A backward mapping method is proposed to calculate the coordinates of flat pixels.
A reference calibration method is adopted to remove the systematic errors.
This technique can accurately measure the low-order aberrations of optical flat.
The proposed technique is very simple and easy to follow.