基于电子倍增CCD的夜视图像重建算法
关键词 夜视图像 电子倍增 CCD 图像重建 图像质量评估
Title Night-Vision Image Reconstruction Algorithm
Based on Electron Multiplying CCD
Abstract
Based on the low contrast and low SNR of night vision image, this article
is aimed at enhancing image quality from both hardware and software
aspects. When it comes to hardware, electron multiplying CCD amplifies weak
signals via multiplying registers and improves contrast and SNR as a
result. As to software, we use image reconstruction algorithm to recover
original information from polluted image. For major image recession is
caused by long shutter time and high ISO, the image reconstruction mainly
focuses on image de-blurring and image de-noising. The algorithms of
deconvolution, linear and non-linear filter and hard/soft-thresholding
wavelet de-noising have been adopted. Furthermore, Peak Signal-to-Noise
Ratio(PSNR) and Structural Similarity Index(SSIM) have been introduced to
objectively measure the variation of image quality and accordingly assess
algorithms’ performance.
Keywords Night-Vision Image EMCCD Image Reconstruction
Image Quality Assessment 目 次
1 引言 1
1.1 研究的发展情况和研究现状 1
1.2 课题研究的背景和意义 1
1.3 论文结构和主要研究内容 2
2 电子倍增 CCD 介绍 2
2.1 EMCCD 的结构以及工作原理 . 2
2.2 EMCCD 的噪声特性 . 4
3 图像重建 5
3.1 图像退化的数学模型 6
3.2 图像去模糊 7
3.2.1 逆卷积算法 . 8
3.2.2 文纳滤波方法 . 9
3.2.3 盲去卷积算法 . 10
3.3 图像去噪 11
3.3.1 空间域 . 11
3.3.2 频域降噪 . 13
3.3.3 小波域 . 14
4 图像质量评价标准 . 21
4.1 基于误差灵敏度的图像评价算法 22
4.2 基于结构相似度的图像质量评价 23
5 算法仿真 24
5.1 图像去模糊重建算法仿真 25
5.1.1 有先验知识条件下图像去散焦模糊 . 25
5.1.2 有先验知识条件下图像去运动模糊 26
5.1.3 无先验知识条件下图像去模糊 26
5.2 图像去噪重建算法仿真 27
5.2.1 中值滤波算法仿真 . 27
5.2.2 均值滤波算法仿真 28
5.2.3 频域低通滤波 29
5.2.4 硬阈值小波变换重建 30
5.2.5 软阈值小波变换重建 31