Messages

16

General / Re: Curved scan

« on: June 21, 2018, 11:45:09 PM »

I had this problem as well in the past.
Run the camera calibration (take images of the checker board image on the screen with the camera as instructed in the user manual) and generate a lens calibration model.
The apply that lens calibration data to the chunk you're processing.

It will fix the curved model you have.

17

General / Re: My observations with a Phantom 4 Pro ; DNG vs JPG

« on: June 21, 2018, 03:51:36 AM »

There is no difference in focus between the images, and the focal length can't change... ihe focal length of the lens is  8.8 mm/24 mm (35mm equivalent)

The camera on the P4P aloows for RAW+JPG to be saved at the same time, which is what we did.
So I'm comparing the RAW vs the JPG that was taken at the same exact moment.

18

General / Re: Working with single images and convert to KML

« on: June 20, 2018, 02:01:28 AM »

Single images can't be used in the process of photogrammetry, which is that the software PhotoScan does, so I'm not surprised there was no exporting option.

The geo-tagging on the image only identifies the location of the camera, not the location of the image contents.

19

General / Re: Mesh too big not enough memory what to do?

« on: June 20, 2018, 01:57:28 AM »

Why can you not merge multiple meshes?

20

General / Re: Mesh too big not enough memory what to do?

« on: June 19, 2018, 11:10:55 PM »

Break up the project int multiple chunks with smaller regions.
Process each region, then merge the chunks back together.

I'd start with breaking up the project into 2 chunks to start. If you're still running into ram problems, break up the project into 4 chunks.

21

General / Re: indoors photography problem

« on: June 19, 2018, 11:08:01 PM »

The shutter speed is WAY to slow if you're hand-holding the camera.
Consider using a tripod instead, and then you can shoot much longer exposures, and be able to gather much more light that way.

22

General / Re: lines in texture files

« on: June 16, 2018, 07:50:42 PM »

MK_1, do you have the Ghosting Filter Enabled when creating the textures?
The check box for it is hidden in the Advanced box in the Build Texture operation window.

23

General / Re: My observations with a Phantom 4 Pro ; DNG vs JPG

« on: June 15, 2018, 08:33:03 PM »

Increasing the distance to the object by a factor of 2 (so the object is smaller in the frame by a factor of 2), creates even more extreme results with the mesh reconstruction.

The mesh reconstructed from the JPG image is all noise, whereas the mesh reconstructed from DNG files actually resembles the real object.

24

General / My observations with a Phantom 4 Pro ; DNG vs JPG

« on: June 15, 2018, 08:27:37 PM »

I know this is a debated subject (RAW vs JPG for photogrammetry), but I just wanted to share some objective evidence I came across. I'm working on a project to figure out how much detail my Phantom 4 Pro can resolve (will post a full write-up on that soon), I came across an interesting observation.
This only applies to images captured with the Phantom 4 Pro as it's the only sensor I've tested so far.

Took DNG+JPG images with the P4P camera, and ran them through PS. (1/200 shutter, f6.5, ISO 400)

The object that is being reconstructed is a calibration chart I made to do the detail resolution test. The object is not very big in the frame (about 480 pixels across) and the image width is 5464 pixels across (so less then 1/10th of the image)

My observations on the source images: (see attached image)
The JPG image (out of camera) looks cleaner. Less image noise. (likely smoothed and sharpened by the jpg compressor)
In contrast, the DNG image looks full or grain and noise.

My observations on the reconstructed mesh:
The mesh created from the DNG files is cleaner while the mesh created from the JPG files is noticeably rougher looking/noisier.

Just to note that DJI doesn't provide ability to change settings for JPG creation.

If you're using a P4P, I'd love to hear from you and what your experience has been with JPG vs DNG.

25

General / Weird message when aligning

« on: June 15, 2018, 06:45:33 PM »

A weird message gets displayed when I align a bunch of DNG images (shot on the Phantom 4 Pro)
The message is:
libtiff error: input: Null count for "Tag 51008" (type 7, writecount -3, passcount 1)

The images seem to align just fine, but I was wondering what that message meant?
I could not find any reference of it on google.

Anyone get this before?

26

General / Re: How to spend $5k USD

« on: June 15, 2018, 01:54:59 AM »

Well for $5k you could put together something similar in terms of CPU power compared to the NC24 system, but not very close when it comes to GPU performance.
 
You won't be able to build a dual CPU system with i7 or i9 CPUs. You need Xeon E5 CPUs for those systems, and Xeons are much more expensive then the i7/i9 counterparts..
A top end i9 (like a 7940X or 7960X) will likely be as fast as the dual E5 in the NC24 system) Those will run you $1200 to $1400 respectively.
224GB of ram will likely be about $2000.
A Motherboard that can handle 256GB ram will be about $500.
The Tesla k80 GPUs are about $2k a piece, and the 1080Ti are not far off at about 1/3 that price.
Then you'll need some storage, a case, PSU.... another $500...

Hopefully this gives you a but of an idea of how far your $5k will go.

27

General / Re: Ikea Lack Photogrammetry Turntable

« on: June 12, 2018, 03:37:06 PM »

Thanks for sharing!

Where did you get the lazy susan from?

28

General / Re: slow processing, workstation issue?

« on: June 12, 2018, 03:36:17 PM »

Of course newer hardware will speed up processing, but you haven't told us what kinds of workloads you're throwing at your system. How many images? What resolution?

Having said that, if it's just a speed issue, you likely already have enough RAM to process all the images, so more ram won't help.
More GPUs will help, but with diminishing returns. They don't scale linearly. For example 4GPUs won't be 4 times as fast.
See this article for more insight and some benchmarks:
https://www.pugetsystems.com/labs/articles/Agisoft-PhotoScan-GPU-Acceleration-710/

As for the CPU, you can certainly improve it for more speed. Just by replacing your CPU you can easily double the speed for a few hundred dollars. Look for a E52680 or E5-2690 which should be a drop in replacement. 
Just be careful which Xeon E5 CPU you choose. The V3 and V4 CPUs are not compatible with the earlier version V1 and V2  motherboards, so choose wisely.

If you decide that you need more horsepower then what a drop in CPU replacement can bring,  you are looking at a whole new system (new motherboard, new ram, new CPU)

29

General / Re: Measuring photogrammetry resolution

« on: June 07, 2018, 03:16:36 PM »

Thanks for the reply SAV,
Both GSD and Spacial Resolution are purely theoretical numbers. And while they are useful to a degree, it's like judging a photograph purely based on the megapixels the sensors. We all know there's more to how sharp an image is, then just the sensor.

To measure the ability of a camera/lens combination to capture data, charts like these are used to determine the resolution of the lens/sensor combination. Change the lens, and the captured chart will show different results (more or less resolution depending if the lens is able to resolve more or less detail)

Camera resolution test chart from ISO 12233:2000

Similarly in photogrammetry, the lens plays a HUGE part in the resolution the camera can capture, which is why I called GSD and Spacial Resolution purely theoretical. Depending on what lens is used (and how the entire system is used) you can get closer or further to that theoretical number.

Would I be wrong to assume that there's no standardized method to measure Actual Spacial Resolution (not theoretical) with some sort of equipment similar to what the chart above does for photography?
 

30

General / Re: Minimum system requirements

« on: June 07, 2018, 03:26:17 AM »

In my case, I will upgrade RAM but also operating systems and linux swap partition will run on a top notch SDD disk in case there is RAM memory pagination.

Strongly recommend against planning to use the swap file when ram gets full.
Even with the fastest SSD, the software slows down dramatically!

I would rather turn off the swap file so the program can tell you it doesn't have enough ram. At which point you can split up the prject into chunks, to reduce the amount of memory required.