Abstract:
This paper builds a raindrop degradation model based on imaging experiment, which models the image distortion caused by raindrop on the lens of camera from four aspects, including blurring, distortion, displacement and attenuation. According to this model, it proposes a restoration algorithm including inverse scaling, overlapping separation, and filtering, etc. Experimental results show that it can depict the distortion caused by raindrop well, and can get a promising result by separating overlapping images and correcting distortion effectively in general.
Key words:
image restoration,
raindrop degradation model,
overlapping image separation
摘要: 基于成像实验建立雨滴退化模型,从模糊、变形、位移和衰减4个方面描述附着在镜头表面的雨滴对图像的退化方式,并据此提出一种包含逆向放缩、叠加分离、滤波等环节的退化图像恢复算法。实验结果证明,所得模型能够较好地刻画镜头上的雨滴对图像的退化现象。在一般情况下,该算法能够有效分离叠加图像,校正变形,从而获得较好的恢复结果。
关键词:
图像恢复,
雨滴退化模型,
叠加图像分离
CLC Number:
LI Chang-Cheng, LUO Yu-Bin, ZHENG Xiao-Meng. Restoration Algorithm of Degraded Image Based on Raindrop Degradation Model[J]. Computer Engineering, 2011, 37(7): 253-254,267.
李长城, 罗予频, 郑晓明. 基于雨滴退化模型的退化图像恢复算法[J]. 计算机工程, 2011, 37(7): 253-254,267.