Day and night cloud detection method and device based on static satellite multichannel data

Abstract

The invention relates to a day and night cloud detection method based on static satellite multichannel data. The day and night cloud detection method comprises the steps: S1, preprocessing data, and generating throughout-the-year month-by-month clear sky background traced drawing data; S2, performing day and night judgment on a current image one-by-one pixel, and carrying out detection by a corresponding detection algorithm. By a multi-step cloud detection algorithm, visible light and thermal infrared band information are comprehensively used, so that the probability of mis-judgment is reduced; cloud detection can be quickly, accurately and stably realized on the earth surface of which the underlying surface is covered by ice snow, desert and vegetations, and high adaptability is achieved. The invention also provides a day and night cloud detection device based on static satellite multichannel data.

Description

technical field [0001] The invention relates to the technical fields of remote sensing image processing and atmospheric science, in particular to a day and night cloud detection method and device based on multi-channel data of stationary satellites. Background technique [0002] At present, cloud detection is an important part of the World Climate Research Program. The existing satellite image cloud detection is generally divided into two categories: statistical method and threshold method, and combines cluster analysis, pattern recognition, wavelet analysis and neural network methods. ISCCP uses visible light and infrared channel data, and uses the difference threshold between clear sky radiation and observed radiation to realize cloud detection; MODIS and NOAA use CO 2 Thin slice method realizes cloud detection. Practice shows that a single algorithm always has limitations, such as the low accuracy of cloud detection at high latitudes and seasonal cloud changes, and the i...

AI Technical Summary

Problems solved by technology

[0006] In summary, various cloud detection methods have their own shortcomings, such as: the threshold method and the statistical equation method are affected by the limitations of time and space; the radiative transfer algorithm is limited by the algorithm model, and often becomes an auxiliary method for other methods; Although the cloud detection method has certain advantages over the basic methods, it is still difficult to detect them due to the special radiation characteristics of thin clouds, low clouds and fog

However, the existing patent documents focus more on cloud detection methods for polar-orbiting satellites and aerial data, and focus on cloud detection when visible light data is available during the day, while there are fewer patent documents on cloud detection for geostationary satellites.

Images