search

rising-turtle / VCU_RVI_Benchmark

Implementations of VIO methods to use the benchmark
31 stars 3 forks source link

Calibration information #1

Closed lywcn closed 4 years ago

lywcn commented 4 years ago

Thank you very much for providing this useful dataset!!!

I am wondering if you can also provide your calibration results (camera, camera-imu etc). Thank you!

rising-turtle commented 4 years ago

Hi @lywcn :

Thanks for your interest. The camera's data (color and depth) were collected with two preprocessing steps: 1. the structure core was configed in the undistortion mode, which means the color images were undistorted by the structure core driver; 2. the raw depth images were used to generate virtual depth images, which are pixel-wise corresponding to the color images, via the transformation between the left IR camera and the color camera. Therefore, for the color camera, its distortion free intrinsic parameters were obtained from the driver:
fx: 459.357, fy: 459.764, cx: 332.695, cy: 258.998, k1: 0. , k2: 0., p1: 0., p2: 0. The camera-imu transformation matrix is: R = [0.00193013, -0.999997, 0.00115338, -0.999996, -0.0019327, -0.00223606, 0.00223829, -0.00114906, -0.999997] t = [-0.00817048, 0.015075, -0.0110795] There are some examples to set up these parameters in VIO, which might be also helpful. https://github.com/rising-turtle/VINS-Mono/blob/master/config/struct_core_config.yaml https://github.com/rising-turtle/VINS-RGBD/blob/master/config/struct_core/struct_core_color_config.yaml

Hopefully, it will be helpful.

lywcn commented 4 years ago

cool! The yaml file is exactly what I need for now.