In order to prevent the failure of the white balance algorithm of color digital reflection microscope and improve the accuracy of color acquisition, a colorimetric characterization method based on non-RAW data is proposed. First, a nonlinearity correction model based on power function is established by using the neutral patches of the color card. The non-RAW data is converted into the data linearly related to the scene radiance. Then, the colorimetric characterization model is established. Finally, experiments are conducted to test the accuracy of colorimetric characterization model before and after correction of non-RAW data, and the influence of luminance change of light source on nonlinearity correction model and the accuracy of colorimetric characterization is analyzed. The results show that the corrected non-RAW data will improve the accuracy of colorimetric characterization, and the improvement effect is especially obvious for the linear colorimetric characterization model. At the same time, the luminance change of the light source will cause the variation of the nonlinear correction model as well as the accuracy of colorimetric characterization, but when the luminance change is not sharp, the general nonlinear correction model can be used.