基于光谱连接空间的彩色相机光谱重构研究

Spectral Reconstruction from Trichromatic Digital Camera Responses Based on Interim Connection Space for Spectral Data

针对彩色相机三通道响应值重构光谱反射率精度低的问题,提出了基于光谱连接空间的彩色相机光谱反射率重构方法。首先通过光谱反射率已知的训练样本集和多项式拟合方法建立相机响应值到光谱连接空间的转换矩阵,然后利用该矩阵将待重构样本的相机响应值映射到光谱连接空间,最后选用合适的光谱重构算法在光谱连接空间内实现光谱反射率重构,并利用色度误差和光谱误差两个指标对重构结果进行评价。在上述过程中,鉴于转换矩阵的重要性,采用了基于反距离加权的最小二乘法计算转换矩阵以提高相机响应值到光谱连接空间的转换精度。实验结果表明:本文方法切实可行且精度可靠,与基于彩色相机三通道响应值的光谱重构方法相比,色度重构精度和光谱重构精度均显著提高,平均色差和谱差分别为1.145 2和0.010 3,可在较大程度上满足数字典藏、高保真颜色复制等的需要。

Aiming at the problem of poor spectral reconstruction accuracy from trichromatic camera responses,a method was presented for spectral reflectance reconstruction based on the interim connection space（ ICS） for spectral data and trichromatic digital camera. Firstly,the training samples with known spectral reflectance and polynomial fitting method were utilized to establish a transform matrix from the trichromatic camera responses to the ICS response values. Then,the target samples or pixels were mapped to the ICS by means of the transform matrix. Finally,spectral reflectance reconstruction of each sample or pixel was implemented via an appropriate spectral reconstruction algorithm using ICS response values and the reconstruction accuracy is evaluated in terms of chromaticity difference and spectral root mean square error. In addition,the inverse distance weighted least square method was applied to calculate the transform matrix to improve the conversion accuracy from the camera response to the ICS response value. The experimental results show that the proposed method is feasible and has reliable precision. Compared to the spectral reflectance reconstruction from trichromatic camera responses,the chromaticity reconstruction accuracy and spectral reconstruction accuracy of the proposed method are significantly improved. The mean chromaticity difference and mean spectral root mean square error are 1. 145 2 and 0. 010 3,respectively. It can meetthe needs of digital archives,high fidelity color reproduction,etc.