Stray Light Reduction and Multiple Spectrum Extraction by Multi-frequency-Modulation and its Application on Oil Painting Identification

In this study, a simple and practical method has been developed to correct the response of spectroscopy for measurement errors arising from the background electric noise and the stray light from the environment and instruments. A sinusoidal modulated light intensity signal was incident on a tested sample, and then the reflected light was collected by the spectrometer. Using the three-parameter sine curve fitting algorithm, we can extract the primary reflectance signal from the total measured signal, which contains the primary reflectance signal and the stray light from the environment. Two cases of one-spectrum extraction and two-mixed spectra extraction were performed to demonstrate the feasibility of this method. The similarity between the extracted spectra by this proposed method with extreme environment stray light is higher than 95%. Finally, we applied this method to the oil painting identification, and the results showed that it could successfully distinguish the printed fake painting from the spectrum similarity comparison. This method can eliminate the affection of the background electric noise and the stray light from the environment. Moreover, it took the same measurement time as single wavelength extraction for multiple wavelengths extraction.