流场的三维速度测量是研究热流体湍流流动的重要技术手段之一。本项目将发展一种基于纹影成像的高速三维粒子追踪测速技术。该技术采用两套二维纹影成像光路，在测试流场中加入粒子，利用两台高速相机同时采集两个角度的纹影图像；基于双视角立体成像和重建原理，获取粒子的三维坐标分布；结合先进粒子追踪算法，包括极线约束、多帧追踪、粒子分布拓扑结构相似性、空间-时间耦合验证等，追踪粒子拉格朗日运动轨迹，实现具有高频响应的高精度三维体空间速度测量。将采用合成图像及水槽构建三维热羽流流动两种方法，对所发展的方法进行验证，对测量精度进行定量化表征。该系统基于纹影成像原理，与采用激光的三维测速方法相比，在硬件上仅需使用普通光源，成本更低，应用更为方便灵活，具有更高的精度，在大区域高速测量方面具有明显优势。在获得三维速度场的同时，所获取的纹影图像还可表征流场中温度或密度的变化，在热流体相关领域具有广阔的应用前景。

The three-dimensional velocity measurement is an essential technique for experimental investigation of turbulent thermal fluid dynamics. In current proposal, a time-resolved three-dimensional particle tracing velocimetry (3D-PTV) is developed based on a novel stereoscopic schlieren imaging system. The system includes two standard schlieren configurations sharing one LED light source. The test section, which is seeded with tracing particles, is placed at the intersection of the two converging light beams. In this way, the stereoscopic information of the seeding particles is preserved on the two schlieren images taken at different angles and may be recovered through 3D reconstruction. An advanced strategy for accurate 3D particle tracing will be established by integrating several state-of-the-art algorithms, such as epipolar geometry constraint, multi-frame particle tracing, particle distribution similarity, temporal-spatial integrated validation, etc. The accuracy of the 3D velocity measurement will be evaluated systematically using synthetic images as well as thermal plume tests. Comparing with the laser-based volumetric velocimetry methods, the proposed technique only requires normal LED light, which is much cheaper, easier for implementation, more accurate and is superior for high frequency test in large areas. Moreover, besides the 3D velocity information, the schlieren images also contain rich information about temperature or density variations. The schliern-based 3D-PTV proposed in current project is believed to have wide applications for fundamental and applied research in thermal fluid mechanics and heat transfer disciplines.