高超声速三维流场光流粒子图像测速新原理

The measurement of hypersonic three-dimensional flow field is the basis for the development of national strategic important equipment such as hypersonic vehicle. Particle image velocimetry (PIV) is an effective method for full field non-contact measurement of hypersonic three-dimensional flow field. The principle of PIV measurement is the most important factor affecting the accuracy and resolution of PIV measurement in hypersonic three-dimensional flow field. The existing PIV measurement principle based on cross correlation can’t be applied to the precise measurement of high spatial resolution of the fine structure of hypersonic three-dimensional flow field, which has become a bottleneck restricting the independent research and development of high performance three-dimensional PIV measurement instruments in China. This project plans to break through weak optical imaging and image enhancement technology, study on low contrast two-dimensional particle image enhancement method, build multi visual spatial geometric model and optical path distortion model and research on accurate calibration method of multi-camera vision, reveal the influence law of particle three-dimensional reconstruction precision; propose a three-dimensional particle image speed estimation method based on stereoscopic optical flow, study on the determinant condition and detection method of three-dimensional velocity outlier; establish a new principle of three-dimensional optical flow PIV measurement for high hypersonic flow field and realize high spatial resolution measurement of hypersonic flow field. This project has very important theoretical significance and application value for the application of three-dimensional PIV in hypersonic vehicle wind tunnel test.

高超声速三维流场测量是研制高超声速飞行器等国家战略重大装备的基础，粒子图像测速（Particle Image Velocimetry，简称PIV测量）是高超声速三维流场全场非接触测量的有效方法，PIV测量原理是影响高超声速三维流场PIV测量精度和测量分辨率的基础和关键。现有基于互相关的PIV测量原理无法适用于高超声速三维流场精细结构的高空间分辨率精确测量，已成为制约我国高性能PIV测量仪器自主研发的瓶颈。本项目拟突破微弱光成像与像增强技术，研究低对比度二维粒子图像增强方法；建立多目视觉空间几何模型与光路畸变模型，研究多目视觉精确标定方法，揭示粒子三维重构精度影响规律；提出基于立体光流的三维图像速度估计方法，研究三维速度误矢量判定条件和检测方法；建立高超声速三维流场光流PIV测量新原理，实现高空间分辨率精确测量；在高超声速飞行器风洞风速测量中验证应用，具有十分重要的理论意义和应用。

高超声速飞行器在未来军事、政治和经济中将发挥重要的战略作用。高超声风洞流场测量是研究高超声速飞行器气动力与气动热问题的基础。目前PIV技术已成为高超声速流场速度测量的重要测试技术，其中PIV测量原理是PIV技术测量精度的关键，然而现有基于互相关的PIV测量原理，存在粒子图像信噪比低、测量分辨率低、测量精度不足等问题，无法适应于高超声速三维流场的高分辨率高精度测量。本项目开展了高超声速流场中层析粒子重建模型研究，三维速度测量算法研究，流场超分辨率重建模型研究等研究工作，提出了基于变分去卷积模型的粒子场修复算法、基于流场分割的变分光流算法、基于速度分解定理的变分光流算法、基于反投影网络的流场超分辨率算法、三维流场PIV测量精度与测量分辨率的误差模型等一系列粒子图像测速算法研究。本项目主要研究成果：在权威SCI期刊和顶级国际学术会议上发表论文18篇；培养3名博士研究生、17名硕士研究生。申请国家专利12项，其中7项已授权。设计了适用于大型高超声速风洞的相机保护罩，研制一套多目视觉系统精密安装与调整装置，并在中国绵阳空气动力研究与发展中心Φ0.5米7Ma高超声速风洞进行了PIV测量应用验证，提升了高超声速流场PIV测量精度，具有十分重要的理论意义和应用价值。