To accurately reproduce colors, it is necessary to calibrate a digital camera that is used to acquire images. The traditional calibration methods have been used to build a color correction model that has a transformation matrix mapping data from a RGB image to standard XYZ tristimulus. Determination of this matrix is performed through matrix transformation from each color checker patch to its corresponding actual XYZ value, using least-squares regression. However, these methods cannot effectively minimize the CIE DE error because of these algorithms' limitations and CCD dark currents, which can be hard to accurately measure. In this paper, a new color correction model is presented for color calibration and helps minimize the CIEDE2000 error in order to achieve higher color fidelity. Experimental results show that our method significantly outperforms other algorithms.