Road extraction method based on fully convolutional neural network ensemble learning

Abstract

The present invention proposes a road extraction method based on fully convolutional network ensemble learning, aiming to obtain road extraction results with coherent roads, complete road network structure and high recall rate. The implementation steps are as follows: first select the sample set in the existing remote sensing database and divide the test samples and training samples, secondly train a fully convolutional neural network with cross entropy as the loss function, and then change the positive sample penalty weight of the loss function. Based on the optimization of network model parameters of a network, a fully convolutional neural network with different penalty weight loss functions is obtained, and then the fully convolutional neural network obtained by training is used for road extraction, and the road extraction results of different fully convolutional networks are obtained. Finally, according to the principle of spatial consistency, the different extraction results are integrated according to the integration strategy based on spatial consistency, and the final result map is output. The invention can improve the recall rate of the road extraction result and has strong robustness.

Description

technical field [0001] The invention belongs to the technical field of image processing, relates to a road extraction method, and further relates to a road extraction method based on fully convolutional neural network integrated learning, which can be used to obtain the road network structure of an optical remote sensing image. Background technique [0002] Road extraction from remote sensing images aims to replace the tedious manual work, using the road extraction method to obtain the semantic segmentation result map of the road and background of the input remote sensing image. The current road extraction methods can be roughly divided into three categories. The first type is knowledge-based road extraction methods, such as: threshold method, template matching method; the second type is based on morphology methods, such as: edge detection method, watershed algorithm, split growth method; these two types of road extraction methods The effect is not ideal, and the overall ac...

AI Technical Summary

Problems solved by technology

Although this method makes full use of the excellent characteristics of the fully convolutional neural network, it can extract discriminative features and extract the road results of these features, but when using this model for road extraction, there are certain deficiencies: first, the When the method is used for road extraction, it does not take into account the unbalanced distribution of roads and backgrounds in remote sensing images. The fully convolutional neural network uses the cross-entropy loss function with the same penalty weight for positive and negative samples; secondly, using the fully convolutional neural network When the network extracts roads, it does not take into account the spatial consistency of remote sensing images

However, since the penalty weight of the loss function largely depends on the road characteristics of the training samples, and the spatial consistency of remote sensing images is not considered when extracting roads, the test samples that are quite different from the training samples are The improvement of road extraction recall rate is limited, and the robustness is poor

Images