mew's profile - activity

I have written the most basic of Vis applications. In essence, here is the main():

viz::Viz3d myWindow("Coordinate Frame");
myWindow.spin()

When I execute this program, the spin method completes instantaneously. Well, not quite: I see a window flash in front of me before it completes. It is my understanding that the event loop should keep spinning until it detects relevant user input.

I tried restructuring my code, replacing the spin() with a while loop:

while(!myWindow.wasStopped()) {
   myWindow.spinOnce(1, true);
}

As expected, the spinOnce exhibits the same behavior, and completes without displaying any window.

Is this a problem with my installation?

Did you ever identify the problem here? Also, what program are you using for visualization?

By accident I tried estimating the relative position of an image to itself (don't ask). I would expect a result of 0 translation and 0 rotation.

Surprisingly, I get a non-zero translation result. In fact I get a rather significant result: 0.0825 -0.0825.

In essence my code is as follows:

cv::Point2d pp(u0, v0);
cv::Mat R, t, mask;
cv::Mat E = cv::findEssentialMat(points1, points2, focal, pp, cv::RANSAC, 0.999, 1.0, mask);
cv::recoverPose(E, points1, points2, R, t);

In the above code, t != 0. My question is: is a non-zero result for recoverPose valid when points1 and points2 are identical? If so, why?

I have a sequence of images. I would like to estimate the camera pose for each frame in my sequence.

I have calculated features in each frame, and tracked them through the sequence. I would now like to estimate the camera pose for each frame. I do so using the following openCV routines:

Mat essentialMatrix = findEssentialMat(pointsA, pointsB, f, pp, RANSAC, 0.999, 1.0, mask); recoverPose(essentialMatrix, pointsA, pointsB, R, T, focalLength, principlePoint); Where pointsA and pointsB are comprised of 2D coordinates of features that are present in both frames. pointsA is associated with the frame before pointsB.

The problem I am encountering is that the R and T estimations are very noisy, to the point where I believe something is wrong with my pose estimation.

My question is, how do I estimate the camera pose from two sets of features?

Note: I am familiar with this answered question. However, I believe openCV3 now includes methods that address this problem more eloquently.