利用积分球对近红外增强相机进行了可见、近红外波段的辐射定标, 并采用该相机采集了典型目标的可见、近红外图像.利用便携式地物光谱仪采集了相同目标可见、近红外波段的光谱反射率曲线, 并分析了目标在可见、近红外波段的光谱反射特性.通过寻找同一天气条件、同一时间段所拍摄的典型目标近红外图像灰度值与可见光图像灰度值及典型目标可见、近红外波段反射率之间的关系, 提出一种基于可见光图像及目标反射率反演近红外图像的算法.在近红外图像反演过程中考虑了大气透过率的影响, 反演结果表明, 本文算法可以较好地体现典型目标近红外图像的全局特征, 为后续不同天气条件下近红外场景仿真的实现提供参考.

An algorithm to inverse near infrared images based on visible light images and reflectance of typical target was proposed, according to the relationship between digital number of near infrared images and digital number of visible light images and reflectance of typical targets. In order to gain its relationship, two experiments were carried out. Firstly, radiometric calibration of camera in visible and near infrared waveband (0.38~1.1 μm) was conducted by using integrating sphere; secondly, visible light images and near infrared images of some typical targets were obtained by the camera; at the same time, the spectral reflectance which ranges from 0.38 μm to 1.1 μm for those typical targets was obtained by spectrometer under the same weather conditions. The atmospheric transmittance was also considered. The inversion result shows that the proposed algorithm can better reflect the characteristics of typical target in near infrared waveband, and this work lays a solid foundation to achieve the near infrared scene simulation.