Color appearance evaluation of agricultural products image based on spectral information of lighting

摘要

A color calibration method of the tomato images taken by a digital camera through the post-ripening process was developed based on the spectral feature information of the lighting to the sample for the final purpose of the continuous monitoring of the surface color changes during cultivation. The influences of the spectral patterns of the lighting on the color appearance were analyzed by comparing with the standard image taken under the similar to natural light (Color Rending Index: 91, Color Temp: 5500K). The RGB values of the Red, Green and Blue sections of the standard color chart were evaluated using the spectral feature information. As the results, very good correlation between the predicted color parameters and those of the standard image taken under the similar to natural light was obtained. Additionally, based on the predicated RGB values, the color appearances of the tomato images were calibrated. The calibrated color appearance of the tomato surface consistently agreed with those acquired under the standard lighting conditions through post-ripening process. Consequently, the developed method based on the visible spectral data could be acceptable to calibrate the color appearance of the tomato image taken under the various lighting conditions for the objects with the complicated color distribution and geometrical structure.