Publication: RLaMs-Dehazing: Optimized Depth Map Improvement Single Colour Image Dehazing
Файлы
Дата
2023
Авторы
Sangita Roy
Journal Title
Journal ISSN
Volume Title
Издатель
НИЯУ МИФИ
Аннотация
Visibility Degradation is a classical problem owing to the presence of Atmospheric Particulate Matter (APM). There are different image dehazing algorithms. Any one method cannot be relied upon as each haze condition is unique. An innovative algorithm has been proposed inverting the image formation atmospheric scattering model [2, 32]. The model has been improvised by one key factor. This is Regularized Lagrangian multiplier (RLaM) based Depth Map (DM) refinement. The algorithm has low time complexity which intrigues real-time efficient applications. Different state-of-the-art visibility algorithms have been studied and their subjective and objective performance evaluations have been evaluated. Extensive investigation shows remarkable improvement with the proposed algorithm. This method is equally applicable to different atmospheric conditions. Time complexity experimented with execution time and Big (O) for real-time effectiveness. Extensive experiment results show the potential of the proposed algorithm independent of the influence of atmospheric conditions and capturing devices adaptive to computer vision applications. Time complexity and quality output trade-off achieved with the removal of ringing artifacts efficiently
Описание
Ключевые слова
Big (O) , Ill-Posed Inverse Problem , APM , Transmission estimation , Airlight