For corner detection of black white checkerboard, a new sub-pixel operator with high speed and high position accuracy was proposed. The principle of Hessian matrix, the execution procedure of the algorithm and basic ways of getting the sub-pixel coordinate in black white checkerboard were presented. All the corners needed in black white checkerboard were detected and all the coordinates of corners were calculated through experiments. Sub-pixel coordinates of corners were calculated with two order Taylor expansion. Finally, under different noise levels, compared with the Harris algorithm, extraction error and extraction time are calculated. The experimental results show that Hessian matrix is more accurate for corner detection of black white checkerboard. The result can be used in camera calibration.