In the imaging of chemical reactions using near-infrared absorbance characteristics, the problem is that the regression model constructed based on the spectrum obtained by the spectroscope cannot be applied to the absorbance image by an actual imaging system. In this study, we corrected the mismatch of spectral information between instruments using direct standardization. A concentration regression model was constructed by machine learning based on the modified absorbance spectrum, and the reaction diffusion phenomenon by the HCl-NaOH aqueous solution flowing through the Y-shaped microfluidic chip was visualized by adapting to the absorbance image.