Victor1234's profile - activity

Is it possible to convert haarcascade to LBP and how?

I get camera assembled on copter and I want to know camera position in coordinate system related with ground. I obtained rotation and translation vectors by SolvePNP function. But those coordinates related with copter. Is I need to rotated translation vector? I how to do it?

Who had experience with running OpenCL on Intel HD on Linux?

I througth about it. But if view angle is too big squares turn into quads and can not be decribed by RotatedRect. But I'll try and let you know.

I thought about it. But if view angle is too big, squares turn into quads and cannot be described as RotatedRect. But I will try and let you know.

I have vector of approximated 4-point closed contours. How I can determine that some of them are concentric?

Unless thus are not need to android build?

>-- Could NOT find JNI
Maybe you need to install Java JNI?

How to know it and is it possible to count them?

If you have approximately flat object to recognize, see feature_homography.py sample

Which version of OpenCV you are using? You code works fine on Linux 3.5.7-gentoo with OpenCV 2.4.5

Can you specify you system, opencv compilation flags, source code, camera model and error?

May be I do not fully understand your question, but what does mean "using static background image"? If your already have it, why you need to separate it?

Where I can find image set to training to reconstruct 3D-model?

I think it's related to support for qt5 in a recent commits

In file included from /var/tmp/portage/media-libs/opencv-9999/work/OpenCV-9999/modules/highgui/src/window_QT.cpp:46:0:
/var/tmp/portage/media-libs/opencv-9999/work/OpenCV-9999/modules/highgui/src/window_QT.h:81:17: fatal error: QTest: No such file or directory

Is it trunk version problem or only my?

System:
Linux machine 3.5.7-gentoo #2 SMP

I have translation, rotation and camera matrices for two camera. How I can obtain two projection matrices for them?

I calibrated two cameras and i know it extrinsic parameters, how i can determine world coordinate of object by two frame from them?

I get some results with SURF and FLANN, but i want to try some machine learning like k-nearest instead of them, where i can find good tutorial to train and use k-nearest algorithm for planar image recognision?

@Guanta, i try to use SURF, but it gave too many false point matches, it is possible to get roi mask by some other alghorithm?

What do you think, what alghoritms developers used in this video http://vimeo.com/50747223, why it so perfect?

Can you post some of the image pairs?

I read "Visual Tracking of Human Visitors under Variable-Lighting Conditions for a Responsive Audio Art Installation", A. Godbehere, A. Matsukawa, K. Goldberg, American Control Conference, Montreal, June 2012.

But i cannot find Filtering and Connected Components, Gale-Shapley Matching or Heuristic Confidence Model realizations in bgfg_gmg.cpp. What's wrong?)

You can use some profiler or getTickFrequency

I think you can use Structure from motion technic, for example

What the horizontal align you want if you move your camera at 10 cm?

Why reprojectImageTo3D is not suitable?

add capture.isOpened() assert

use <pre> tag to format you question

run lsusb and ls /dev/video*

I think you can calculate distance between point cloud and it projection to plane, multiply by proper factor and obtain your disparity point.

As i know key code return qt, not OpenCV, try to set it up, i have some strange result too at once.