Reflectance and light source estimation for indoor AR Applications
Reflectance and light source estimation for indoor AR Applications
复制标题
室内 AR 应用的反射率和光源估计
DOI：
10.1109/vr.2014.6802072
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
H. Takemura
中科院分区：
文献类型：
--
作者：
Alexander Plopski;T. Mashita;K. Kiyokawa;H. Takemura
我们提出了一种方法，使实时增强的环境组成的材料具有不同的纹理和反射特性，而不需要特定于应用程序的硬件或广泛的准备。我们的解决方案使用一组重建模型的RGB图像来优化反射参数和光位置。每个图像被分解成它的镜面反射和漫反射组件，我们估计多个光源的位置从镜面高光。环境存储在体素网格中，我们通过逆渲染优化每个体素的反射特性和颜色。我们验证了我们的方法与模拟环境，并提出了相应的重建环境的结果。
We present an approach which enables real-time augmentation of an environment composed of materials with different texture and reflectance properties without the need of application-specific hardware or extensive preparation. Our solution uses a set of RGB images of a reconstructed model to optimize the reflectance parameters and light location. Each image is decomposed into its specular and diffuse components and we estimate the location of multiple light sources from specular highlights. The environment is stored in a voxel grid and we optimize the reflectance properties and colour of each voxel through inverse rendering. We verify our approach with a simulated environment and present results from a corresponding reconstructed environment.