H520 does not have the possibility of vertical missions. Hopefully someday it will have that possibility :p

The interferences, I imagine that you refer to zones of shadow of coverage with respect to the satellite positioning, and since it does not have mission for it, in this case it does not influence much. Normally at the time of starting the flight, at ground level is where more problem would exist, it would be a matter of looking where you have more failure before starting the taking of images and determine if the flight will be done in manual or Angle mode (preferably).

What kind of camera to use? Undoubtedly the choice would be the E90 for its higher quality and greater area covered by each image, but it can also be counterproductive if there is not enough space. This is determined by the distance you can separate from the facade.

It is necessary to make calculations by hand, to calculate distances to the objective, depending on the focal distances of each camera, to determine which is the best footprint.

That distance, in theory should change depending on the desired overlap, at least 70%, and thus determine the distance of the next vertical accession path with respect to the previous one. A lot of planning work, I think it's a challenge, I've never done with lack of space, but I think it's very interesting.

Bearing in mind that the error in Z is usually 2.5 times the horizontal one, you should ask the developer of the photogrammetric software you use, the problems that may arise when making the stich. In theory should not be a problem, but we know that the theory often fails if not prepared for it.

These points are the first that I would at least take into account, and from there, move forward.

Is it possible? Yes, it needs a lot more work? Yes, please tell us how it is :D
 
Thank heaps for the advice #arruntuus
We re considering few options at the moment.
There is plenty of space to work with in terms of distance from the façade.
It seems the most obvious option is manual flight grid style, floor by floor without stiching, therefore supplying single images of each floor.
Another issue is the max altitude, approx. 600FT, This takes 21 days period and $800 in fees for CASA to get clearance for class C 400ft + flight.
One more question. Does 520 have front obstacle avoidance to aid with safety?
Thank you for any input
 
Great to know,
Front is all that needed for this mission, only to ensure no contact with the building as the drone will be facing front for the entire flight.
Thanks heaps for your help
 
Thank heaps for the advice #arruntuus
We re considering few options at the moment.
There is plenty of space to work with in terms of distance from the façade.
It seems the most obvious option is manual flight grid style, floor by floor without stiching, therefore supplying single images of each floor.
Another issue is the max altitude, approx. 600FT, This takes 21 days period and $800 in fees for CASA to get clearance for class C 400ft + flight.
One more question. Does 520 have front obstacle avoidance to aid with safety?
Thank you for any input

Having space in front of the façade simplifies the matter a lot.

Regarding the 400 feet of height limit that we have, you must think that is from the home. Do you have another building or a mountain or something that leaves you at least 200 feet above the floor of the building? The limitation is always from the takeoff point, you can fly -200 feet and then +400 feet, and thus get the total height and comply with the regulations without having to ask for more permits.

In my country we cannot increase the height, we can never exceed 400 feet but we can fly to 400 feet of height of the highest object in a radius of 500 feet from the home. So if we have a 300-foot high building less than 500 feet away from the home we can really rise to 700 feet (300+400).

Civil aviation regulations clearly state that any manned vessel in unsegregated airspace must fly at least 500 feet above the highest point in the area. 100 feet are left as a margin of safety and hence we have, in the drones, a ceiling of 400 feet.
 
Concerning vertical grid, I will try it soon, but I thought of it already of course it is done with datapilot.

Create waypoints in a horizontal manual grid: Give heights of say 10m for first line, 20m for second line and so forth,
then and that's the trick, bring all the points of the grid to overlap.
Of course you will only see the waypoints on top of the grid, those under will be hidden but still there.
 
  • Like
Reactions: Mc-Murph
Concerning vertical grid, I will try it soon, but I thought of it already of course it is done with datapilot.

Create waypoints in a horizontal manual grid: Give heights of say 10m for first line, 20m for second line and so forth,
then and that's the trick, bring all the points of the grid to overlap.
Of course you will only see the waypoints on top of the grid, those under will be hidden but still there.

I think it's not that simple, you forget the orientation of the drone or of the camera.
 
it looks ok, but have to be carefull that waypoints in front, stay in front (and not behind otherwise camera turns around)
 

Attachments

  • Capture1.JPG
    Capture1.JPG
    54.2 KB · Views: 30
  • Capture2.JPG
    Capture2.JPG
    31.4 KB · Views: 28
  • Capture3.JPG
    Capture3.JPG
    17.3 KB · Views: 26
That's what I mean, with the controls we have we can't control the direction of the drone, the drone follows the direction of the waypoints. It is necessary to modify the Yaw of the camera so that it always points to the facade. I think it can be done but the work it takes is a lot. It's a nice exercise to try.

The other big problem is to establish the distance, which would be the footprint, to calculate the overlap, etc. Which can also be calculated automatically in a normal flight plan and then change the horizontal distances by vertical, both in distance between passes and in distance between photos..........

Can it be done? No doubt, but........ Once obtained, it would be a repetitive work........ it is better to have a automated process ;)
 
It would certainly be better to have an automated process.

For distance and overlapp see the grid results depending on camera, but since it would be as high as 50 or 70 meters, in a vertical sruface you will rarely get the place to step back that distance. So hand calc or trigger photos every x meters.
 
  • Like
Reactions: fullnoise uav
I'm going to try to simulate a practical exercise. If you come up with ideas to improve it, they will be welcome :D

I have taken as an example a building of a city, trying to follow the example of @ithacapool , but a little more complicated, with the space a little fair.

I used Google Earth to do the horizontal and vertical measurement and added a little margin. 70 meters wide and 65 meters high.

Horizontal measurement.png
Vertical measurement.png

We have the dimensions of the rectangle that will perimeter the mission. We create a polygon with these dimensions and obtain the references to create the polygon in DataPilotPlanner.

Horizontal mission area.png

We measure the distance to the facade, it will be necessary to determine the height of the mission and determine the rest of the parameters.

Distance to the façade.png

We have all the necessary data, let's go to DataPilotPlanner. We create the survey area and determine the height at the distance to the façade.

Horizontal Mission in DataPilotPlanner.png

We already have all the data to create our vertical mission. We save the data from the Trigger Distance, every 4.43 meters horizontally the camera has to take a picture. We count the number of passes the drone has to make to cover the entire area. In this case 12. As we know that the highest point in this area is the building itself that we are going to scan, 65 meters high, we set the height of the RTL to 75m, which in case of problems does not crash into any building.

We proceed to create the definitive mission. Things to keep in mind and where work is boring. I have created 12 passes. I have set the Yaw to the left, in this case, in all Waypoints (that the camera points all the time to the facade). In the first Waypoint I set that I take distance-based photos, every 4.4 meters. It doesn't matter if the lines or past are crooked, we lack the last step.

Temporary final mission.png

We're going to modify what we have and we're going to put all the Waypoints on one side above Waypoint 1 and all the ones on the other side in Waypoint 4. We have to modify the height of all the Waypoints again. How do we calculate the height? We know the height of the building 65 m, we know the number of passes that is 12, we divide and gives us a difference of height of 5.4 meters between passes. The first pass I have established at 3 meters high, I only have to add 5.4 m to the previous pass.

Final mission.png

More things to take into account, so that the points of the corners are exactly in the same horizontal position you can copy the coordinates, for one side copying the coordinates of the first Waypoint in the rest of Waypoints of that side, and with the other the same. With this all the points of one side would be exactly in the same horizontal position, only having to modify the height.

All we need to do is start the mission and see if it works :cool:

P.D.: I'm very tired and I've probably been confused about something or assumed something that is not or I don't know. Let's see if among all of us, we find problems with this method of doing vertical missions.

P.D.2: I'm so tired that I'm not even going to revise what the translator has translated for me, if you understand any nonsense, it could be that the translator has left me bad again, or as my wife says, you only say nonsenses :p:p:p:p
 
  • Like
Reactions: Mc-Murph
I'm going to try to simulate a practical exercise. If you come up with ideas to improve it, they will be welcome :D

I have taken as an example a building of a city, trying to follow the example of @ithacapool , but a little more complicated, with the space a little fair.

I used Google Earth to do the horizontal and vertical measurement and added a little margin. 70 meters wide and 65 meters high.

View attachment 12625
View attachment 12626

We have the dimensions of the rectangle that will perimeter the mission. We create a polygon with these dimensions and obtain the references to create the polygon in DataPilotPlanner.

View attachment 12629

We measure the distance to the facade, it will be necessary to determine the height of the mission and determine the rest of the parameters.

View attachment 12628

We have all the necessary data, let's go to DataPilotPlanner. We create the survey area and determine the height at the distance to the façade.

View attachment 12634

We already have all the data to create our vertical mission. We save the data from the Trigger Distance, every 4.43 meters horizontally the camera has to take a picture. We count the number of passes the drone has to make to cover the entire area. In this case 12. As we know that the highest point in this area is the building itself that we are going to scan, 65 meters high, we set the height of the RTL to 75m, which in case of problems does not crash into any building.

We proceed to create the definitive mission. Things to keep in mind and where work is boring. I have created 12 passes. I have set the Yaw to the left, in this case, in all Waypoints (that the camera points all the time to the facade). In the first Waypoint I set that I take distance-based photos, every 4.4 meters. It doesn't matter if the lines or past are crooked, we lack the last step.

View attachment 12636

We're going to modify what we have and we're going to put all the Waypoints on one side above Waypoint 1 and all the ones on the other side in Waypoint 4. We have to modify the height of all the Waypoints again. How do we calculate the height? We know the height of the building 65 m, we know the number of passes that is 12, we divide and gives us a difference of height of 5.4 meters between passes. The first pass I have established at 3 meters high, I only have to add 5.4 m to the previous pass.

View attachment 12639

More things to take into account, so that the points of the corners are exactly in the same horizontal position you can copy the coordinates, for one side copying the coordinates of the first Waypoint in the rest of Waypoints of that side, and with the other the same. With this all the points of one side would be exactly in the same horizontal position, only having to modify the height.

All we need to do is start the mission and see if it works :cool:

P.D.: I'm very tired and I've probably been confused about something or assumed something that is not or I don't know. Let's see if among all of us, we find problems with this method of doing vertical missions.

P.D.2: I'm so tired that I'm not even going to revise what the translator has translated for me, if you understand any nonsense, it could be that the translator has left me bad again, or as my wife says, you only say nonsenses :p:p:p:p

Your wife is always right. Correct, arruntus?

[emoji41]
 
  • Like
Reactions: arruntus
Hello that looks good and translation worked perfectly as well.

Under waypoint 46 it seems that camera is going the wrong way, see me previous alert : "carefull that waypoints in front, stay in front (and not behind otherwise camera turns around) "
maybe trying a slightly slanted vertical surface works better. Like first pass at 12m distance second pass at 11.5 and so on.

Photogrammetry rules say that you may never have more than twice the distance between farthest and closest point, so if you start at 12m away, your minimum final distance is 6m.

Also taking as many photos as 192 is not that usefull. Your camera takes perfect shots 12m away, no need to get so close.

Bear in mind that windows and their transparency / reflection will cause errors in 3d mesh. Your sample building is exposed to the south, so be carefull at what time you fly, avoid noon time, as you might get the sun exactly into windows reflection, or pick a cloudy day.

Last but not least,
Don't forget to enable OBS avoidance !! And hands on remote at all times ready to get manual control. Have another person to watch distance from building, watch for traffic
 

Attachments

  • setup.png
    setup.png
    1 MB · Views: 18
  • Like
Reactions: arruntus
Hello that looks good and translation worked perfectly as well.

Under waypoint 46 it seems that camera is going the wrong way, see me previous alert : "carefull that waypoints in front, stay in front (and not behind otherwise camera turns around) "
maybe trying a slightly slanted vertical surface works better. Like first pass at 12m distance second pass at 11.5 and so on.

Yes, because the orientation of the camera depends on the orientation of the drone, when moving the points the orientation changes slightly and that can be a problem, maybe applying the coordinate copying system, the problem is corrected, more tests are needed.

This could be avoided with an option so that the drone is not oriented in the direction of the next Waypoint. Keeping the same orientation would simplify things a lot.

Photogrammetry rules say that you may never have more than twice the distance between farthest and closest point, so if you start at 12m away, your minimum final distance is 6m.

In this case, the example is simple, a linear data capture, to complicate it more with all the work that gives the simplest example would be masochistic. :D

Also taking as many photos as 192 is not that usefull. Your camera takes perfect shots 12m away, no need to get so close.

The footprint is determined by the characteristics of the camera and the distance to the object of study or height. In the example you have buildings behind you, you are limited. It also depends a lot on the safety regulations, if you want to fly further back and you have to do it over the road, that forces you to cut the traffic, puffffffffffff not even think about the paperwork and permits that entails. In that sense is an example of complicated flying.

Bear in mind that windows and their transparency / reflection will cause errors in 3d mesh. Your sample building is exposed to the south, so be carefull at what time you fly, avoid noon time, as you might get the sun exactly into windows reflection, or pick a cloudy day.

You are right, the problem of reflections is big, this building does not have much glass but modern buildings or office buildings force you to put a polarized filter.

Last but not least,
Don't forget to enable OBS avoidance !! And hands on remote at all times ready to get manual control. Have another person to watch distance from building, watch for traffic

Indeed, the safety plan for this flight obliges you to take many safety measures. This must always be taken into account.

This is just one example and it is clear that the amount of work it takes to do it by hand is very large. We need vertical flight missions. Yunnec are you around? :p
 
Last edited:
  • Like
Reactions: Sarah
Can’t wait to see your final output.
Will you be using Pix4D to collect the materials?
 
@record.play.rewind I don't know if you mean me or @ithacapool. In my case it is an example, a challenge that I have set myself to see if it is possible to do it. It's not a job.

Continuing with the previous example and given that we have a problem with the orientation of the camera..........................

Taken from the manual

use the K1 to set the pan position of the gimbal camera. When the switch position is down, the gimbal camera is in Global Mode. The pan (Yaw) direction of the gimbal camera will be fixed regardless of the aircraft’s movements.

If we start the flight in the global mode of the yaw and with the camera pointing perperdiculamente to the facade, the camera, regardless of the route that follows the drone is going to point to the facade all the time. Besides, we are not going to have to touch that parameter in all Waypoints, something that simplifies the work :p

P.D.: I forgot, yes, I use Pix4D.
 
@arruntus here’s an interesting 360 workflow. Create your waypoints in google earth “relative to ground” export the file as KML to Pix4D mission hub (if it’s capable of reading KML files.

By using google earth to create your points, you can have a visual and precise measurements.
 
@arruntus here’s an interesting 360 workflow. Create your waypoints in google earth “relative to ground” export the file as KML to Pix4D mission hub (if it’s capable of reading KML files.

By using google earth to create your points, you can have a visual and precise measurements.

Yes, it's possible, but there's a problem. Google Earth's accuracy is very poor. It cannot be used commercially if it is not accompanied by GCP's measured with quality equipment.

Having a gap of between 1 and 1.5 m (the accuracy given by any GPS of a normal drone) is not accepted for calculation by any engineering or surveying company. To make animations, proposals, bids, yes it would be valid. Nowadays, for calculations, and to use it in work they ask you for RTK precision, that is to say precision less than 5cm.

For flight plans ummmmmmmmmmmmm if you do not have objects (buildings, high tension lines, etc) in a space less than 10 meters I would not trust either, you can have an accident. Google Earth is photogrammetry with satellites, we make photogrammetry with drones, much more precision ;)
 
Yes, it's possible, but there's a problem. Google Earth's accuracy is very poor. It cannot be used commercially if it is not accompanied by GCP's measured with quality equipment.

Having a gap of between 1 and 1.5 m (the accuracy given by any GPS of a normal drone) is not accepted for calculation by any engineering or surveying company. To make animations, proposals, bids, yes it would be valid. Nowadays, for calculations, and to use it in work they ask you for RTK precision, that is to say precision less than 5cm.

For flight plans ummmmmmmmmmmmm if you do not have objects (buildings, high tension lines, etc) in a space less than 10 meters I would not trust either, you can have an accident. Google Earth is photogrammetry with satellites, we make photogrammetry with drones, much more precision ;)

Since your flight is experimental, revert to the article that I shared in the past, also google earth is used all the time in our field (or with the surveyors here in the US) when setting preliminary waypoints, and adjust accordingly.

Setting GCP, without any data from ground surveyors is by putting your own makers based on the information gathered from your county recorder's office or the assessor's office. They can provide detailed records and measurements.
 
  • Like
Reactions: arruntus

New Posts

Members online

Forum statistics

Threads
20,955
Messages
241,599
Members
27,285
Latest member
hendrtiz