基于高分辨率光场成像技术的单相机三维流场测量系统研究

The proposed research seeks to develop a novel single-camera volumetric flow measurement technique. Based on the latest light field photography technology - the focused light field, a single-light-field-camera will be constructed by precisely encapsulating a CCD chip and a microlens array. A self-calibration algorithm will be simultaneously developed to compensate any misalignment between the CCD chip and the microlens array. According to linear Gaussian optics and the focused light field, a synthetic image generator (SIG) will be developed to simulate light field particle images produced by the single-light-field-camera. After accomplishing these three tasks, a 3D particle image reconstruction algorithm will be developed by combining the Full Resolution Light Field Rendering (FR-LFR) technique and the MART method. Based on the fact that seeding particles are sparsely distributed, computational efficiency of the reconstruction algorithm will be greatly improved by only reconstructing particle images. With the availability of 3D particle images, a Kalman Filter based particle tracking algorithm will be developed to compute one velocity vector from each particle image. The algorithm combines coarse cross-correlation, Gaussian fitting based particle identification with the Kalman Filter, which predicts particle's location at next frame and guide the high resolution particle tracking process. The 3D reconstruction and particle tracking algorithms will be numerically verified by a series of Lamb-Oseen vortex images generated by the SIG. The entire hardware system and light field image processing algorithms will be further validated by a vortex-ring impingement upon convex surface experiment. Compared to the current 3D PIV techniques, which feature complex system setup and expensive hardware system, the single-camera volumetric flow measurement technique will greatly simplify experimental procedures and reduce hardware cost, thereby provides researchers an easy-to-use, cost effective high resolution 3D flow measurement system to explore new possibilities in experimental fluid mechanics.

本项目旨在发展一种基于高分辨率光场成像原理的单相机三维瞬态流场测试技术。首先根据高分辨率光场技术，精密封装微透镜阵列与CCD芯片，构建单相机粒子光场图像采集系统；同时发展微透镜阵列自校准算法，以矫正CCD芯片与微透镜阵列之间的细微位置偏差。发展光场粒子图像数字合成平台，精确模拟示踪粒子所散射的激光经镜头、微透镜阵列到达CCD芯片后所形成的光场图像。采用全分辨光场渲染技术，结合流场示踪粒子稀疏均布的特性，发展基于乘积代数重建技术的高精度、高效率三维粒子图像重构算法。最后，开发基于Kalman滤波的三维粒子跟踪算法，实现对流场的高分辨三维瞬态速度场测量。以Lamb-Oseen vortex模型和圆形涡环-圆柱碰撞实验为研究对象，分别开展数值验证和模型实验，对所发展的硬件平台和先进算法进行系统的验证和修正。该技术将能极大地提升三维PIV测试技术的通用程度，提高流体力学实验研究水平及其理论深度。

本项目旨在发展一种基于高分辨率光场成像原理的单相机三维瞬态流场测试技术，主要研究内容包括：.1、光场成像分辨率理论分析；.2、光场相机设计及硬件封装；.3、光场畸变校准算法；.4、密集光线追踪光场三维重构算法；.5、GPU光场重建加速算法；.6、层析PIV对比验证实验；.7、超声速射流实验研究；.8、高超声速曲面压力实验研究。.取得了以下研究成果：.1、光场相机光学设计技术；.2、光场相机硬件封装技术，开发国内首套自主知识产权光场PIV相机；.3、高分辨、高效光场三维流场重构技术；.4、发展了全新一代的受限空间三维流场测试技术；.5、形成了完整的自主知识产权光场三维PIV测试技术；.6、国际上首次实现单相机三维超声速射流流场测量；.7、国际上首次实现单相机锥体模型三维尺寸及表面压力分布同步测量；.8、在“两机”专项压气机内部流场实验研究中获得初步应用。