Deriving Nonnegative Spectral Transmittance Based on PCA-LDA
Hans Journal of Biomedicine , Volume 11, pp 14-22; doi:10.12677/hjbm.2021.111003
Abstract: 本文针对高光谱成像，提出了一种基于主成分分析(Principal Component Analysis, PCA)和线性判别分析(Linear Discriminant Analysis, LDA)生成非负光谱透过率的方法。该非负光谱透过率可应用于可编程高光谱成像系统的光学成像结果等效于PCA-LDA数字模型后处理的高光谱数据的结果。该方法通过滤除训练过程中的负值，无需补偿透过比的二次采集，可以直接针对高光谱数据应用，即高光谱数据采集和PCA-LDA后处理可以通过光学成像的物理过程一次性同时实现，有助于在光学和遥感领域的信息应用中实现更智能便捷的光学检测与传感。 In this paper, a method to derive nonnegative spectral transmittance based on Principal Compo-nent Analysis (PCA) and Linear Discriminant Analysis (LDA) is proposed for hyperspectral imaging. The nonnegative spectral transmittance can be applied to the optical imaging of programmable hyperspectral imaging system, and the collected images are supposed to be equivalent to the re-sults after PCA-LDA post-processing. By filtering out the negative value in the training process, the method can be directly applied to hyperspectral data, that is, hyperspectral data acquisition and PCA-LDA post-processing can be realized simultaneously through the physical process of optical imaging, which is helpful to realize more intelligent and convenient optical detection and sensing in sensing applications.
Click here to see the statistics on "Hans Journal of Biomedicine" .