基于高速摄影的气液两相流图像处理和重建算法研究

发布时间：2018-04-17 01:08

本文选题：气液两相流 + 小波去噪　；参考：《天津大学》2016年硕士论文

【摘要】：最近几十年,随着科学技术不断推动,气液两相流广泛应用于石油、医学、制造、航天、生物等多方面领域。气液两相流测量技术得到了广泛重视,从二维测量逐渐发展为三维测量,为气液两相流研究提供了更广阔的研究平台。图像处理与重建,以及可视化测量成为主要的研究手段之一,并不断取得理论及技术成果。本文基于高速摄影法,通过实验所得图像测量气液两相流参数,搭建了基于OpenGL的软件平台,针对气液两相流特征提出了三维重建算法。同时设计三维重建结果的评估方法,通过对图像处理结果深入研究和比较,实现算法优化和误差补偿。其中本课题相关研究工作总结如下:1、设计和优化了图像处理算法。图像处理算法设计,包括图像灰度、差影、反色、中值滤波、阈值分割等。针对课题中图像去噪难点进行深入分析。对几种去噪算法特性比较,基于小波变换理论分析,通过小波去噪对中值滤波以及阈值分割进行算法优化。2、研究和分析图像匹配技术国内外现状,实现快速、精确的图像匹配算法。结合课题实验研究,针对匹配算法的高精度需求,采用基于轮廓和对数极坐标变换的匹配算法。基于气泡轮廓特征,使匹配速度大幅提高。对数极坐标变换适用于实验课题下高密度气泡的基于轮廓特征匹配,通过设计实现图像匹配评价方法,表明对数极坐标匹配精准度更高、匹配成功率更高。3、基于图像预处理和图像匹配算法,设计图像插值算法。通过对高密度多相流气泡特性分析,实验验证多项式插值可满足精度要求,并取得较好的可视化效果。设计了图像插值评价方法,通过结果数据分析,多项式插值精度高、程序运行时间短。4、基于OpenGL三维重建平台设计。通过对实验课题深入研究,OpenGL软件平台适用于气液两相流三维重建算法研究以及可视化实现。根据实验参数精度和可视化需求,平台实现了图像窗口大小、旋转变换、背景渲染、多视角视图、硬件控制等软件设计。
[Abstract]:In recent decades, with the development of science and technology, gas-liquid two-phase flow has been widely used in many fields such as petroleum, medicine, manufacture, spaceflight, biology and so on.The technology of gas-liquid two-phase flow measurement has been paid more and more attention, from two-dimensional measurement to three-dimensional measurement, which provides a broader research platform for gas-liquid two-phase flow research.Image processing and reconstruction, as well as visual measurement, have become one of the main research methods, and continue to achieve theoretical and technical results.In this paper, based on high-speed photography, the parameters of gas-liquid two-phase flow are measured by experimental images, a software platform based on OpenGL is built, and a three-dimensional reconstruction algorithm for gas-liquid two-phase flow characteristics is proposed.At the same time, the evaluation method of 3D reconstruction results is designed, and the algorithm optimization and error compensation are realized through the in-depth study and comparison of image processing results.The related research work is summarized as follows: 1. The image processing algorithm is designed and optimized.Image processing algorithm design, including image grayscale, differential image, inverse color, median filter, threshold segmentation and so on.The difficulty of image denoising is analyzed deeply.The characteristics of several denoising algorithms are compared. Based on the wavelet transform theory, the median filtering and threshold segmentation are optimized by wavelet denoising. The present situation of image matching technology is studied and analyzed.Accurate image matching algorithm.The matching algorithm based on contour and logarithmic polar transformation is adopted to meet the high precision requirement of the matching algorithm.Based on the bubble profile feature, the matching speed is greatly improved.Logarithmic polar coordinate transformation is suitable for contour feature matching of high-density bubbles under experimental project. The method of image matching evaluation is designed and realized, which shows that logarithmic polar coordinate matching is more accurate.The success rate of matching is higher. 3. Based on image preprocessing and image matching algorithm, image interpolation algorithm is designed.By analyzing the bubble characteristics of high density multiphase flow, the experimental results show that polynomial interpolation can meet the precision requirements and achieve a good visualization effect.The image interpolation evaluation method is designed. Through the analysis of the result data, the polynomial interpolation accuracy is high, the program running time is short. 4. The program is designed based on OpenGL 3D reconstruction platform.The OpenGL software platform is suitable for 3D reconstruction and visualization of gas-liquid two-phase flow.According to the precision of experimental parameters and the requirement of visualization, the platform realizes the software design of image window size, rotation transformation, background rendering, multi-view view, hardware control and so on.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】