A novel underwater de-scattering method based on sparse non-negative matrix factorization
| Autor: | Xiaopeng Liu, Junyu Dong, Hina Saeeda |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Scattering
Computer science Acoustics Feature extraction 020206 networking & telecommunications 02 engineering and technology Light scattering Non-negative matrix factorization Matrix decomposition 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Underwater Image restoration Sparse matrix |
| Zdroj: | SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI |
| DOI: | 10.1109/uic-atc.2017.8397463 |
| Popis: | Underwater image processing is a challenging problem due to the scattering effect in the ocean. Floating particles and light attenuation cause underwater images to appear as bluish and blurred. According to the underwater light propagation model, the underwater image can be divided into the direct component and scattering components. In this paper, we propose the Sparse Non-negative Matrix Factorization(SNMF) to remove the scattering effect of the degraded underwater image, which separates the direct light from the scattering light. Through SNMF, the input underwater image can be divided into two additive parts, direct component and scattering component. While the scattering component is constrained to have a relatively low sparseness to obtain the restored underwater image. The experimental results demonstrate that the proposed method has a improved visual quality and equivalent to the state-of-the-art underwater image processing methods. Moreover it is simple and easy to implement. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|