I work for a heavy / highway construction company, we are looking into survey drones for topographic surveying, as well as point references and 3d mapping etc. I bought a dji phantom when they first came out, weve been using a dji vision plus for aerial photos of our jobsites. Consider me a newb for the surveying end of this, but correct me if I'm wrong, but for surveying, don't I just need an auto pilot heavy lifting copter, with a gps camera that I can control from the ground? Then get the correct software(Pix4d)? I am being quoted 50000 dollars and I just think that is ridiculous.
HELP
Replies
You're exactly right. You need a drone capable of collecting GPS tagged imagery and a post-processing software package. Depending on the accuracy you need for the final survey and on your level of access to the area of interest on the ground, you probably will not need RTK GPS on the drone itself. If you already own equipment to collect accurate GPS coordinates on the ground, the accuracy of the coordinates in the air is of less importance.
If your survey areas are larger than about 80 acres, chances are you'll want to go with a fixed wing aircraft to cover it all at once. As John mentioned above, we offer such a drone at Event 38 (full disclosure: I'm the founder of this company).
To post-process the data from your drone, whichever you choose, the two obvious choices are Agisoft Photoscan ($3500) and Pix4d ($8700 or rental programs). These programs will export DEMs and Orthomosaics which you can then analyze using any standard GIS package.
If you have any other questions just send us an email and we'll be happy to help explain the technologies available and why we've chosen the ones we offer. I agree that $50,000 is over the top for what you're trying to do. Maybe they offer some features that you don't really need. If that's the case, we can probably help you out on a more reasonable budget.
3DR is beginning to really branch in this direction.
http://store.3drobotics.com/products/aero-m/
Have you checked to see what they're offering?
-Mike
Make sure you understand the costs beyond the acquisition. Look at the contract of what you are allowed to replace yourself and when you need to send the vehicle back (warranty issues I reckon?) and how much is charged for engineer time (vehicle check). Once you have figured out how this relationship works out, then make the decision (not based on the price of the acquisition alone). Find out what the prices are for spare parts. The ones that break most: prop, parts of the wing, etc).
Sensefly is not the only company out there. Others like precisionhawk offer solutions where you basically get the platform/vehicle, but you must acquire the processing software separately, or you need to download opensource toolsets for ground station control.
They're not based in the US. Consider whether sourcing from a US manufacturer would make you happier in terms of how fast you can ship stuff to them and back, source new components and develop a customer relationship with them.
Also, consider paying a local guy or someone in your neighborhood to put a platform together for you. Make sure they have the skills (have they done a mapper themselves?). The idea is to spend less money and get somewhere quickly without having to learn all the stuff yourself. Let them demo it. If you feel like it's not professional enough for your needs, then upgrade to one of the big boys.
Or... figure out which guy provides these services in your area and invite them for a visit? Maybe after all you don't really want the equipment and the device, but just pay this guy on a frequent basis to perform the mapping for you? One issue here to consider is the risk of owning the equipment... if it fails frequently, you end up paying a lot of money, which perhaps through a service contract with a 3rd party could have reduced the risk more. in other words... knowing how much you pay for the work instead of having a high risk associated with it...
Thanks for the input, I am by no means a surveyor, but the head surveyor in my company says that with the rtk technology, it give very precise locations, are there other more inexpensive routes?
If you use ground control points determined by your own RTK equipment, you put some markers on the ground that you survey along with the land and then specify the exact location of them later in post-processing step. So you'd select a pixel which you recognize as one of the markers you put down and then specify what position that is. This way you get a model and ortho that has the 5cm accuracy, even though the gps on your vehicle may have been 2-3m precise. The disadvantage is that it takes more time and effort: putting the markers down, getting the points from rtk equipment and then specifying in the range of photos for those pixels what the gcp is before post-processing happens.
I have no direct experience with sensefly's equipment, but if there's a bit of latency between recording the position and the photo is actually taken, then you already won't have photos with 5cm precision. If the plane flies at 12 m/s, then 50ms latency would already be 60cm. You'd need to manage latencies down to 4ms precise in order to stick within the 5cm RTK 95% circle. So it would be good to ask them how this rtk ebee compares with taking ground control points in terms of accuracy and the absolute precision (in real world coordinates) of each point of the orthophoto compared to the gcp method.
Ok another question, we are thinking about getting an Ebee RTK, this is a very expensive route, but with the accuracy, could be what we need.
THOUGHTS?
Hi Jim,
Well, those solutions are complete in the sense that you don't need to work on the vehicles themselves and all the configuration, calibration and complexities are done for you. I have to say though that 50k is only the acquisition, during the use you'll get some rough landings, which require you to spend more money on maintenance. To give you an idea, some manufacturers require you to mail the entire vehicle back to them for their engineers to have a look or replace wings and other stuff. Any operation like that is the cost of shipment + $1000 for the engineer "review"/ check + high prices above the market for replacement parts (I've seen quotes of $100 for a propeller).
Anyway, what the other guys said is your alternative, but you'll need to spend more time figuring out how things work and should be configured and calibrated. There are now wikis full of that info, but it will take time to get started. Mistakes like incorrect CG then become possible that could cause damages when you get started.
What do you call "accurate" by the way? With GPS's like this or of those cameras, you need to think 3 meters precision. However, as you take many more photos of the same area, the software correlates them together and the final error you should count on is anything between 1 meter and 20cm, depending on how the photos were taken. The precision of the model itself however is mostly determined by the GSD. If you do use some high accuracy ground points (RTK?), you can get to an absolute precision (planimetric accuracy?) around 5cm or slightly lower. Given the pixel size is usually around 2-3cm on the ground, you wouldn't get better accuracy than that.
If the areas you're measuring are around 100x40m or sizes like that, the use of a multirotor is much easier. You can get the cheaper ones that fly up to 15 minutes or get a long endurance hexa of some sorts that flies up to 35 minutes. The bigger hexa can map areas up to 400x400m. Planes cover areas like that easily though and if you get them cheap enough (skywalker, even a bixler?), you just replace the frame every so often.
this was generated with pix4d trial, and a phantom vision plus
2.pdf
Dear Jim, You dont need to spend $50k, but rather less than $3,000 for a solution. Here is how one DIY member is using a less expensive solution by www.event38.com
Hope this helps a little.
I am correct in thinking my major need is a gps camera with a good resolution? I need to be able to get accurate coordinates.
If I build heavy load octocopter, get a good gimbal, a gps camera, good autopilot software, set the camera to take pictures at a certain interval, isn't that the meat and bones of the operation?