Cloud detection method of optic remote sensing image

Abstract

The invention provides a cloud detection method of an optic remote sensing image, which comprises the following steps of: 1) diving an optic remote sensing image into basic processing units; 2) diving the basic processing units into two types which are possible to be cloud or not to be cloud according to a gray threshold; 3) diving the basic processing units which are possible to be the cloud according to fractal dimension; and 4) judging an undetermined unit in a fractal dimension dividing result to be a cloud layer or an earth object material according to angle second moment. The method combines with three judging characteristic parameters according to a reasonable order, and the different characteristic parameters have different functions when detecting, thereby not only being capable of obtaining higher resolution precision, but also being capable of obviously improving the running efficiency thereof in average meaning.

Description

technical field [0001] The invention relates to the technical field of optical remote sensing image processing, in particular to a method for detecting cloud slices and extracting cloud related parameters from optical images acquired by a sensor of a remote sensing platform. Background technique [0002] With the rapid development of science and technology, satellite remote sensing technology has entered a new stage that can quickly and timely obtain massive information on earth observation. These rich data information are widely used in various fields such as agriculture, meteorology, environmental protection, resource investigation, etc., and have important application value. In actual situations, due to the complex and changeable weather conditions, remote sensing satellites acquire satellite images, especially when acquiring large-scale remote sensing images, there will be clouds in most cases, and it is difficult to obtain real cloudless clear sky ground target images. ...

AI Technical Summary

Problems solved by technology

Although the texture analysis method can obtain higher discrimination accuracy than the threshold method, this method requires enough sample data for training to obtain the feature value intervals of different types of objects, and the quality and quantity of samples directly affect the accuracy of the training interval. Then affect the discriminant accuracy of the method

Moreover, the texture analysis method ignores the unique spectral characteristics of clouds

Therefore, there are still a certain number of misjudgments and omissions

Images