High-spectrum automatic cloud detection method based on space and spectral information

Abstract

The invention discloses a high-spectrum automatic cloud detection method based on space and spectral information, which can be applied to automatic cloud judgment of aerospace and aviation hyperspectral remote sensing images, is capable of reducing storage cost of hyperspectral remote sensing data and saving transmission bandwidth. The high-spectrum automatic cloud detection method comprises the following steps of: carrying out preprocessing and waveband selection according to spectral samples of cloud, snow, water and the like and then carrying out characteristic extraction; carrying out sample training by adopting a classifier to obtain a cloud spectral classifying model; in a hyperspectral image cloud detection stage, classifying each pixel of hyper-spectrums through preprocessing, waveband selection and characteristic extraction which are same in a training stage; and checking consistency of pixels and neighborhood pixels, finally determining whether the pixels are cloud pixels, and working out the proportion of the cloud pixels and giving a cloud judgment result. By the high-spectrum automatic cloud detection method, in conjunction with image segmentation, target classifying identification and machine learning technologies, the problem of hyperspectral cloud judgment is solved and the defect of low detection rate caused by only utilizing textural information or spectral information can be avoided.

Description

technical field [0001] The invention relates to the technical field of remote sensing, in particular to a hyperspectral automatic cloud detection method based on spatial and spectral information, which can be applied to automatic cloud detection of hyperspectral remote sensing images acquired by aerospace and aviation sensor platforms. Background technique [0002] With the rapid development of space technology and sensor technology, remote sensing image data has become an important means for people to obtain information, playing an increasingly important role in military reconnaissance, environmental monitoring, resource investigation, land use, and urban planning. However, clouds, as an obstacle to the propagation of optical signals, greatly affect the quality of remote sensing information acquisition, thereby reducing the utilization of data. At present, except for the synthetic aperture radar sensor that can directly image through clouds, other sensors have not completel...

AI Technical Summary

Problems solved by technology

These methods have the disadvantages of large amount of calculation, the natural texture is difficult to express with a single model, the generalization performance of the algorithm is not good, and most methods are only suitable for some specific images

It is difficult to meet the needs of modern remote sensing mass data processing and real-time rapid interpretation of image intelligence

Images