xlz
commented
5 years ago You don't want to "calibrate registered images" because registered images are synthetically produced with existing calibration parameters to represent what the RGB camera would see if it were placed at the exact pose of the depth camera.
You would want to calibrate using the source images directly from the sensors, i.e. the RGB image and the IR image, to derive the intrinsics of the two cameras and the extrinsic transformation between the two.
This is not an error, but I have a question. I want to use the registered image to perform some tracking, and then I can easily find the corresponding points, however the registered image is dirty, it contains shadows of the depth information.
For my marker tracking this is ok, but my marker tracking systems requires I provide camera calibration information. So far my attempts at calibrating the registered image with opencv chessboards failed (The same code with other images sources even low resolutions, is giving me great results).
I am guessing that actually, the registered image is already undistorted, so the process I am going through is not needed but I am not sure.
Does anyone have an opencv camera calibration result for the registered image? Or if it is already undistorted does anyone know how I can make a calibration file that makes no changes when applied?
My calibration files are formatted as follows:
`%YAML:1.0 calibration_time: "dom 27 feb 2011 17:16:27 CET" nframes: 5 image_width: 640 image_height: 480 board_width: 8 board_height: 6 square_size: 2.8999999165534973e-02 flags: 0 camera_matrix: !!opencv-matrix rows: 3 cols: 3 dt: d data: [ 628.158, 0., 324.099, 0., 628.156, 260.908, 0., 0., 1. ] avg_reprojection_error: 8.4719362981201396e-01 per_view_reprojection_errors: !!opencv-matrix rows: 5 cols: 1 dt: f data: [ 4.28777784e-01, 9.05612946e-01, 1.08689988e+00, 6.98072255e-01, 9.57101762e-01 ]
a set of 6-tuples (rotation vector + translation vector) for each view
extrinsic_parameters: !!opencv-matrix rows: 5 cols: 6 dt: f data: [ 6.47093183e-33, 6.47093183e-33, 0., 0., 0., 0., 0., 0., 0., 0., 0., 6.47067322e-33, .Nan, 2.80259693e-45, 3.73649902e+02, 128., 6.47104938e-33, .Nan, .Nan, 1.44968018e+02, 6.47105820e-33, 0., 1.52838165e+02, 2.27430847e+02, .Nan, 2.21454285e+02, 2.61589783e+02, 2.16458176e+02, 3.19432526e+02, 2.12284958e+02 ] ` Cheers
Fred