摘要动态范围指的是在一个场景中，最亮区域的亮度值和最暗区域的亮度值的比值。在实际的自然场景中动态范围可以达到14个数量级，人眼能够识别的最大动态范围是9个数量级。而传统的照相机和图像显示设备只能采集和显示不到3个数量级的低动态范围图像。这使得人们无法得到高质量的高动态范围图像。本文所研究的光电数字图像高动态范围成像技术，通过对相同场景不同曝光程度的多张图像合成，然后在保证其图像质量的情况下压缩其动态范围，从而得到能在低动态范围显示设备上输出的高质量图像。67457

在论文中，首先分析了人眼的视觉特性和高动态范围成像的数据特点，然后主要介绍了高动态范围成像技术的实现方法。通过配准来消除在合成前由于拍照时抖动产生轻微的移动或者旋转误差，然后进行HDR响应曲线的标定，接下来对采集到的多幅同一场景不同曝光量的图像进行合成，其中主要分为灰度图像的合成和彩色图像的合成，最后通过色调映射使结果可以在低动态范围显示设备显示出来。并对最终的实验结果和所用的算法进行分析。

毕业论文关键词：高动态范围、配准、响应曲线、色调映射

Title    Optical Digital Images High Dynamic Imaging Technology       

Abstract

Dynamic range refers to a scene, the ratio of the luminance values of the brightest area and the darkest areas of the luminance value. Natural scenes in the actual dynamic range up to 14 orders of magnitude, the human eye can recognize the maximum dynamic range of 9 orders of magnitude. The traditional camera and the image display apparatus can only capture and display the low dynamic range image less than three orders of magnitude. This makes it impossible to obtain high-quality high dynamic range scenes. Studied this optical digital image high dynamic range imaging technology, the same scene with different exposure levels of multiple image synthesis, and ensure image quality in the case of dynamic range compression, to obtain a low dynamic range in the display device high-quality image output.

In the paper, we first analyzed the characteristics of human visual and high dynamic range imaging data, and then introduce how to realize the high dynamic range imaging techniques. Through registration to eliminate the synthesis ago because of a slight camera shake that occurs when moving or rotation error, then calibration HDR response curve, next though collected pieces of the same scene with different exposure amounts image synthesis, which is pided into grayscale and color image synthesis synthesis, and finally through the tone mapping results in a low dynamic range display device display. And analysis the final experimental results and algorithms used .

Keywords  High dynamic range, Registration, Response curve, Tone mapping

目   次

1  引言1

1.1  研究背景1

1.2  高动态范围成像技术简介1

1.3 本文研究内容4

2  高动态成像基础6

 2.1 人类视觉系统6

2.1.1  对比灵敏度 6

2.1.2  分辨率 6

2.1.3  同时对比现象 6

2.1.4  人眼的视觉特性 7

2.2  提高输入信号的动态范围 7

2.3  高动态范围图像的数据分布10

2.4  输出信号动态范围的压缩12

3  高动态范围成像的算法和实现13

3.1  高动态范围成像实现步骤分析13

3.2  图像配准13

3.2.1  图像配准的思路14

3.2.2  图像二值化与配准14

3.3  响应曲线的标定16

3.3.1  Mann and R.W.Picard算法 16