高光谱数据融合技术在物证检验中的应用

Application of Hyperspectral Data Fusion Technology in Evidence Examination

摘要: 高光谱成像技术不仅可以记录物证的空间信息，还可以记录物证的光谱信息，图谱合一的特点在物证检验中发挥着重要的作用。由于高光谱成像仪的光学结构特点，导致高光谱成像仪获得物证的空间分辨率较低。本文提出高空间分辨率图像与高光谱数据融合的方法，通过插值算法使高空间分辨率图像的每个空间点对应一条光谱曲线。实验采用相机拍摄的2020万像素的高清图片与400~1700 nm高光谱成像仪采集的高光谱数据，通过插值融合算法获得空间分辨率2020万像素、光谱范围400~1700 nm的高光谱数据立方。结果表明，高光谱数据融合技术不仅能够获取高空间分辨率的空间图像而且能够获取每个空间点的光谱曲线，解决了高光谱成像仪在物证检验过程中空间分辨率低的问题。

Abstract: Hyperspectral imaging technology can record not only the spatial information but also the spec-tral information of the evidence. The combination of atlas plays an important role in evidence examination. Due to the optical structure of the hyperspectral imager, the spatial resolution of the evidence is low. In this paper, a fusion method of high spatial resolution image and hyperspectral data is proposed. An interpolation algorithm is used to make each spatial point of high spatial resolution image correspond to a spectral curve. In the experiment, the high-definition images of 20.2 million pixels taken by the camera and the hyperspectral data of 400 - 1700 nm hyperspectral imager were collected. The hyperspectral data cube with spatial resolution of 20.2 million pixels and spectral range of 400 - 1700 nm is obtained by the interpolation fusion algorithm. The results show that hyperspectral data fusion technology can acquire not only the high spatial resolution images, but also the spectral curves of each spatial point, which solves the problem of low spatial resolution of hyperspectral imager in the process of the evidence examination.