基于宽光谱光学遥感图像的细分光谱光学遥感图像的模拟

Simulation of an image with a subsection of spectral band using an image with a wider spectral band

本文分析了现有遥感图像模拟方法的特点,介绍了地物波谱知识库的建立情况与应用意义,论述了光学传感器的成像特点及地面实测数据与遥感反演数据之间的转换关系。本文在对TM图像的光谱波段进行细划分的基础上,提出利用波谱知识库的数据支持来模拟获取具有更高光谱分辨率的细分光谱光学遥感图像的方法。本方法依据光谱数据库等先验知识的支持,通过遥感成像时辐射传输的模拟运算,在宽光谱光学遥感数据提供的图像纹理信息和地类信息的基础上,获取细分光谱波段的模拟光学遥感图像,同时避免了难以接受的计算量。细分光谱模拟图像的实验结果表明,本方法能够较为可靠的模拟出真实高光谱分辨率图像的信息,模拟方法可信。

The characteristics of current Remote Sensing （RS） image simulation methods were analyzed; the application significance and the existing status of spectrum knowledge base of objects were described; the imaging principles of the optics image were discussed; the translation relationship between the data computed from RS images and the data measured directly from the earth surface were analyzed. A method was proposed by using images with lower spectrum resolution to simulate images with higher spectrum resolutions. Based on previously obtained data （i.e. spectrum library） and simulations of the radiation transmission process, our method uses the texture and sorting information of objects which is derived from a RS image with a wider spectral band, and to finally simulate a RS image with a subsection of spectral band. During the simulation process, unacceptable calculations were avoided. Our results show that our technique is effective.