A shadow removal method based on color model and undecimated wavelet transform

This paper proposed a shadow removal algorithm based on color model and undecimated wavelet transformation according to properties of the shadow and characters of each band. Firstly, the shadow area in original image was accurately extracted based on HSV color model. Secondly, Euclidean distance was used to determine the adjacent non-shadow area in order to compensate the shadow area. The statistics of shadow and non-shadow area was combined in undecimated wavelet transform to eliminate shadows. Finally, we compared our method with the algorithm based on HSI histogram equalization and RGB normalized. Experiment results showed that our method had better performance.