Dealing with shadows: Capturing intrinsic scene appearance for image-based outdoor localisation

摘要: In outdoor environments shadows are common. These typically strong visual features cause considerable change in the appearance of a place, and therefore confound vision-based localisation approaches. this paper we describe how to convert colour image scene greyscale invariant where pixel values function underlying material property not lighting. We summarise theory shadow images discuss modelling calibration issues which important for non-ideal off-the-shelf cameras. evaluate technique with commonly used robotic camera an autonomous car operating environment, show that it can outperform use ordinary task localisation.