The Method for Removing Rain in Multiple Images

DOI

10.2991/icmcs-18.2018.44How to use a DOI?

Keywords

Rain removal; Image recovery; Wavelet multi-level decomposition; Wavelet fusion

Abstract

The outdoor vision system under the condition of rain weather, especially when the intensity of rain is relatively high, the contrast of the images obtained is low and unclear which resulting in serious degradation. The traditional way to remove the rain is restricted by its strength, and the result is not satisfactory. According to the feature about that the vision system can obtain a plurality of consecutive different degraded images in a short time, this paper deals with multiple images to achieve the recovery. And according to the dynamic characteristics of rain and snow, the direction, strength, and shape of rain lines or snow lines are not fixed. It is difficult to establish a unified physical model in the spatial domain, but they are not affected by their dynamic characteristics in the frequency domain. This paper analyzes from the perspective of frequency domain, uses wavelet multi-level decomposition and wavelet fusion method to determine the specific layer of rain noise and formulates a fusion rule based on the degree of rain noise pollution. A certain number of layers of continuous degraded images are wavelet-fused to achieve the purpose of removing rain (snow). The simulation results show that the method proposed in this paper has satisfactory recovery results and is not subject to noise intensity.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Hong Ji
AU  - Zhen Chen
PY  - 2018/10
DA  - 2018/10
TI  - The Method for Removing Rain in Multiple Images
BT  - Proceedings of the 8th International Conference on Management and Computer Science (ICMCS 2018)
PB  - Atlantis Press
SP  - 216
EP  - 222
SN  - 2352-538X
UR  - https://doi.org/10.2991/icmcs-18.2018.44
DO  - 10.2991/icmcs-18.2018.44
ID  - Ji2018/10
ER  -