Activity Feed › Discussion Forums › Photogrammetry, LiDAR & UAS › Drones the future?
Drones the future?
Posted by Adam P on September 28, 2017 at 8:28 pmHas anyone been using drones lately? How does it work for surveying, is it accurate? Are the process complicated at all? My boss has been thinking that within the next 5 years, we??ll lose a lot of business if we don??t have our own drones.
chris-mills replied 5 years, 6 months ago 25 Members · 71 Replies- 71 Replies
Depends on what method you approach it with; photogrammetry is great for bare material stockpiles and volumes, but it doesn’t take much brush or ground cover to skew the results. LiDAR is great for measuring through foliage, but has trouble picking up really tight resolutions, and in some situations needs to supplemented with photogrammetry. There’s a massive price gap between the two to be considered, as well as the necessity of a drone pilot’s license obtained through the FAA.
As far as boundary is concerned, I’m not particularly sure that I would ever trust aerial measurements against on-the-ground measurements, so I think that aspect of the field is still relatively safe for the time being.
I guess what it really comes down to is what your company’s primary focus is; if it’s mainly boundary and title work, I don’t think you have anything to worry about. If it’s topo, then yeah, you might begin to lose business within the next few years.
Adam P, post: 448681, member: 10824 wrote: Has anyone been using drones lately? How does it work for surveying, is it accurate? Are the process complicated at all? My boss has been thinking that within the next 5 years, we??ll lose a lot of business if we don??t have our own drones.
If you can use photogrammetry to enhance your current deliverables, then definitely go for it. The barriers to entry are so low, in both cost and training, that there’s no reason to sit on the sidelines. In many places, drones are the present.
We use them, and the data is accurate as long as you have quality products and understand how to use them. Do your homework before you buy, there’s a lot out there. And you really need to sit down and figure out how you envision using it, because determining your needs will drive a lot of the decision making process.
I am looking at possibly contracting out a drone survey instead of a traditional survey due to accessibility/safety issues.
I think it is definitely a technology that will be broadly used by most surveyors. Will the surveys be cheaper for the clients?…Probably not.A friend was showing off his drone.
Set it in the front yard, then went back into his house. The drone took off, flew a grid and returned the point of departure in 10 minutes with stereo photography, gps photo centers, imu orientation. He dumped the SD card into his computer and with little more than setting up a directory structure, and defaults, the software generated: orthos, 3d wire frame, dem, point cloud, (contours optional, break lines would have be traced).
I asked: what did you use for control? That’s when it got impressive. He had 6 targets, laser printed paper, circular bar code targets. The software scanned the images and automatically identified the 3D control, auto correlated and bundle adjustment without operator assistance. (On a good day, but he had the process down pat.)
8-10 lot subdivision map in an hour. As a base map, it was good for 1:2400.
The control of course were surveyed found corners, nothing more.
Yes, it’s coming.
Larry Scott, post: 448702, member: 8766 wrote: A friend was showing off his drone.
Set it in the front yard, then went back into his house. The drone took off, flew a grid and returned the point of departure in 10 minutes with stereo photography, gps photo centers, imu orientation. He dumped the SD card into his computer and with little more than setting up a directory structure, and defaults, the software generated: orthos, 3d wire frame, (contours optional, break lines would have be traced).
I asked: what did you use for control? That’s when it got impressive. He had 6 targets, laser printed paper, circular bar code targets. The software scanned the images and automatically identified the 3D control, auto correlated and bundle adjustment without operator assistance.
8-10 lot subdivision map in an hour. As a base map, it was good for 1:2400. The control of course were surveyed found corners.
Yes, it’s coming.
And of course traditional QC shots to verify the accuracy statistics the software is spitting out?
imaudigger, post: 448703, member: 7286 wrote: And of course traditional QC shots to verify the accuracy statistics the software is spitting out?
As much QC as the purpose requires. More control is always a plus. Blind control (de-weighted barcode targets and check the residuals) too. Post processed gps/imu blended solution and gyro stabilized camera.
No different than aerial mapping a county at 8,000 feet in a Navajo. Just a small area at 100 feet. All the same software and photogrammetric rules.
Nobody would map a few lots by conventional aerial. But the drones are packing heat for farm surveys.
Speaking of farm surveys…I actually am very interested in how drones are being used to optimize crop yields.
The example I watched showed how they can create a fertilizer grid based on georeferenced aerial photos.
The equipment has GPS and regulates the fertilizer quantities based on the attributes of that grid.I think they also used infrared sensors.Pretty amazing stuff.
I know they have realized higher crop yields simply by planting row crops in straight, accurately spaced rows.
Larry Scott, post: 448702, member: 8766 wrote: A friend was showing off his drone.
Set it in the front yard, then went back into his house. The drone took off, flew a grid and returned the point of departure in 10 minutes with stereo photography, gps photo centers, imu orientation. He dumped the SD card into his computer and with little more than setting up a directory structure, and defaults, the software generated: orthos, 3d wire frame, dem, point cloud, (contours optional, break lines would have be traced).
I asked: what did you use for control? That’s when it got impressive. He had 6 targets, laser printed paper, circular bar code targets. The software scanned the images and automatically identified the 3D control, auto correlated and bundle adjustment without operator assistance. (On a good day, but he had the process down pat.)
8-10 lot subdivision map in an hour. As a base map, it was good for 1:2400.
The control of course were surveyed found corners, nothing more.
Yes, it’s coming.
Was this Pix4d, Agisoft Photoscan or ?
FrozenNorth, post: 448710, member: 10219 wrote: Was this Pix4d, Agisoft Photoscan or ?
Agisoft photoscan.
He reports pix4d is even better.He’s run several. And precision Ag, gps/imu camera, chem sprayer, is off-the-shelf.
But site maps of small area, drone-borne photo and lidar could be the next big thing. And more and more off-the-shelf.
I noticed a long time ago my Navcom Starfire realtime gps (satellite linked) was a John Deere company.
How does a drone set a corner? or dig up a section corner…or locate a blazed or fenced line?
Mike Mac, post: 448713, member: 2901 wrote: How does a drone set a corner? or dig up a section corner…or locate a blazed or fenced line?
[SARCASM]Is someone trying to do that with a drone? Will pay to watch.[/SARCASM]
Mike Mac, post: 448713, member: 2901 wrote: How does a drone set a corner? or dig up a section corner…or locate a blazed or fenced line?
Find a couple corners, make a map, at least you’ll know better where to look. And spot fence line too.
It’s an augmentation, enhancement, and now cost effective. Not an end-all-do-all. Maybe one of those services that can be worthwhile – in-house service.
Larry Scott, post: 448711, member: 8766 wrote: But site maps of small area, drone-borne photo and lidar could be the next big thing. And more and more off-the-shelf.
Is there such a LiDAR unit available now for something like the DJI Phantom 4, which is getting a good rap here at the moment, or is it still limited to bigger drones?
Drones are extremely useful, for now. They work for a lot of workflows with very little effort. However, in 5 years, I expect drones to be falling out of favor due to satellite imagery and LiDAR. Smart cities are coming and with them comes data..all of the data you’ll need.
Raybies, post: 448720, member: 9029 wrote: Drones are extremely useful, for now. They work for a lot of workflows with very little effort. However, in 5 years, I expect drones to be falling out of favor due to satellite imagery and LiDAR. Smart cities are coming and with them comes data..all of the data you’ll need.
Satellite imagery won’t meet 1:2400 or larger. And timely, and dem.
You don’t think so? DigitalGlobe is already offering 18 hour turnaround for some data products with most in the 2-5 day window.
I’ll admit that resolution and accuracy arent quite there, but in 5 years? I think it’ll be a different ballgame. That’s why Trimble bowed out of drones.
Raybies, post: 448724, member: 9029 wrote: You don’t think so? DigitalGlobe is already offering 18 hour turnaround for some data products with most in the 2-5 day window.
I’ll admit that resolution and accuracy arent quite there, but in 5 years? I think it’ll be a different ballgame. That’s why Trimble bowed out of drones.
You could take a picture of your girlfriend with a 2500 mm telephoto lens at a hundred yards. Compare that a photo of your girlfriend with 100 mm lens 4 feet away.
Larry Scott, post: 448702, member: 8766 wrote: I asked: what did you use for control? That’s when it got impressive. He had 6 targets, laser printed paper, circular bar code targets. The software scanned the images and automatically identified the 3D control, auto correlated and bundle adjustment without operator assistance. (On a good day, but he had the process down pat.)
This is fine for small areas – a drone flying a house at 100′. If you are flying large areas at a greater height the size of barcode target needed becomes unwieldy.and you need a lot of flat ground to set them on. You also need to be sure that the targets will stay in place, so you can’t use road markings, etc. It’s then probably easier to manually identify the targets on the photos. Once you’ve picked one in Photoscan it puts up all the relevant photos so you can check each quite quickly.
We flew 150 ha. of quarry and landfill last Tuesday, 26 ground targets at 0.4m square and 6 road markings.12 hours total site work. This was an extended site where we had previously done ground survey, so the untouched areas gave a good ground check – typically +/- 50mm (during the setting of the ground targets we took additional GPS points in areas of vegetation to check the accuracy of the filtering and to provide “area” corrections for the undergrowth). When the site was first surveyed it took about 10 days on the ground..
Boundaries – fine for general topo, but not suitable for exact delineation. Even when fences go under tree cover you can generally get quite a good line from little glimpses through the canopy. As with all methods, you need to be selective. If you are setting ground control then the GPS can pick up additional information in the areas where you know the vegetation will obscure important points.
Log in to reply.