Single image desmogging using Gradient channel prior and Information gain based bilateral
| Autor: | Kamlesh Lakhwani, Jeevan Bala |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Haze
business.industry Computer science ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 02 engineering and technology Atmospheric model 01 natural sciences Kernel (image processing) 0103 physical sciences 0202 electrical engineering electronic engineering information engineering Benchmark (computing) 020201 artificial intelligence & image processing Computer vision Artificial intelligence Bilateral filter Information gain Single image 010306 general physics business Image restoration |
| Zdroj: | 2020 3rd International Conference on Emerging Technologies in Computer Engineering: Machine Learning and Internet of Things (ICETCE). |
| DOI: | 10.1109/icetce48199.2020.9091768 |
| Popis: | Smog is defined as an intense air pollution. It is a combination of haze and fog and it degrades the visibility of outdoor image at a great extent. Therefore, existing imaging systems are unable to obtain the potential information from these weather degraded images. Recently many image restoration models have been implemented to remove the effect of haze and fog from images. But, removing the smog from images is defined as an ill-posed problem. Therefore, in this paper, a novel desmogging model is proposed. Initially, gradient channel prior is used to estimate the optical information of smoggy images. Thereafter, a information gain based bilateral filter is designed to refine the transmission map. Finally, the restored image is obtained using an improved restoration model. Extensive experiments have been carried out by considering benchmark hazy images. Performance analysis reveal that the proposed technique outperforms the existing visibility restoration techniques. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|