Target detection of hyperspectral remote sensing image is one of the important research contents of hyperspectral remote sensing application, and its application fields include environmental monitoring, urban investigation, mineral mapping and military reconnaissance. The great connotation and uniqueness of hyperspectral remote sensing lies in that it can start with the high-dimensional spatial characteristics of data and carry out more effective target detection and classification processing based on the physical properties of ground objects themselves. Based on this, many target detection algorithms have been developed in recent years. However, at present, there is little research on hyperspectral image target detection technology based on harmonic analysis. This chapter will explore and experiment the small target detection and recognition technology of hyperspectral images from the perspective of hyperspectral dimension harmonic analysis.

Harmonic analysis technology is widely used in power pollution detection and treatment. For example, Tang Sheng Qing et al. (2006) detected power system harmonics based on BP(Back Propagation) neural network, and gave the training and simulation of the third harmonic, which can be used for real-time detection and harmonic dynamic compensation of power system. In the field of remote sensing, harmonic analysis is widely used to analyze time series data at present, which was first put forward by Jakubauskas and others, 200 1, 2002. In order to describe the seasonal changes of natural and agricultural land use, they also studied the application of harmonic analysis in crop species identification, and conducted experiments in southern Kansas. Toshi-Hiro et al. (2005) used harmonic analysis to detect crop phenology of MODIS time series data; Bethany et al.(2007) explored a curve fitting technique based on harmonic analysis, and extracted the annual phenological law from noisy NDVI time series data. Aiming at the problem of hyperspectral image target detection, there is no literature on harmonic analysis at present.