Portal de Periódicos da CAPES

Spectral Reflectance Reconstruction Based on Weighted Root Polynomial

2021; Springer Nature; Linguagem: Inglês

10.1007/978-981-16-3180-1_1

2190-3026

Mingjiang Ji, Junsheng Shi, Yonghang Tai, Wanli Jiang,

Industrial Vision Systems and Defect Detection

The reconstruction of spectral reflectance by the RGB values of the camera response can not only improve the efficiency of spectral reflectance reconstruction but also reduce the reconstruction cost. However, the accuracy of spectral reflectance reconstruction is not high only through the three-channel response value. The accuracy of spectral reflectance reconstruction is also related to camera exposure. In this paper, a method of a weighted root polynomial is researched using Canon EOS 500D, Digital Color Checker SG as the training samples, Color Checker Rendition Chart as the test samples, and root-mean-square error (RMSE) and goodness of fit coefficient (GFC) as the evaluation criteria. According to the experimental results, the method can maintain the reconstruction accuracy under the same conditions and reduce the impact of the exposure changes on reconstruction accuracy when the camera exposure time changes.