Topics

1

Feature Requests / Visual DTM Alignment for DEM Export

« on: November 23, 2011, 10:48:49 AM »

I have two models of the same area (riverbank) that were created at two different time periods. I want to try and measure the difference between these two models, however, my surveyed coordinates were a few feet off, which causes models to be really of when exporting them as DEMs. Would there be a way so that you could view two DTMs in the same visual space, move them both around, and export them as DEMs? There are enough similarities in the models that a visual overlap would be sufficient, but I cannot figure out a good way to do this.

Thanks,

Tyler

2

General / Geometry Accuracy Calculations/Validation

« on: November 19, 2011, 05:52:07 AM »

Howdy all,

Does anyone know if there is a "best" method for calculating or estimating the overall accuracy of a model that is generated? From all that I have read the only way to do that is with comparing actual elevation values that are surveyed independently or determined from a different method. Lets say I only have 15 GCPs that each have their own RSME value, is there anyway to estimate the accuracy across the model into a single value from the 15 RMSE values that I have? I haven't found anything in the literature and I just wanted to see if anyone knew.

Thanks,

Tyler

3

General / Importing Photos w/ masks and GCP's from another project file

« on: November 18, 2011, 04:48:34 AM »

Howdy,

I have two separate files that are of the same scene, but were created with two different cameras. Is there a way to import all of those picture files into the same chunk while keeping all of the GCP's and masks of the images? I couldn't figure out how to if there is and this would save a lot of time having to redo the GCPs manually.

Thanks,

Tyler

4

General / Accuracy of Point Cloud Changed?

« on: November 16, 2011, 02:42:19 PM »

I am running on a computer that has little local space on the hard drive so I've been going back and forth from a network drive to the local drive. I ran the same exact chunk in the same exact file with one file location on the local drive and another on the network drive because I just wanted to test performance for Aligning photos (I also ran a Build Geometry yesterday that took 20+ hrs on the network drive when it used to take about 12 on the local). The RMSE for my GCPs was significantly different between the two files...this seems super weird.

Screen shots of the two GCP windows are attached.

5

General / DEM Exporting Validity

« on: November 15, 2011, 01:51:56 AM »

Howdy again,

This question involves how that elevation value of a DEM is calculated if there are multiple polygons that exist for that particular projection. This could happen in real life if there was a cliff and an overhang. So, the DEM's that I am producing are not from aerial photos, but photos taken about 10m away from a stream bank where there is a nearly vertical outcrop of soil with a lot of boulders and cobbles at the bottom.

If there is a large rock in the foreground of the projection, will a DEM export the elevations of that rock, or the elvations of the bank behind the rock?

Here is some visual help. It is most helpful to look at the bottom right XYZ Axis to make sure you know what angle you are looking at, but the first picture is looking directly along the X axis so the Z-axis is completely Vertical, and the y-axis is completely to the right. In the second picture, the axes have been rotated so we can see behind this boulder, and that yellow "X" mark would not be visible if the DEM was produced looking along the X axis (like the first picture).


http://www.tylerdepke.com/wp/wp-content/uploads/2011/11/DEM-Q1.jpg


http://www.tylerdepke.com/wp/wp-content/uploads/2011/11/DEM-Q2.jpg

This would be very important for analyzing changes in volume of sediment and rocks from one time frame to another.

Thanks again,

Tyler

6

General / Problems Building Geometry

« on: November 15, 2011, 01:32:59 AM »

Howdy all,

I've encountered the problem a few times now and I am unsure why it is occuring and how to fix it. After doing a Build Geometry, the results say that there were 200,000 faces when this is clearly not the case. Here is a screenshot of the project. There are 3 chunks, and as you can see, the middle chunk did not come out right.

http://www.tylerdepke.com/wp/wp-content/uploads/2011/11/SS-3-Chunks-Middle-Missing.jpg

I'm wondering why the middle portion of the model, where the cameras had what I would consider better angles, did not come out correctly. I am looking at my GCP's right now, but they don't look too bad either so I am stumped. I ran all three chunck as batch processes where they were to all build geometry and build textures.

Thanks,

Tyler

7

General / Difference Map or 3D Model

« on: November 10, 2011, 09:40:12 AM »

Howdy,

As far as I know, the only 3D analysis that can be done in this program are volume calculations and distance measurements. I would like to be able to take two separate models, overlap them, and be able to calculate the NET erosion or deposition AND calculate the total erosion and total deposition of this area. I would also like to have a model that quantitatively shows areas of erosion and areas of deposition (probably using some kind of color gradient draped onto a 3D model).

Has anyone done this using models from Agisoft or know the best way to go about doing this?

Thanks,

Tyler

8

General / Cannot Find "Optimize Point Cloud" Option

« on: November 03, 2011, 03:05:01 AM »

Howdy,

I've loaded 19 oblique aerial photos, put about 30 GCP's total (each photo has about 10-15 points), loaded the actual coordinates for the GCPS, Ran the "Image Adjustment" to create a point cloud and when I go into Control Points to try and optimize there is no option to "Optimize" as it states in the manual. Is there something else I have to do in order to optimize?

Thanks,

Tyler

9

Camera Calibration / "Best" Calibration Parameters?

« on: November 03, 2011, 03:00:25 AM »

I am not an expert in photogrammetry, although when doing research a few years back I was told that if possible when loading calibrations it is best to have AT LEAST k1, k2, and k3 and beyond that solving and loading more variables into software that is just going to re-solve the cameras interior parameters could create more problems. Can someone explain what the most optimal calibration solution is or what the advantages and disadvantages are for solving for different variables?

This software only has the ability to solve for:

k1, k2
k1, k1, k3
k1, k2, p1, p2
k1, k2, k3, p1, p2

Please only refer to these variable options and if you can point me to any literature that would be helpful as well.

Thanks,

Tyler

10

Camera Calibration / Question about Calibrating Long Focal Lengths and Focusing Lens

« on: November 03, 2011, 02:51:29 AM »

Howdy all,

I am a masters student working on a terrestrial photogrammetry project for my thesis. I have some photos that I took from oblique angles on cliffs using a 50mm fixed and a 70mm both focused at infinity.

I am trying to calibrate a 50mm fixed a 70-300mm set at 70mm using the checkerboard pattern and a projector, but I cannot fill the entire image with the pattern and focus at infinity.

What would the best option be:

1. Not fill the entire image with the checkboard pattern but keep the lenses focused at infinity
2. Move closer to the projector to fill the image, but have the images blurry
3. Move closer to the projector to fill the image, adjust the focus to keep the images sharp (even though that was not how the lens was focused for the images I plan to create models with)


I am not sure how much the focus will affect the focal length while zooming the lens will obviously change the focal length drastically.

Thanks,

Tyler