Remote sensing image cloud cover calculation method based on composite neural network

Abstract

The invention relates to the technical field of remote sensing satellite image processing and transmission, and provides a remote sensing image cloud cover calculation method based on a composite neural network. The method has the characteristics of high efficiency, good optimization performance, high accuracy and the like, and is suitable for the fields of remote sensing cloud cover calculation and detection and the like. The remote sensing image cloud cover calculation method based on a composite neural network comprises the following steps: (1) establishing a cloud cover calculation sample library in a cloud cover grading manner; (2) constructing a composite neural network; (3) carrying out training and learning of the composite neural network, adjusting weights of the miniature neural network and the small neural network respectively, and obtaining a network model capable of being used for cloud cover calculation; (4) calculating the cloud amount of a remote sensing image by adopting a thumb graph priority strategy to obtain a cloud amount value. The method is mainly applied to remote sensing satellite image processing and transmission occasions.

Description

technical field [0001] The invention relates to the technical field of remote sensing satellite image processing and transmission, in particular to a remote sensing image cloud amount calculation method based on a composite neural network, which can be used in application scenarios such as remote sensing image cloud amount calculation, remote sensing satellite imaging quality evaluation, and remote sensing satellite data transmission. . Background technique [0002] In the field of remote sensing image cloud amount calculation, there are mainly the following methods for cloud amount evaluation, but they all have some defects in accuracy and calculation efficiency: [0003] (1) Brightness threshold method is the oldest and simple cloud area extraction and cloud amount calculation algorithm. It is based on the difference in the brightness value of clouds and ground objects. For images of a specific type of sensor, the cloud area is extracted by thresholding. For different ima...

AI Technical Summary

Problems solved by technology

In addition, the Japanese scholar Otsu proposed an automatic threshold based on the maximum inter-class variance classification criterion (hereinafter referred to as the Otsu threshold), which is also commonly used to extract cloud areas, but these methods are only effective for scenes with high cloud content, and cannot automatically identify cloudless Scenes

[0004] (2) Using brightness for analysis will inevitably misjudgment bright features such as snow, sand, and rocks

These methods improve the detection accuracy to a certain extent, but generally need to use a large number of artificially interpreted and different types of cloudy images as samples to train the classifier, and the manual interpretation of cloudy images is an extremely time-consuming, labor-intensive task. strenuous work

Moreover, most methods have low efficiency and high requirements for computer memory, and often do not have the ability to process massive remote sensing images.

Images