燃料电池工作过程三维成像算法研究与实现

发布时间：2018-05-21 03:36

本文选题：质子交换膜燃料电池 + 三维可视化　；参考：《北京邮电大学》2017年硕士论文

【摘要】：近年来,随着人们对清洁能源越来越多的关注,科学家们对燃料电池的研究也逐渐深入。燃料电池工作过程中的水管理是影响燃料电池工作效率的重要因素之一,当水量过多时会导致电池内部的质子交换膜被“淹死”;而水量过少,则会导致质子交换膜“干死”;这两种情况都会使得燃料电池不能正常工作。如何能够显式地观察燃料电池中的水、气状态,对于燃料电池的研发与改进具有十分重要的意义,本课题从图像处理的角度,以三维可视化图像的方式来显示燃料电池内部结构,并且在图像中确定水和气体的分布状态。文中利用CT层析透照技术对工作过程中会出现的几种不同水量情况的质子交换膜燃料电池进行扫描,获取大量燃料电池层析图像,用于完成燃料电池内部结构的重构和呈现。本文首先对这些层析图像进行了初步的灰度分析与研究,研究发现燃料电池内部不同成分灰度值显示有较明确的灰度区间,并且各个区间之间有明显的分割阈值,而这些阈值与灰度区间可以为后续的燃料电池成分分割提供很好的分割依据。由于获取到的层析图像为射线图像,存在噪声点等影响图像质量的因素,文中重点研究了图像滤波方法,通过比较中值滤波、均值滤波、高斯滤波这几种降噪方法在燃料电池层析图像中的降噪效果,发现高斯滤波算法更适用于层析图像的降噪处理操作。为了更好地进行电池内部的成分划分,通过对比各种边缘检测算法在燃料电池层析图像中的检测结果,提出了一种基于Canny边缘检测算子的二维图像区域分割方法,利用各成分的灰度阈值在燃料电池层析图像中对电池结构以及水和气体等成分进行边缘检测与区域划分,通过对燃料电池内部不同灰度区域进行色彩标记,在层析图像中清晰的判定燃料电池内部水和气体成分。在此基础上,文章还对光线投射算法进行了优化,在新的算法中,背景数据被忽略,只需要提取层析图像分割结果中燃料电池的目标数据,大大减少了原算法的计算量,并且通过对不同灰度区间的色彩以及透明度进行设置来实现三维图像的区域划分以及透视效果。为了更好地实现三维效果的显示,本课题利用VTK (Visualization Toolkit)可视化工具进行三维图像体数据的分析以及三维图像的渲染,同时采用硬件加速机制来提高算法效率。本文将CT图像三维可视化技术应用于燃料电池工作过程中水管理的研究中,利用改进的图像分割与三维可视化算法对获取到的燃料电池层析图像进行可视化操作,通过对图像不透明度以及色彩的设置,实现对燃料电池内部的不同结构进行差别标记,很好的完成了燃料电池三维图像的成分分割工作,达到了在水管理研究中对燃料电池内部成分状态分析的要求,并且通过对光线投射算法的改进,在算法效率提高方面效果明显。
[Abstract]:In recent years, as people pay more and more attention to clean energy, scientists' research on fuel cells has gradually deepened. Water management is one of the important factors that affect the efficiency of the fuel cell. When the water is too much, the proton exchange membrane inside the fuel cell will be drowned, but too little water will lead to the "dry death" of the membrane. In both cases, fuel cells don't work properly. How to observe the state of water and gas in fuel cell is very important for the research and improvement of fuel cell. The internal structure of the fuel cell is displayed by a three-dimensional visual image, and the distribution of water and gas is determined in the image. In this paper, CT tomography was used to scan several proton exchange membrane fuel cells (PEMFC) with different water content in the process of operation, and a large number of chromatographic images were obtained to complete the reconstruction and presentation of the internal structure of the fuel cell. In this paper, we analyze and study the gray scale of these chromatographic images, and find that the gray value of different components of fuel cell has a clear gray range, and there is an obvious segmentation threshold between the different sections. These thresholds and gray levels can provide a good basis for the subsequent fuel cell component segmentation. Due to the fact that the computed image is a radiographic image and there are some factors affecting the image quality, such as noise points, this paper focuses on the image filtering method, and compares the median filtering, the mean filter and so on. The effect of Gao Si filtering on the noise reduction of fuel cell tomography image is discussed. It is found that Gao Si filtering algorithm is more suitable for the noise reduction operation of the tomographic image. In order to divide the components of the battery better, a two-dimensional image region segmentation method based on Canny edge detection operator is proposed by comparing the detection results of various edge detection algorithms in the fuel cell tomography image. The gray threshold of each component is used to detect and divide the edge of the structure of the fuel cell and the components of water and gas in the chromatographic image of the fuel cell, and the different grayscale areas of the fuel cell are labeled with color. The internal water and gas components of the fuel cell are clearly determined in the chromatographic image. On this basis, the ray-casting algorithm is optimized. In the new algorithm, the background data is ignored, only the target data of fuel cell is extracted from the segmentation result of the tomography image, which greatly reduces the computational complexity of the original algorithm. By setting the color and transparency of different gray levels, the region division and perspective effect of 3D images are realized. In order to display the 3D effect better, the visualization tool of VTK Visualization Toolkit is used to analyze the volume data of 3D images and render 3D images. At the same time, the hardware acceleration mechanism is used to improve the efficiency of the algorithm. In this paper, the 3D visualization technology of CT image is applied to the research of water management in the process of fuel cell operation, and the improved image segmentation and 3D visualization algorithm is used to visualize the obtained fuel cell tomography image. By setting the image opacity and color, the different structure of the fuel cell can be marked differently, and the component segmentation of the three-dimensional image of the fuel cell can be completed well. It meets the requirement of fuel cell internal component state analysis in water management research, and through the improvement of ray casting algorithm, the efficiency of the algorithm is improved obviously.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TM911.4

【参考文献】