-
摘要在EMCCD发展以及彩色夜视的应用需求背景下，本文研究了基于EMCCD的彩色夜视算法。首先，文章介绍了国内外EMCCD和彩色夜视技术的发展，概括了EMCCD成像原理和已有的彩色夜视算法。之后，对于Bayer 彩色滤波器采样图像，提出了基于区域色调一致性的插值算法和基于边缘的插值修正算法，与传统的双线性插值进行了各种分量的信噪比对比，MATLAB仿真结果显示本文的算法在色调改善上要优于双线性插值。对于微光与红外的融合处理，本文采用一种优于MATLAB里中值滤波函数的中值算法对微光图像预处理并利用共生矩阵的二次统计量对融合图像评价。在小波域引入了最大值法、均值法、区域分析法、主成分分析法对图像进行融合；同时，也对比了Laplace和FSD金字塔的融合算法效果。在一系列融合算法的基础上更进一步提出了基于小波域统计量传递和直方图匹配的色彩传递算法，仿真表明经色彩传递后的图像在能量，对比度，清晰度上都得到了提高。33552
- 上一篇：MEMS角度传感器的测量电路与ARM控制软件
- 下一篇：基于嵌入式系统的CF卡读写操作设计+驱动程序
毕业论文关键词 EMCCD 彩色夜视 插值 融合 评价 MATLAB
毕业设计说明书外文摘要
Title Color Night Vision Algorithms Based on EMCCD
Abstract
Under the development of EMCCD and the application of color night vision,this paper studies color night vision algorithms based on EMCCD.Firstly,the paper introduces the development of EMCCD and color night vision technology at home and abroad and summarizes the imaging principle of EMCCD and the existing algorithms of color night vision.Then, we propose a interpolation algorithm based on hue consistency and a correction algorithm based on edge for Bayer color filter sampling images.By the comparison with traditional bilinear interpolation about the signal to noise ratio of various components,MATLAB simulation results show that this algorithm is better than bilinear interpolation in hue improvement.For the fusion of low light level images and infrared images,this paper adopts a median filtering algorithm in the low light level image preprocessing.This algorithm is better than the median filter function of MATLAB.We use second degree statistics of co-occurrence matrix to evaluate fused images.The maximum value method, average method ,regional analysis method and principal component analysis method are introduced in the wavelet domain for image fusion.Meanwhile, the fusion algorithm of Laplace and FSD in Pyramid are compared.The color transfer algorithms based on statistics transfer in the wavelet domain and histogram matching are proposed on the basis of a series of fusion algorithms.Simulation results show that the energy, contrast and resolution of images by color transfer can be improved.
Keywords EMCCD;color night vision;interpolation;fusion;evaluation;MATLAB
目 次
1 绪论1
1.1 EMCCD简介1
1.2 EMCCD的研究意义和应用1
1.3 彩色夜视技术的发展及研究意义2
1.4 本论文主要的研究工作及布局3
2 EMCCD成像系统5
2.1 EMCCD的结构与工作过程5
2.2 EMCCD的倍增原理5
2.3 EMCCD的噪声6
3 彩色夜视算法7
3.1 基于Bayer滤波器的彩色夜视算法7
3.2 已有的红外与微光融合算法7
4 彩色图像评价9
4.1 插值图像评价9
4.2 融合图像评价9
5 基于Bayer滤波器的彩色夜视算法实现11
5.1 彩色滤波阵列CFA11
5.2 常见的插值算法实现11
5.3 实验结果及分析14
6 微光与红外彩色夜视融合算法实现17
6.1 颜色空间17
6.2 红外与微光图像预处理17