高光谱目标分类技术发展与展望

Development and Prospect of Hyperspectral Target Classification Technology

高光谱成像技术源于遥感探测,具有谱图合一的独特优势,在农林、地矿、防伪和环境保护等领域的应用日益广泛。作为高光谱技术最常见和最基本的功能,目标分类在各个高光谱领域的应用均具有至关重要的作用。综述了高光谱目标分类技术的发展现状,分析了高光谱数据的结构特点,归纳了高光谱目标分类的一般流程,并详细阐述了数据读取、图像预处理和目标分类等高光谱目标分类流程的主流方法及其基本原理。结合近年来的典型案例,分析了高光谱目标分类的过程,并对其发展趋势进行了分析和展望。

Hyperspectral imaging technology originated from remote sensing,has the unique advantage of spectral integration.It is increasingly widely used in agriculture and forestry,geology and mining,anti-counterfeiting and environmental protection.As the most common and basic function of hyperspectral technology,target classification plays a vital role in the application of hyperspectral fields.In this paper,the development status of hyperspectral target classification technology is reviewed,the structural characteristics of hyperspectral data are analyzed,the general process of hyperspectral target classification is summarized,and the mainstream methods and basic principles of data reading,image preprocessing and target classification are described in detail.Combined with the typical cases in recent years,the process of hyperspectral target classification is analyzed,and the development trend is analyzed and prospected.