Transmission Binary Mapping Algorithm with Deep Learning for Underwater Scene Restoration
Authors:
- 10 August 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1545-1549
Abstract
Deep learning based image resolution reconstruction has gained popularity recently. Using sensors, underwater pictures are captured. Due to the water's exponential attenuation of the transmitted light, underwater photographs have poor visibility, including low contrast and color blurring. The identification of targets and further evaluation are made more difficult by the underwater images' degradation. This paper offers a hybrid underwater image processing method called deep learning based transmission binary mapping algorithm that combines a color balance algorithm with an underwater picture restoration technique to solve these issues. This research work proposes an enhanced picture super-resolution reconstruction approach. It is based on transmission based binary mapping algorithm for enhancement in the efficiency of machine learning concepts. To calculate transmission maps, RGB attenuation for each channels are decomposed. These suggested approach can efficiently determine light at the background and combined transmission binary mapping at underwater images, which can then be used to recover these image. In addition, color attenuation balancing technique helps to provide upgraded scenes that seems realistic. Finally, thorough testing are put into place to evaluate how well our method works. The suggested method surpasses a number of popular underwater image enhancement and restoration techniques, according to tests on photographs taken in various underwater situations.Keywords
This publication has 10 references indexed in Scilit:
- Automatic Traffic Sign Board Detection from Camera Images Using Deep learning and Binarization Search Algorithm2023
- Underwater image illumination estimation via an evolving extreme learning machine by an improved salp swarm algorithmJournal of the Optical Society of America A, 2023
- UMGAN: Underwater Image Enhancement Network for Unpaired Image-to-Image TranslationJournal of Marine Science and Engineering, 2023
- Real-time detection of uneaten feed pellets in underwater images for aquaculture using an improved YOLO-V4 networkComputers and Electronics in Agriculture, 2021
- Automatic Skin Tumor Detection Using Online Tiger Claw Region Based Segmentation – A Novel Comparative TechniqueIETE Journal of Research, 2021
- A Wavelet Based Deep Learning Method for Underwater Image Super Resolution ReconstructionIEEE Access, 2020
- Underwater Image Super-Resolution by Descattering and FusionIEEE Access, 2017
- Region-specialized underwater image restoration in inhomogeneous optical environmentsOptik, 2014
- Underwater Image Processing: State of the Art of Restoration and Image Enhancement MethodsEURASIP Journal on Advances in Signal Processing, 2010
- Underwater color constancy: enhancement of automatic live fish recognition2003