pranav's profile - activity

Hi,

Can anybody discuss possible reasons why imageData member in IplImage structure has pixel elements of char datatype instead of unsigned char? Specifically, what does it mean to have pixel intensity with anything other than 0 or +ve values?

Hi,

I am planning an experiment to evaluate accuracy of homography computed using cvfindhomography. I want to compare RANSAC & LMEDS options. I will select N known points(coordinates are known) on a planer screen and will capture its view using a webcam. Then i will select a subset of points from them for computing homography between screen and camera sensor plane. Using this homography i will project all other screen points(which were not used for homography estimation) to camera image plane. Error criteria for homography then will be radial distance of projections using estimated homography and their true position as detected in the undistorted image(assuming corner detector detects corners accurately).

Actually the target is to model the behaviour of error in estimated corner locations so as to determine an error compensation term which can be used to correct the corner locations estimated by homography.

Has anyone tried similar experiment with RANSAC & LMEDS for cvfindhomography or has any suggestion for a better approach to model such error term?

Thanks waqar for useful reference.

So, in the figure in question, if e' and x' are represented in homogenous coordinates then their cross-product will represent a normal to the plane which passes through e',x' and origin (0,0,0) of homogenous coordinate system. And intersection between this plane and Z=W, where e'=(x1,y1,W) and x'=(x2,y2,W), will be the line l' hence l' can be represented as cross product of e' and x'.

Hi, I have observed that equations relating undistorted(or corrected) image points and distorted image points given here link text and here link text are contradictory.In first link,(x',y')(undistorted coordinates) are mapped to (x",y")(distorted coordinates) using same factors for tangential & radial distortion as (x,y)(distorted coordinates) are mapped to (x_corrected,y_corrected)(undistorted coordinates) in the second link.Is it an issue or i have interpreted the equations wrongly?

Hi, There are some methods reported in literature for example try googling: A Comparative Review Of Camera Calibrating Methods with Accuracy Evaluation and you can also refer Accuracy assesment section of Tsai's seminal paper were he proposes such tests: A versatile camera calibration technique for high accuracy 3d machine vision metrology using off-the-shelf tv cameras and lenses (1987). Further, you can refer section-2 in paper : Accuracy Improvement in Camera Calibration by FaJie Li, Qi Zang and Reinhard Klette, which also mentions some approaches for evaluating camera calibration accuracy.

Hi, I want to know whether we should pass original(distorted) image-points to cvFindExtrinsicCameraParams2 or we have to undistort them before passing as an argument? Although i think we need to give original image-points because we are also passing distortion coefficients but still wanted to confirm since it is not mentioned explicitly in: http://opencv.willowgarage.com/documentation/camera_calibration_and_3d_reconstruction.html#findextrinsiccameraparams2