Region-based adaptive single image dehazing, detail enhancement and pre-processing using auto-colour transfer method
作者：
Highlights：
• The dark portions in a hazy image become over dark after dehazing.
• This phenomenon can be eliminated by enhancing of dark-regions of a hazy image before dehazing.
• The haze effect of an image varies according to the depth or object distance.
• Adaptive dehazing method recovers the haze free image according to haze levels that are affecting various regions of the hazy image.
• In adaptive dehazing, the number of iterations and values of parameters are adaptively chosen in accordance with haze level in different regions. (for example, less affected/moderately affected/more affected by haze).
摘要
•The dark portions in a hazy image become over dark after dehazing.•This phenomenon can be eliminated by enhancing of dark-regions of a hazy image before dehazing.•The haze effect of an image varies according to the depth or object distance.•Adaptive dehazing method recovers the haze free image according to haze levels that are affecting various regions of the hazy image.•In adaptive dehazing, the number of iterations and values of parameters are adaptively chosen in accordance with haze level in different regions. (for example, less affected/moderately affected/more affected by haze).
论文关键词：Dark regions enhancement,Adaptive single image dehazing,Detail enhancement,Auto-colour transfer method,Hazy image categorization
论文评审过程：Received 2 September 2020, Revised 18 September 2021, Accepted 3 October 2021, Available online 12 October 2021, Version of Record 27 October 2021.
论文官网地址：https://doi.org/10.1016/j.image.2021.116532
