基于傅里叶变换高光谱成像的BRDFs信息获取与颜色再现

BRDFs (Bidirectional Reflectance Distribution Function Spectrum) information acquisition and color reproduction is the key to the reproduction of real object appearance and color, it is important for the representation of object appearance, realistic surface rendering, 3D object surface color reproduction and so on. It is becoming the research focus in the field of color science and technology. According to the fact that the acquisition system suffer from the problems of less spectral channels, narrow band range, and low algorithm accuracy of spectral reconstruction and data compression, in this research a novel approach focusing on the principles and methods for BRDFs Fourier transform hyperspectral imaging data acquisition, spectral reconstruction, representation model and data compression for color reproduction is proposed . With the aim of solving the problems of low spectral resolution, image registration and dependent on sample training for spectral reconstruction accuracy, the principle and methods of BRDFs Fourier transform imaging acquisition system is deduced, and the corresponding algorithms for the data acquisition, spectral reconstruction, representation model and full spectrum restoration. Afterward, in view of the existing problems in storage and computation of mass BRDFs data, based on the mechanism of human vision perception, the novel data BRDFs compression method is researched on the condition of ensuring the accuracy of spectrum and color. On the basis, it provides effective theoretical support and technical support for BRDFs full surface representation and true three-dimensional rendering of spectral color reproduction.

BRDFs信息获取和颜色重现是真实物体外貌和颜色信息还原及绘制的关键，在物体全外貌表征、真实感物体表面绘制、三维物体表面颜色再现等方面有非常重要的应用，正成为颜色科学与技术领域的研究热点。本项目拟探索BRDFs傅里叶变换高光谱成像数据获取、全光谱复原、表征模型和颜色再现数据压缩的原理和方法。针对传统BRDFs多光谱成像获取系统需图像配准、光谱分辨率低、光谱重建精度依赖训练样本等问题，探索构建宽波段傅里叶变换高光谱BRDFs成像采集系统的原理和方法；并研究相应的数据获取、光谱重建、BRDFs表征模型和全光谱复原算法；针对高光谱BRDFs数据存储和计算量大的问题，依据人眼睛视觉感知机理，在保证光谱精度和颜色再现精度的条件下，探索海量BRDFs数据进行光谱维和空间角度维数据压缩的新方法。以此为依据，为BRDFs全外貌表征和真三维绘制的光谱颜色再现提供有效的理论支持和技术支撑。

BRDFs信息获取和颜色重现是三维物体外貌和颜色信息还原绘制的关键，在物体全外貌表征、真实物体表面绘制、三维物体颜色再现等方面具有非常重要的应用，已经成为颜色科学技术领域的研究热点。本项目中构建BRDFs傅里叶变换高光谱成像系统，设计加工了多自由度旋转载物平台，以平台的多自由角度旋转，实现目标的成像式多角度采集双向反射光谱分布（BRDFs）；对采样干涉图像数据以快速光谱重建算法得到了二维成像目标的BRDFs全光谱信息，并进行了标准照明体的颜色重建；重点分析了光源照明对物体BRDF颜色的影响；以空间维度的光谱聚类分析和光谱维度的因式分解、成分分析等方法进行空间和光谱维度降维，以克服采集和重建得到的目标BRDFs数据量大，不易于存取和处理的缺点。本项目系统研究了目标BRDFs的获取、光谱重建、颜色再现和数据压缩表征方法；分析了光源特性对BRDFs颜色的影响与改进；以提取光谱颜色特性和空间光谱降维的方法进行BRDFs海量数据的稀疏表征；充分利用了傅里叶高光谱成像系统的高光谱分辨率优势，实现了基于颜色和光谱质量的目标BRDFs信息压缩表征。发表学术论文8篇，其中4篇SCI检索，3篇EI检索；申请国家发明专利6件，其中3件已授权。部分研究内容拓展性的应用于视觉密度测量，获得2018年度“中国计量测试学会科学技术进步奖”一等奖和2019年度“北京市科学技术进步奖”二等奖。