how to get test-data files???

228 views
Skip to first unread message

Rui Liu

unread,
Jun 21, 2015, 3:47:48 PM6/21/15
to multica...@googlegroups.com
Hi, I am wondering how to get those test-data, like points.dat, Res.dat and IdMat.dat. Obviously they are the inputs for the gocal.m file. In my opinions, they are from the images we actually took for the laser pointer. And I observed the sample data sets provided in the website are image samples. So, can anyone tell me how to extract those (points.dat, Res.dat and IdMat.dat) from the images? 

Really appreciate any answers!!!



-Rui

Rui Liu

unread,
Jun 21, 2015, 6:21:54 PM6/21/15
to multica...@googlegroups.com
Should we follow the README.txt under the MultiCamSelfCal folder? If so, I wonder how should we set the local variables for the BlueC  project? I did not find the CamSelfCal code under BlueCCal and auxiliary scripts under any folder called blueCAcquire. I know I might miss something. Anyone can give any suggestions or tips?

I can get the calibration result by using the gocal.m with the provided data set under strawlab folder.  
Message has been deleted

Rui Liu

unread,
Jun 22, 2015, 12:33:44 PM6/22/15
to multica...@googlegroups.com
Never mind. I think I figured it out. We can extract those data by using the .m files under FingPoints folder. But I still wonder how the environment variables should be set?


On Sunday, June 21, 2015 at 2:47:48 PM UTC-5, Rui Liu wrote:

Rui Liu

unread,
Jun 23, 2015, 9:16:31 PM6/23/15
to multica...@googlegroups.com
I still did not extract the points.dat and so on information from the images, hope anyone can give me hints what i should notice when I use im2points.m...

Sean Gart

unread,
Jun 24, 2015, 4:55:26 AM6/24/15
to multica...@googlegroups.com
Hi Rui,

im2points.m should generate the files like points.dat, res, and IdMat. You don't have to worry about the readme. Are you getting an error when you run im2points or is it just not outputting the files? One thing is to make sure you have lots of calibration images or it doesn't like to work properly. 

-Sean

Rui Liu

unread,
Jun 24, 2015, 10:38:47 AM6/24/15
to multica...@googlegroups.com
Hi, Sean, thanks for your reply. I think I probably made a mistake somewhere. Here is how I tested the provided images with 4 cameras from their website (http://cmp.felk.cvut.cz/~svoboda/SelfCal/SampleData/).

As I am using Matlab, I used a configuration file path as the input of the read_configuration() inside im2points.m. My configuration file is like the following:

[Paths]
Data: C:/Users/Rui/Documents/MATLAB/MultiCamSelfCal-master/MultiCamSelfCal/FindingPoints/TestDataOneSide.images_only
Camera-Filename: arctic%d

[Files]
Image-Name-Prefix: arctic%d.pvi.*.
Basename: arctic
Num-Cameras: 4
Num-Projectors: 0
projector-Data: files.txt
Image-Extension: jpg

[Images]
LED_Size: 7 
LED_Color: green
Subpix:0.3333
Camera_Resolution: 640 480
Projector_Resolution: 640 480

[Calibration]
Num-Cameras: 4
Num-Projectors: 0
Nonlinear-Parameters: 30    0    1    0    0    0
Nonlinear-Update: 1   0   1   0   0   0
Initial-Tolerance: 10
Do-Global-Iterations: 0
Global-Iteration-Threshold: 0.5
Global-Iteration-Max: 100
Num-Cameras-Fill: 2
Do-Bundle-Adjustment: 1
Undo-Radial: 0
Min-Points-Value: 30
N-Tuples: 3
Square-Pixels: 1

TestDataOneSide.images_only contains 4 folders arctic1, arctic2, arctic3 and arctic4, each folder contains 500 images . Image name is like: arctic1.pvi.0000.jpg

I typed im2points in matlab command window to run it. Then I got this:


Error using read_configuration (line 124)
could not open points data file "points.dat"

Error in im2points (line 22)
config = read_configuration('../../strawlab/test-data/cam4sample/cam4sampletest');



I am wondering where you put those images and what on earth is the input you set for the read_configuration() function inside im2points.m?


Again, thanks for your reply, Sean. Waiting for your reply, man. 



-Rui



--
You received this message because you are subscribed to a topic in the Google Groups "MultiCamSelfCal" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/multicamselfcal/DSRN8X0LBEY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to multicamselfc...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Message has been deleted

Rui Liu

unread,
Jun 26, 2015, 6:11:53 PM6/26/15
to multica...@googlegroups.com
Hi, Sean, never mind about these questions anymore. I think I figured them out. I did not put a right path in the configuration file. When we want to compute the images, the ending part of the path should be like: /arctic%d/

I got the average images and std images for each cameras, however, I think there maybe something wrong with the code later on. On line 138 of im2points.m, it reports error like:

Finding points in camera N     1Error using imresize
Expected input number 1, A, to be nonempty.

I checked that the input of imresize should not be empty, but here the A inside the provided code somehow is empty. Are we supposed to make some changes here?

On Wed, Jun 24, 2015 at 3:55 AM, Sean Gart <gar...@gmail.com> wrote:

--
Message has been deleted

Rui Liu

unread,
Jun 26, 2015, 6:19:39 PM6/26/15
to multica...@googlegroups.com
Sean, the line in im2points.m complains is:  [pos,err] = getpoint([sprintf(im.dir,seq(i).camId),seq(i).data(idx2data).name], SHOWFIG, config.imgs, avIM, stdIM);

Rui Liu

unread,
Jun 26, 2015, 7:21:57 PM6/26/15
to multica...@googlegroups.com
Hi, Sean, never mind, I just make the version 1 code works. Thanks a lot.

--
Reply all
Reply to author
Forward
0 new messages