Remote sensing image change detection method based on deep learning

Abstract

The invention discloses a remote sensing image change detection method based on deep learning. The method comprises the steps: converting a remote sensing image change detection problem into an imagesemantic segmentation problem for processing, and constructing a change detection model based on U-net; introducing a ConvLSTM layer to optimize a U-net structure to obtain a change detection model based on L-Unet, introducing hole convolution, and referring to a hole space pyramid pooling structure to obtain a change detection model based on A-Lunet; based on the public remote sensing data set, training and testing the constructed change detection model to obtain a trained change detection model; and inputting the remote sensing images of different time phases into the trained change detection model to obtain a final detection result. According to the change detection model based on L-Unet and the change detection model based on A-Lunet provided by the invention, relatively high accuracycan be achieved, and relatively high adaptability to an image offset problem is achieved.

Description

technical field [0001] The invention relates to the technical field of remote sensing image change detection, in particular to a remote sensing image change detection method based on deep learning. Background technique [0002] Remote sensing image change detection uses correlation algorithms to analyze two or more remote sensing images in different phases of the same area, and find out the changed area, which can provide people with large-scale change information on the earth's surface, and is one of the important research directions of remote sensing technology. 1. Application in many fields such as landform feature monitoring, natural disaster monitoring, environmental monitoring (early fire detection based on thermal imaging nonlinear multi-period forecast, environmental remote sensing, 2007, 110(1): 18-28), forest resource monitoring, etc. wide and play a vital role. [0003] However, in fact, in most change detection applications, the common methods are still visual i...

AI Technical Summary

Problems solved by technology

At present, the existing methods basically realize the change detection of remote sensing images and improve the performance of change detection. However, with the continuous development of remote sensing technology, change detection of remote sensing images still faces many problems that need further research.

Images