Water body contour automatic extraction method and system based on full convolutional neural network

Abstract

The invention relates to a water body contour automatic extraction method and system based on a full convolutional neural network. The water body contour automatic extraction method comprises the following steps: respectively constructing a training sample library and an image prediction library; iteratively training the training sample library through a multi-receptive field feature combined full convolutional network to obtain a network model; and performing water body contour extraction on an image prediction library by using the network model to obtain a surface water body contour extraction result, wherein the training sample library is constructed by surface images and water body annotation data, and the image prediction library is constructed by surface images. The multi-receptive-field feature combined full convolutional neural network is high in scale robustness, can adapt to water extraction of high-resolution remote sensing images of different scales under different complex conditions, and can realize continuous iterative optimization.

Description

technical field [0001] The present invention relates to a coordinate fitting method, and more specifically refers to a method and system for automatically extracting water contours based on a fully convolutional neural network. Background technique [0002] Water body extraction is of great significance for applications such as water resource monitoring, natural disaster assessment, and environmental protection, and the use of remote sensing technology to obtain surface information is the most commonly used technical means. Traditional methods mainly include band threshold method, supervised classification method, water body index method, and spectral relationship method. For water body recognition, little attention is paid to the shape, size, texture, edge, semantics, shadow and other characteristics of the water body, which seriously affects the accuracy of water body extraction. In addition, for massive remote sensing data processing, traditional methods generally have l...

AI Technical Summary

Problems solved by technology

In the production of various topographic maps and thematic maps, cartographers need to manually outline various contours of water bodies on satellite or aerial images, which is obviously a relatively heavy and inefficient work

Images