Shadow detection using chromaticity and entropy in colour image

Shadows in an image often bring a significant problem which can cause unintended negative outcome, so how to detect shadow is an important issue of computer vision tasks. This paper proposed a method to detect shadows from real images. Due to shadows in image have a dark pixel value, shadow candidates are defined. Shadow candidates have been estimated and detected by chromaticity of colour image and threshold image using entropy. Some experiments are conducted so as to verify the proposed method, and results show that the proposed method can detect shadows in colour image. © 2018 Inderscience Enterprises Ltd.