Open AccessCombined constraint for single image dehazing
Author(s) -
He Renjie,
Wang Zhiyong,
Fan Yangyu,
Feng David Dagan
Publication year - 2015
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
ISSN - 1350-911X
DOI - 10.1049/el.2015.0707
Subject(s) - constraint (computer aided design) , image (mathematics) , channel (broadcasting) , pixel , computer science , transmission (telecommunications) , artificial intelligence , computer vision , mathematics , telecommunications , geometry
Existing single image dehazing techniques have limitations on estimating accurate transmission maps for pixels which do not meet the assumption of the dark channel prior. A combined constraint is presented for single image dehazing by taking both geometric and natural characteristics of outdoor images into consideration. The combined constraint can be applied to various regularisation‐based optimisation frameworks. Both qualitative and quantitative comparisons with the state‐of‐the‐art dehazing approaches indicate the superiority of the proposed constraint.