Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - ozbigben

Pages: [1] 2 3 ... 5
Python and Java API / Re: Setting bounding box via 2 markers
« on: July 18, 2018, 07:15:17 AM »
Previous steps:
Load image folders to chunks
Detect markers
Import reference
Align images

Python and Java API / Setting bounding box via 2 markers
« on: July 18, 2018, 07:01:37 AM »
Hi all

I'm refining a workflow for doing high resolution scanning using photogrammetry and am at the stage of adding some more automation. Most of the generic processing stuff I can probably work my way through but I have one things that's a bit more complex that's similar to

I'm using an L-shaped target with coded CPs to set the orientation of the image. This is working nicely and I'm looking to add cropping to this.  Having looked through a few posts this might be best done by defining the bounding box after alignment so that only the required area is processed.  I detect markers and import cords prior to alignment so rotation of the model is not required. Bounding box would need to be aligned to the coordinate system (already a script for that) and then positioned/sized to match CPs defining the top/left and bottom right corners.  CP labels would be constant and named during detection "target NNN"  Height of bounding box is OK after alignment

Current setup is attached. Tope Left CP would be added to L-shaped scale bar, bottom right would be a single CP placed manually. I could hard -code the top/left cords but the script would be a little more re-useable getting the cords from a CP.

Any help would be greatly appreciated and happy to share details of the rest of the setup.  Using a Hasselblad H5D60 to scan at 600dpi up to ~2x4m including foldouts from very large books which don't lie flat.

General / Re: Running Photoscan on a VM
« on: August 26, 2014, 08:22:09 AM »
As with the other GPU-related error in posts here it seems to be the graphics driver.  NVidia graphics driver was 340.52 (or something like that). Looked at the local machine that didn't have any problems and it was running 331.65 so I did a clean install of that version on the VM

Code: [Select]
finished depth reconstruction in 318.432 seconds
Device 1 performance: 105.843 million samples/sec (CPU)
Device 2 performance: 106.71 million samples/sec (GRID K2)
Total performance: 212.553 million samples/sec

Code: [Select]
finished depth reconstruction in 148.631 seconds
Device 1 performance: 140.158 million samples/sec (CPU)
Device 2 performance: 434.571 million samples/sec (GRID K2)
Total performance: 574.73 million samples/sec

[edit] dammit... thought I got rid of that doppelganger login on one of my PCs... ;)  and looked up the correct version it was running previously

General / Re: CPU and GPU benchmarks
« on: August 23, 2014, 10:36:29 AM »
I've just been handed a VM by our IT department primarily for testing Photoscan. Interesting weeks ahead  ;D

PC: Dell Precision T5600 , 635W,  Xeon E5-2630 (Six Core, 2.30GHz Turbo, 15MB, 7.2 GT/s), 32GB (4x8GB) 1600MHz DDR3 ECC RDIMM, 3 GB NVIDIA Quadro K4000

VM (via PCoIP): 16 vCPU cores and 64 GB RAM + equivalent to Quadro K5000 GPU (dedicated)

General / Re: Optimization generates heavy noise in sparse cloud
« on: August 19, 2014, 12:10:43 AM »
I think he suggested the export so you could import after optimisation.

General / Re: 1980s Shipwreck Excavation - help with camera calibration?
« on: August 16, 2014, 09:15:02 AM »
I am about to start looking at some similar data for 3D reconstruction. However mine is c. 1970 black and white negatives.

if you haven't scanned the negs yet, we've been using a Canon 5DmkII, 100mm macro f2.8 IS and lightbox to do this efficiently, with an effective resolution around 3800dpi. We use Adobe Bridge to do the initial colour adjustment and save as 16bit greyscale initially until the contrast adjustments and any processing are finalised.  The IS model of the macro lens is noticeably sharper than the earlier model (optically, not from the image stabiliser)

General / Re: how to relate cameras and wrong parts of a cloud ?
« on: August 16, 2014, 09:05:09 AM »
I've been doing some experimenting with shooting scenarios using screenshots from Skyrim (so no GPS coordinates to provide initial camera positions). I've had occasions where I've had two oblique planes and fixed it by setting the initial lens settings with the calculated values from the alignment and repeating the alignment step.  I've also had situations where large groups of images in a sequence weren't aligned, but selecting them and aligning just the selected images worked nicely. Image sets were about 500 in these cases

General / Re: Developing RAW to JPEG
« on: August 16, 2014, 08:55:39 AM »
i use LZW TIFF for most applications including Photoscan.  I'm just using Bridge, setting Black 0 and Contrast -20 on top of the usual white balance etc...

General / Re: single line processing
« on: July 23, 2014, 01:56:28 AM »
You should still get a result as long as the camera is reasonably perpendicular to the flight path

Good to see your persistence paid off. At least now you have some invaluable knowledge of what doesn't work and why which in the long run can be more useful than if things work the first time.  :)

General / Re: Fixed camera,turning object workflow.
« on: July 06, 2014, 03:16:02 PM »
Yes, although a white or black background may make more sense. Green could give you colour casts in your texture. You would still want to shoot multiple rows at different heights.

General / Re: Capturing a good dataset
« on: July 06, 2014, 03:16:52 AM »
Climb higher than the glacier and shoot across it, but even then you'd need to go along both sides of the glacier for a good result

Forgot to mention that one of the good things about Exiftool id that even though it may not be able to write a specific tag, it will extract any metadata it can find in an image. I use:
Code: [Select]
exiftool -s %1 > metadata.txtfor a drag and drop batch file to list the available metadata in an image into a text file. When you're setting up your format file just take the field names from this and prefix them with "$" wherever you want a field value written.

You could try using EXIFTool to format the EXIF data from the images e.g.

General / Re: Anyone tried Mars stereo pairs?
« on: July 04, 2014, 12:56:02 AM »
Cool, might have to have a play. Preparing some data for rendering but there are some annoying gaps in the DTMs where stereo pairs are available.

Pages: [1] 2 3 ... 5