| 摘要 |
A machine vision system (20) color camera is calibrated via two algorithms. Separate red, green and blue image gray level histograms (44) in a calibration target (40) are utilized to generate corresponding standard cumulative probability functions (46). The probability functions and a target copy is provided to the system. Repeating the above steps (48) result in comparing cumulative probability functions are compared to corresponding standard cumulative probability functions (50) to obtain red, green and blue channel gains of the uncalibrated camera (52). In a second calibration algorithm, a target patch of pure white is observed (70, 72) while determining camera outputs for red, green and blue (74). The average gain for each output channel of the color camera is determined (76). The calculated average white level value for each channel is divided into the full scale value of the uncalibrated vision system. The three gain factors of either algorithm applied to a subsequent image yields a color-corrected image (52, 78).
|
