Robust and Accurate Calibration Point Extraction with Multi-scale Chess-Board Feature Detector - Advances in Image and Graphics Technologies - page 162

almost all cameras of the first set, while we fail more over the second set due to some
ill-captured images. We think this experiment is very necessary since our algorithm is to
serve the users who might care about the performance only. It also helps us to find out
the limitations and problems, or at least to make the rules of capturing clearer to users.

data 2 (by users)

/ 通用格式

succeed
fail

data 1

/ 通用格式

/ 通用格式 / 通用格式 / 通用格式 / 通用格式 / 通用格式

Fig. 16. Experiments on user-captured data

We show some results of chess-board vertices detection and the 3D models cor-
respondingly reconstructed. With the help of our method, we can reconstruct quite
accurate and vivid models. (We choose 4 images from each scenario)

Fig. 17. Original images

Fig. 18. Vertices detection results