Research And Implementation Of 3D Imaging Algorithm For Fuel Cell In Working Process

With more and more attention on clean energy, researchers have gradually deepened the research of fuel cell. The amount of water inside the fuel cell in work is an important factor affecting the efficiency of fuel cell. When the water is too much, the proton exchange membrane would be "drowned", and when the water is too small circumstances will lead to proton exchange membrane "dry death", both of them will make the fuel cell can not work properly. Therefore how to observe the water vapor state of the fuel cell is very important for the research and development of proton exchange membrane fuel cells. Based on the research of this field, this paper uses the three-dimensional visualization method to display the internal structure of fuel cell from the perspective of image processing, and show the distribution of water and gas inside.The CT tomography technique was used to scan the proton exchange membrane fuel cells with different amounts of water in the research, and a large number of fuel cell tomography images were acquired for reconstructing and presenting the internal structure of the fuel cell. In the process of analyzing the gray value of the tomography images, we found that the gray value of different components in the fuel cell showed a clear gray-scale interval, and each segment has a clear threshold, which could be a great basis of image segmentation.Because the obtained tomographic images are ray images, there are noise points and other factors that affect the quality of images. To deal with these problems,the paper focuses on some image filtering methods by comparing the median filtering, mean filtering, Gaussian filtering and their work on noise reduction in the fuel cell images. The results show that Gaussian filtering algorithm is more suitable for noise reduction of fuel cell tomographic images. In order to achieve a better partition in the internal components of the battery, 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 images. And according to the gray value of images, the water and gas in the tomographic image of fuel cells could be clearly distinguish through color tag. Besides, the article also optimized the ray casting algorithm. The background data is neglected in the new algorithm, and only the target data of the fuel cell in the segmented image is extracted, which greatly reduces the calculation. The color and transparency settings of different gray levels help to realize the 3D image segmentation and perspective effect. To get a better display of 3D images, this paper uses VTK (Visualization Toolkit)tool to analyze the three-dimensional volume data and display the 3D rendering effect. At the same time, hardware acceleration mechanism is used to improve the efficiency of ray casting algorithm.Using the image segmentation and three-dimensional visualization algorithm of this paper, the visualization operation of the obtained fuel cell tomography images is carried out. And the fuel structure is well marked by different opacity and color of the fuel cells. The three-dimensional image segmentation of the fuel cells achieves the requirement of analyzing the internal state of the fuel cell in different work state, and through the improvement of ray casting algorithm, it has achieved outstanding results in improving the efficiency of the algorithm.