Building classification method based on satellite remote sensing images

Abstract

The invention belongs to the technical field of satellite remote sensing image processing, and particularly relates to a building classification method based on the satellite remote sensing images. The method comprises the following steps of carrying out multi-azimuth accurate data processing on the satellite remote sensing image data; classifying according to the feature extraction degrees of thebuildings; establishing a satellite remote sensing image feature judgment layer; building a classification model for the buildings in the satellite remote sensing images according to a convolutionalneural network; and carrying out weighted production on the classification data with a larger training error function value obtained after testing to obtain a new training data set, carrying out iterative training on the classification model again, storing an optimal model, and packaging the optimal model into a classifier; and finally, displaying different types of detection results in a classified manner according to the actual distribution condition of the satellite remote sensing images and the actual application field. The method solves the problems of low classification speed, low accuracy and the like of the satellite remote sensing images by adopting a traditional method, is high in landing applicability and wider in application field, can be used for the urban design planning andlayout, enables the urban building distribution to be more reasonable, and can be used for the urban public safety, emergency shelter detection and the like.

Description

technical field [0001] The invention belongs to the technical field of satellite remote sensing image processing, in particular to a building classification method based on satellite remote sensing images. Background technique [0002] Satellite remote sensing image data is the reflection of electromagnetic waves and the electromagnetic waves emitted by remote sensing satellites when they detect objects on the earth's surface in space, so as to extract the information of the objects, convert these electromagnetic waves, and identify visualized images. The difference between satellite remote sensing image and ordinary optical image is that its resolution is low, the target object in the image is deformed greatly, and there are many wave bands. Classification methods are difficult to implement a satellite remote sensing image classification model with high accuracy. [0003] With the development of artificial intelligence, machine learning and deep learning have become a majo...

AI Technical Summary

Problems solved by technology

However, most of the images are ordinary optical images, and there are fewer processing methods that combine satellite remote sensing images with deep learning techniques

Images