I am an LS and operate a few drones for survey mapping in Los Angeles.?ÿ My surveys, which are photography based, are typically small in scope and are nearly always based on a local, assumed coordinate system and bench mark.?ÿ I have been using Aerotas to produce the orthophoto and Autocad base map since they can handle local, assumed systems, but would like to get into my own processing of the data.?ÿ
I have noticed that most of the software for processing drone photographic data is on a global geodetic system, such as latitude/longitude or EPSG, etc.?ÿ Does anyone know of a software package that will allow me to use entirely local, assumed coordinates and elevations on my GPC's, and produce an data on that local, assumed system??ÿ My company is small, so price is a consideration.?ÿ Any help would be appreciated.
?ÿ
I guess I wonder why you would not do everything on SPC.?ÿ With VRS and OPUS it is so easy nowadays - even if you just went with a local elevation, or whatever GPS gave you.?ÿ Makes it easy to get back on the system in the future for repeat business, and makes sharing or selling the info much easier.
Pix4d along with other drone processing applications will work with arbitrary coordinates on your GCP.?ÿ But as others has said, it makes no sense not to use SPC when you have GPS. You are making it more difficult on your self and others to tie into your project in the future.
On the subject of GCP, does size/colour/shape etc make any difference to the "quality" of the image processing? On a recent project of surveying some existing bridge sites, our GCP comprised?ÿ 6" circular orange spray-paint blobs, with a small stone temporarily placed in the middle to leave a hole in the spray. The blobs and the hole showed up very clearly on the images but they are small.
Look into Global Mapper. TBC has a decent Photogrammetry package but its not IDEAL, Applanix(trimble owned) was one of the packages we used for fixed wing operations and lots of other Higher dollar software.
USE a really contrasted Photo Identifiable object like a Stop bar, Handicap Parking Space Box Corner, Manhole lids, Anything that is easily seen, and just place your own white panels based on the ASPRS guidelines,?ÿ and can be recaptured or over flown from an Orthagonal process, to give best coverage, and woohoo!?ÿ Get going!
GSD is really the end goal, so in assumed coordinates, if the Drone is collecting GPS youre stuck in the EPSG/ITRF realm.?ÿ Higher dollar set ups with an IMU might be able to free up the GPS connection, but it is what it is, and works best when used in tandem.
The great advantage of existing features like road markings is that they remain for use on a future date (until the highways people repaint them!).
In open areas with only vegetation we use corrugated brown card (12-15" square) with a white cross with 3" arms. If you can't collect them afterwards then at least they degrade quickly. A wooden template to hold the card makes painting very quick and leaves good sharp edges to the arms.
@richard-imrie I really don't think so.?ÿ Just depends on the pixel, height of your flight and being able to see it to pick it when selecting GCP's when processing.?ÿ Keep it away from from edges where there's elevation change.?ÿ We found out quickly that curbs and walls don't make good GCP locations!
@jflamm?ÿ
The algorithms tend to smooth out steps, so near to a wall or adjacent to a curb then you get some degree of height error - if that is on your GCP then the error reflects through the survey since your control point height is "wrong".
We try and ensure that the control point height is typical of the ground within a metre or so. Deosn't have to be flat, just an even grade.
@jflamm?ÿ
Based upon the velocity,?ÿ path of flight and altitude too, the Nadir, the angle of the swath etc etc etc, you need to plan on these things.?ÿ
Another key component is the imaging device, what's the rate of collection there? Like Lidar, OrthoPhoto platforms collect in the same fashion now in the digital age so depending on the camera you could have tremendous amounts of data or not nearly enough.
I'm not a photogrammetrist, but got rigorously trained by a few for the work they hired me to do, and it's pretty fascinating the details they come up with and then the thing they don't care about at all.
?ÿ
@richard-imrie It's nice to have an "x" of some sort. It's easier to approximate the center. Size only matters relative to altitude but 12"-18" square covers me fine up to 400'. In a pinch you can make something in CAD print them yourself but white can wash out easily so black and light gray works best. Just make sure that when working in a farm field that the farmer is not also working that day or you might end up with this.
?ÿ
Your phone will give you close enough coordinates to come up with SPC coordinates that any software would accept as good. Just don't accept it yourself. No matter how close they are, they are still assumed coordinates. Others who see those coordinates may assume they are valid full blown SPC values and proceed accordingly. Just be forewarned.
Also be aware, that you may get better elevation values from a paper USGS map than your phone.
Paul in PA
That's a good point and an interesting one as well, Paul. Suppose your POB is an NGS mark, and you measure distance and azimuth to a second point. Then you calculate the coordinates of the second point using the SPC of the POB, observed azimuth adjusted for convergence, and observed distance adjusted by a combined factor. Are your coordinates for the second point State Plane Coordinates?
The answer is yes they are, within the accuracy of your instrument and your personal skill. What if you didn't adjust the azimuth for convergence? Well, the projected coordinates are still State Plane Coordinates. Both representations have error ellipses and the one for the second set of coordinates is bigger--there's more uncertainty--but they're still State Plane Coordinates.
A phone-measured lat/lon converted to state plane produces State Plane Coordinates for the point, but the uncertainty is indeterminate and likely very high. However, there is an error ellipse that will contain both the measured coordinates and the point they're supposed to represent.
Such a measurement should be accompanied by an accuracy statement to avoid the confusion you describe.
The same problem exists when assumed coordinates are assigned to reference points based on questionable GPS. They're not likely to be mistaken for SPC, but they may not be replicable by a ground survey either.
?ÿ
?ÿ
?ÿ
