Abstract Image analysis technique is a common method to measure the void fraction and bubble size distribution in the bubbly flow. However, the image recognition algorithm will become invalid finally when the the projection overlaps of bubble become more heavily. In this paper, a new approach on three-dimensional (3D) reconstruction and parameters measurement of overlapping bubbles is developed based on laser scanning. The measurement principle of laser scanning is represented firstly, and appropriate scanning parameters are analyzed for 3D reconstruction at certain temporal and spatial resolution. Several image-preprocessing procedures are performed to extract the features effectively from the original slices of bubbly flow. A compensation method for slices offset is investigated to remove the unaligned feature noise of bubbles. An optimization algorithm based on image convolution of second-order differential averaging operator is proposed to remove the image noise caused by multiple exposures. In experiment, the 3D structure of bubbly flow field is reconstructed, and the void fraction and bubble size are measured with the analysis of distribution. The experimental results show that the method is valid for non-intrusive measurement and reconstruction of bubbly flow without the limitation of projection overlap. The optimization algorithms can effectively reduce the reconstruction deformation and improve the reliability, and the measurement error of void fraction is less than 13%, which indicates the relatively high characterization precision for the bubbly flow.

中文翻译：

---

基于激光扫描的重叠气泡实验研究

摘要 图像分析技术是测量气泡流中空隙率和气泡尺寸分布的常用方法。然而，当气泡的投影重叠越来越严重时，图像识别算法最终会失效。在本文中，基于激光扫描开发了一种用于重叠气泡的三维（3D）重建和参数测量的新方法。首先介绍了激光扫描的测量原理，分析了合适的扫描参数，以在一定的时空分辨率下进行3D重建。执行几个图像预处理程序以有效地从气泡流的原始切片中提取特征。研究了一种切片偏移补偿方法，以去除气泡的未对齐特征噪声。提出了一种基于二阶微分平均算子的图像卷积的优化算法，用于去除多次曝光引起的图像噪声。在实验中，重建了气泡流场的三维结构，并通过分布分析测量了空隙率和气泡尺寸。实验结果表明，该方法对于气泡流的非侵入式测量和重建是有效的，不受投影重叠的限制。优化算法能有效减少重构变形，提高可靠性，孔隙率测量误差小于13%，表明对气泡流具有较高的表征精度。