气动热障图像校正方法研究
文献类型：学位论文
| 作者 | 李征威
|
| 学位类别 | 硕士 |
| 答辩日期 | 2016-05-25 |
| 授予单位 | 中国科学院沈阳自动化研究所 |
| 导师 | 向伟 ; 徐保树 |
| 关键词 | 气动热障 流场 热辐射 红外成像 图像校正 |
| 其他题名 | Research on Correction Method of Aerodynamic Thermal Barrier Image |
| 学位专业 | 控制工程 |
| 中文摘要 | 本文的研究内容是气动热障图像的校正方法，大致分三个部分展开。首先，分析了气动加热带来的复杂的物理环境；然后，分析了气动热障条件下红外成像退化机理并给出了退化图像模型；最后，根据退化图像特点提出了气动热障图像的校正方法。以下分别阐述本文的主要工作：1. 分析了2.5马赫条件下某飞行器的气动热环境。使用有限元法结合ANSYS仿真工具得到头罩周围流场的温度、压力、分布状态等参数，为研究气动热环境下红外成像的退化机理打下基础。2. 分析了气动热障图像的退化机理。首先，提出了一种分析光线传输过程的气体等效透镜方法，结果表明气动流场对红外光线传输的影响基本可以忽略。然后，根据热辐射源的特点，得出气动热辐射噪声在退化图像上的表现形式。最后，建立了气动热障图像退化模型。3. 提出了四种气动热障图像校正方法。基于小波多尺度分解的方法：根据气动热辐射的低频分布特点，对退化图像多个尺度分解并抑制低频信息后重构，再对重构图像加权融合。基于小波改进暗原色的方法：根据气动热障图像和雾天图像的退化模型之间的相似性，使用图像的小波分解后的低频信息代替暗原色，并将原雾天成像退化模型中固定的大气光值改为区域变化的热辐射强度，这样原有算法基于暗原色的单幅图像去雾算法改进为气动热障图像校正方法，从气动热辐射掩盖下复原出清晰图像。基于引导滤波的方法：提出一种新的大窗口下、增大正则项的自适应引导滤波参数选择方法，实现了对退化图像中非均匀、变化缓慢的热辐射噪声的提取；然后在小窗口下使用引导滤波器增强去热辐射后图像。基于加权最小二乘滤波器的方法：克服了使用引导滤波器提取气动热辐射时产生的“光晕”现象。综合四种校正方法，基于引导滤波的气动热障图像校正方法，对热辐射噪声的去除效果显著，校正后图像的清晰度和对比度好，且算法的复杂度低，处理速度快，易于工程实现。综上所述，本文的工作更进一步加深了对气动热障图像退化机理的认识，提出了高效的退化图像校正方法。 |
| 英文摘要 | Content of this paper, broadly divided into three parts, is the correction method of aerodynamic thermal barrier image. Firstly, we simulate complex physical environment caused by aerodynamic heating; then, we analyze degradation mechanisms of infrared imaging under the aerodynamic thermal barrier and give degraded image model; finally, according to the characteristics of the degraded image, we propose correction methods of pneumatic thermal barrier mage. The following describes are the major work of this paper. 1. This paper analysis aerodynamic heating environment of aircraft with 2.5Ma. We use the finite element method with ANSYS simulation tool to give temperature, pressure, distribution and other parameters of flow field around the hood, these lay the foundation of the study of degradation mechanisms of infrared imaging under aerodynamic heating environment. 2. This paper analysis mechanisms of aerodynamic thermal barrier degraded image. First, we present an equivalent lens method to analysis the influence of the aerodynamic flow field on the light transmission can be neglected. According to characteristics of aerodynamic heat radiation, we obtain expressions of aerodynamic thermal radiation noise on degraded image. Finally, the degradation model of aerodynamic thermal barrier image is established. 3. This article proposes three aerodynamic thermal barrier image correction methods. The method based on wavelet multiscale decomposition：According to the characteristic that aerodynamic thermal radiation distributed in low-frequency, we use wavelet to decompose the degraded image at multiple scales and reconstruct image after inhibiting information, then the reconstructed image is weighted fusion. The method of improved image haze removal based on wavelet: According similarity of degradation model between aerodynamic thermal barrier image and haze image, we use the reconstructed low-frequency image after wavelet decomposition to replace dark-channel pixel and the stable value to the atmosphere light of haze image model will become thermal radiation intensity changed in different regions. The original haze removal method based on dark-channel pixel transform converted to aerodynamic thermal barrier image correction method based on wavelet, clear image is recovered from aerodynamic heat radiation covering. The method based on guided filter: we proposed a new adaptive parameters setting method that increases regularization under large window, the inhomogeneous changed slowly thermal radiation are extracted and restrained in degraded images. Then we use guided filter to enhance images removed thermal radiation under small window. The method based on weighted least squares filter(WLS Filter): it overcomes "Halo" phenomenon produced when using guided filter to extract aerodynamic heat radiation. Comparing four kinds of aerodynamic thermal barrier image correction methods, the method removal thermal radiation noise based on guided filter has significant effect, the sharpness and contrast of restored image are better, the complexity of the algorithm is low, and the method has fast processing speed, easy to achieve in engineering. In summary, the work of this paper further deepens the understanding of the degradation mechanism of aerodynamic thermal barrier image and proposes an efficient correction method of degradation image. |
| 语种 | 中文 |
| 产权排序 | 1 |
| 页码 | 67页 |
| 源URL | [http://ir.sia.cn/handle/173321/19629]  |
| 专题 | 沈阳自动化研究所_光电信息技术研究室 |
| 推荐引用方式 GB/T 7714 | 李征威. 气动热障图像校正方法研究[D]. 中国科学院沈阳自动化研究所. 2016. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。