Automatic and Accurate Shadow Detection Using Near-Infrared Information

Dominic Rufenacht,Clement Fredembach,Sabine Susstrunk

DOI: https://doi.org/10.1109/tpami.2013.229

IF: 23.6

2014-08-01

IEEE Transactions on Pattern Analysis and Machine Intelligence

Abstract:We present a method to automatically detect shadows in a fast and accurate manner by taking advantage of the inherent sensitivity of digital camera sensors to the near-infrared (NIR) part of the spectrum. Dark objects, which confound many shadow detection algorithms, often have much higher reflectance in the NIR. We can thus build an accurate shadow candidate map based on image pixels that are dark both in the visible and NIR representations. We further refine the shadow map by incorporating ratios of the visible to the NIR image, based on the observation that commonly encountered light sources have very distinct spectra in the NIR band. The results are validated on a new database, which contains visible/NIR images for a large variety of real-world shadow creating illuminant conditions, as well as manually labeled shadow ground truth. Both quantitative and qualitative evaluations show that our method outperforms current state-of-the-art shadow detection algorithms in terms of accuracy and computational efficiency.

computer science, artificial intelligence,engineering, electrical & electronic