Messages

1

Bug Reports / Startup crashes on Photoscan/Metashape

« on: May 23, 2019, 07:51:21 AM »

Various versions of Photoscan crashing on startup?

Hi--I had an old version of photoscan that I last used a year or two ago with no issues, but now I'm getting a "not responding"  error.  Just on a whim, I decided to download Metashape standard to try out the demo, and similarly crashes on startup, with a blank "not responding" window.  I'm wondering if it's conflicting with a graphics driver or some other software--any recommendations?

Thanks!

Steve

2

General / Re: Surveying trees: canopy comes out well--tree trunks are way off.

« on: June 01, 2016, 05:38:38 PM »

Thanks for the reply--although this is actually arbitrary--not height field reconstruction. I think it's just that the quality of the points is lower (there's a lot of shadow, etc--when you look at the dense cloud you can see many more gaps closer to the hround).  But you're probably right that I'll have better luck w/the dense cloud than the model.

3

General / Surveying trees: canopy comes out well--tree trunks are way off.

« on: May 31, 2016, 10:38:11 PM »

HI forum.

Question about surveying forest/woodland.  I'm trying to survey a lemon grove in southern Spain.  Tree cover is moderate, and we're trying to reconstruct the irrigation systems in the area.  I'm using the point cloud and 3D model to basically "see through the trees", via oblique angles, to piece together some of the walls/ditches that are hidden from overhead view.  However, while the canopy and landscape reconstruction is generally good, the trees themselves (either trunks, or situations where the canopy comes down to the ground) are off significantly.  In other words the vertical error seems to be way off, while the x/y errors are acceptable.

This is a medium-quality alignment, medium dense point cloud from 139 images.

I know that there are fewer points obtainable beneath the canopy, so I wouldn't expect the resolution to be nearly as good as the canopy--but is there any way to improve it?

THanks!

4

General / Re: Photos keep aligning on 2 (or more) intersecting planes

« on: May 30, 2016, 10:25:03 PM »

update: may have solved my own problem--I set 3 markers (with DGPS coordinates), then shifted from Ground Control to Generic.  Seems to have fixed the intersecting plane problem 

5

General / Photos keep aligning on 2 (or more) intersecting planes

« on: May 30, 2016, 10:21:43 PM »

HI forum,

Another UAV-survey related question. I've got this UAV survey with about 3,000 images, which I'm now doing in chunks (to troubleshoot, more than anything else).  Each chunk represents a somewhat-systematically acquired part of my survey area, and contains 100-200 photos.  Some of the chunks align very well (not perfectly, but 90-95% alignment), and I've been getting some nice dense point clouds/orthophotos.

In other chunks, I may get 80-90% ailgnment, but the point cloud is generating points along intersecting planes (or in some cases parallel planes).  To troubleshoot, I"ve selected some of these points, run a Filter Photos by Points to isolate the offending images and remove them, then I try to realign those images.  But I get the same thing.

Now I'm going to try adding some markers prior to the alignment stage, and place them on 4-6 photos (with GCP coordinates).

So my questions:

1.) what tends to cause this intersecting plane issue?
2.) can setting a few markers in advance help this?
3.) does it matter if I used generic, Ground Control, or disabled point selection?  (I"m thinking that when I use the Ground Control pair-preselection it is using the UAV GPS coordinates which aren't all that accurate--maybe that's causing a conflict somewhere?

Any thoughts woudl be appreciated!

6

General / Re: Followup question--can markers throw off my alignment?

« on: May 30, 2016, 03:46:41 AM »

Thanks, jwoods--

Yeah, it's a sizeable project, and because of the nature of the acquisition it's a bit all-over-the-place.  SO I'm trying to get what I can out of it, and use that as a basis for a possible grant for a full-coverage, systematic (cleanly gridded) survey.  Definitely not straightforward process though! Some sections align perfectly, some are awful, some GCPs got stolen--the works

Ultimately what I'll probably do is take about a tenth of the whole thing, somewhere with good coverage, get the best results I can, then use that small section to see what we can accomplish over the full area with a systematic approach.

7

General / Re: Followup question--can markers throw off my alignment?

« on: May 27, 2016, 06:50:50 PM »

Hi Jwoods--thanks for the reply.

Markers were obtained via DGPS to an error of about 2-5 cm, which is fine for my purposes.  They're in decimal degrees. I'll check and see if I've got the reference system set incorrectly in Photoscan--where would I do that?

Thanks!

8

General / Followup question--can markers throw off my alignment?

« on: May 25, 2016, 06:21:39 AM »

Hi--I posted earlier today about an irregular UAV survey.  But until today I haven't had any problems with photo alignment--generally there's good overlap in specific areas.  When I run the align photos stage, usually I'll get about 90-95% alignment.

Today something changed, and I"m getting 0-5% of the photos aligned.  The only thing that is different is that I have added some guided markers, and added geographic reference info (lat/long/elev).  But now when I run the align photos stage, it doesn't work anymore.  I did check on some of the images, and a lot of the folded gray-flag markers are clumped together in individual pictures, so obviously I've done something wrong here.  Any idea where I've screwed up?

THanks!

9

General / Advice on Processing an irregular UAV survey area (~3500 images)

« on: May 24, 2016, 06:57:55 PM »

HI all,

I'm using Photoscan as part of a project to trace irrigation systems in parts of Spain.  The coverage is very irregular  - - we were using a quadcopter with shaky battery life over obstructed terrain.  We also only have a few GCPs -- about 6 or 7 we can use, fairly well spread out.  The GCPs were acquired via DGPS so those are pretty accurate, but if the final model is a few dozen cm or even a meter off in some places it's not the end of the world, since this is  more for survey than mapping. 

(photo coverage map attached--coverage area is about 0.25 sq. km.  Yellow pins/circles are GCPs)

I'm doing this on a (reasonably good) laptop--an Acer ASPIRE V15 Nitro--16MB ram (going up to 32) and an NVIDIA GeForce GTX 960M. 

Photos are 12 mpx, taken with a fisheye lens (may have been a GoPro but will have to check)


HEre are my questions:

1.) Is it even feasible to run a project of this size (3500 images) on my laptop without breaking it into chunks?  Say using ground control pair preselection, medium resolution point cloud/models?

If Not, and I need to break it into chunks:

2.) Can different chunks geographically overlap?  We sometimes re-flew the same area to improve coverage, and it's difficult to really break this up where each chunk is a different region.

3.) Since there are only a few GCPs, and they are very spread out, there's no way to have small chunks (say 300 images or less) have more than 1 or 2 of them.  Can ground-control based pair preselection still be used even with only a couple of markers, to then be merged later?

4.) Somehow, when I ran the first step (alignment) on all the images at lowest quality, it worked ok (about half aligned, which is about what I expected), but when I ran it again on medium alignment, it failed--only aligning 5 out of 3500 photos.  ANy reason why this might be the case? 

If anyone has thoughts on these Q's, or ideas for workflow, I'd really appreciate it!

--Steve

10

General / .wrl/.3ds or orthophoto/DEM?

« on: July 31, 2014, 06:28:28 PM »

Hi--

Probably a newbie question as I'm new to the 3D world--I've been having a really tough time exporting a georeferenced/textured model to something like ArcScene from Photoscan.  We have an existing GIS, and I wanted to provide some new photorealistic data that can be incorporated into that GIS.  Should I be exporting the whole model from Photoscan, or the orthophoto and the DEM as separate files to be opened in ArcScene?

Thanks--I've been stuck on the workflow for awhile and I'm completely confused.

--s

11

General / Re: Photoscan--:ArcScene workflow for real-world coordinates

« on: July 21, 2014, 12:10:17 PM »

Thanks, KJellis.

Thing is there is an existing GIS that's been in development for several years (the whole thing is in ArcGIS--and they're just starting to get into 3D) so I wanted to help them add some photorealism to what they're doing.

But I can't seem to figure out how to get the coordinates that I assigned to the markers in Photoscan into ArcScene--I think I'm missing something obvious but haven't been able to get it.

--s

12

General / Photoscan--:ArcScene workflow for real-world coordinates

« on: July 20, 2014, 06:54:55 PM »

Hi-

I know there have been tons of posts on this but somehow I?m still confused.  I have a nice textured model in Photoscan, and have set 8 markers and set coordinates according to our total station m easurements.  I can?t for the life of me figure out how to get this model--and its proper coordinate system--into ArcScene.  I?ve tried exporting as .dae and .wml--the models import ok but I don?t know how to position it. 

I know this is a basic question but please help--I?ve been stuck for days!

Many thanks.

Steve