Twin network change detection model based on deep learning

Abstract

The invention provides a twin network change detection model based on deep learning, which comprises a double-branch calculation model used for acquiring a difference image, the double-branch calculation model comprises a twin network, a second branch convolution network and an up-sampling convolution network, the twin network is used for respectively extracting time phase feature maps of two time phases, and the up-sampling convolution network is used for carrying out up-sampling on the time phase feature maps of the two time phases; the second branch convolutional network is used for calculating a difference feature map according to the two time-phase feature maps and a difference feature map of the two time-phase feature maps, and the up-sampling convolutional network is used for carrying out up-sampling and / or deconvolution operation on the difference feature map to obtain a difference image. According to the method, the ResNet18 model is transformed to establish a twin network ResAtNet for a change detection scene, the difference feature extraction capability is improved through a double-branch difference feature graph generation method, the model can be suitable for target learning high-dimensional change features, suitable feature expression does not need to be selected by expert knowledge, the method is adaptive to various change scenes, and compared with other existing models, the method has the advantages that the method is simple and convenient to implement, and the method is suitable for large-scale popularization and application. The method has an obvious precision advantage.

Description

technical field [0001] The invention belongs to the technical field of change detection, and in particular relates to a twin network change detection model based on deep learning. Background technique [0002] With the development of aerospace and UAV technology, remote sensing is also being used more and more. For remote sensing data, the label changes of multi-temporal image scenes covering the same area and collected at different times can reflect the change of land use from the semantic level. Therefore, remote sensing change detection is widely used in many fields, such as land use, urban expansion, farmland change, geological disaster monitoring, ecological environment protection, wetland monitoring, forest protection, etc. [0003] In recent years, artificial intelligence, especially deep learning, has gradually emerged. As a new type of machine learning model, its concept originated from the research of artificial neural networks. Among them, "depth" is relative to...

AI Technical Summary

Problems solved by technology

[0005] (1) Traditional high-resolution remote sensing change detection methods mainly include pixel-based methods and object-based methods, but pixel-based remote sensing change detection methods tend to generate more false and noise points in small areas, while object-oriented remote sensing detection methods The method is easily affected by the image segmentation algorithm, and requires a lot of expert knowledge to design and change the feature library;

[0006] (2) The complexity of remote sensing images greatly increases the difficulty for deep learning models to learn robust and discriminative representations from scenes and objects. In this case, existing change detection models have improved learning capabilities for a small amount of training data. space;

[0007] (3) Deep learning needs to learn object features from a large-scale sample library, and the manual labeling of samples is usually an expensive and time-consuming process. Large-scale images usually only have a small part of the changed area, the amount of data is small, and there is generally a lack of high-quality images. quality training images

Images