Optical Remote Sensing Image Cloud Detection Method

Abstract

The invention discloses a cloud detection method for an optical remote sensing image, comprising the steps of: obtaining a brightness map of the image to be detected; converting a multispectral image in the image to be detected into a single-band brightness image; Cloud sample images, calculate the corresponding highest and lowest brightness thresholds; calculate accurate brightness thresholds: analyze the histogram of the image to be detected, qualitatively filter the cloud-free images, and for the remaining cloud-containing images, use the roughly estimated brightness double threshold as the limiting condition to execute based on Calculation of the maximum inter-class variance to obtain an accurate brightness threshold; cloud area morphology synthesis: perform morphological operations on the cloud area after threshold segmentation, eliminate noise points caused by cloud-like objects, fill cloud gaps, optimize cloud area contours, and output the final Cloud mask and cloud content. The present invention quickly acquires image cloud mask and cloud content with high accuracy and qualitatively identifies cloudless images on the premise of less manual participation, and is applicable to both panchromatic and multi-spectral images.

Description

technical field [0001] The invention relates to the technical field of optical remote sensing image processing, and more specifically, the invention relates to a cloud detection method for optical remote sensing images. Background technique [0002] Represented by Zi Zi No. 3, Gao Fen No. 1, and Gao Fen No. 2, the design indicators of domestic optical remote sensing satellites have gradually reached the international advanced level. increase year by year. However, not all remote sensing images can meet the requirements of intelligent processing of image information, and one of the most important factors is the cloud cover on the image. [0003] In the current field of cloud detection, the spectral threshold method is the simplest and most effective algorithm. It is based on the difference in the spectral characteristics of clouds and ground objects in the visible light band, and realizes the distinction between cloud and non-cloud targets through brightness thresholds. Com...

AI Technical Summary

Problems solved by technology

This type of method regards the cloud as a changing target in the image, and uses the idea of ​​change detection to detect the cloud. It is often used in conjunction with the cloud detection algorithm based on a single image, which can effectively improve the detection accuracy, but the disadvantage is that it requires high data. The image itself needs to have relatively accurate geographic information

Images