River and lake remote sensing image segmentation method and system based on convolutional neural network and Transform

A convolutional neural network and remote sensing image technology, applied in the field of river and lake remote sensing image segmentation methods and systems, can solve the problems of affecting the accuracy of segmentation, target loss, and small coverage area, and achieve improved speed and accuracy, good effects, Robust effects to environmental changes

Active Publication Date: 2021-07-30
SHANDONG UNIV +1
View PDF9 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since the remote sensing data is generally collected by satellites or drones, and the "four chaos" phenomenon around rivers and lakes generally covers a small area, the target may be lost after multiple downsampling, thus affecting the accuracy of segmentation. Therefore, under the premise of not affecting the segmentation accuracy and avoiding the loss of targets, it is a very challenging problem to identify small areas of "chaos" in rivers and lakes from remote sensing images.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • River and lake remote sensing image segmentation method and system based on convolutional neural network and Transform
  • River and lake remote sensing image segmentation method and system based on convolutional neural network and Transform
  • River and lake remote sensing image segmentation method and system based on convolutional neural network and Transform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0034] Such as Figure 1-2 As shown, the present embodiment provides a method for segmenting river and lake remote sensing images based on convolutional neural network and Transformer, including:

[0035] S1: Obtain remote sensing images of rivers and lakes containing category labels, and construct a training set;

[0036] S2: preprocessing the training set;

[0037] S3: Use convolutional neural network to extract multi-layer feature maps on the preprocessed training set;

[0038] S4: Use the Transformer encoder to extract the attention feature for the extracted feature map of the last layer, and use the Transformer decoder to obtain the self-attention feature map for the attention feature;

[0039] S5: After splicing the self-attention feature map and the first layer feature map, train the image segmentation model;

[0040] S6: Based on the trained image segmentation model, the category segmentation result of the target in the remote sensing image of the river and lake to ...

Embodiment 2

[0059] This embodiment provides a river and lake remote sensing image segmentation system based on convolutional neural network and Transformer, including:

[0060] The data acquisition module is configured to acquire remote sensing images of rivers and lakes containing category labels, and construct a training set;

[0061] The feature map extraction module is configured to extract multi-layer feature maps using a convolutional neural network for the training set;

[0062] The Transformer module is configured to use a Transformer encoder to extract attention features for the extracted last layer feature map, and use a Transformer decoder to obtain a self-attention feature map for the attention features;

[0063] The model training module is configured to train the image segmentation model after splicing the self-attention feature map and the first layer feature map;

[0064] The image segmentation module is configured to obtain the category segmentation result of the target ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a river and lake remote sensing image segmentation method and system based on a convolutional neural network and Transform, and the method comprises the steps: obtaining a river and lake remote sensing image containing a class label, and constructing a training set; extracting a multi-layer feature map from the training set by using a convolutional neural network; extracting an attention feature from the last layer of extracted feature map by adopting a Transform encoder, and obtaining a self-attention feature map from the attention feature by adopting a Transform decoder; after the self-attention feature map and the first-layer feature map are spliced, training an image segmentation model; and obtaining a category segmentation result of the target in the remote sensing image of the river and lake to be detected based on the trained image segmentation model. Transform is introduced into the field of remote sensing image segmentation, a self-attention mechanism is used for replacing convolution operation, the receptive field area during operation is enlarged, due to the fact that down-sampling and up-sampling operation does not exist, image scale change cannot be caused, the problem of target loss is solved, and the defects of an existing deep learning segmentation method in the field of remote sensing image segmentation are overcome.

Description

technical field [0001] The invention relates to the technical field of convolutional neural networks and remote sensing images, in particular to a method and system for segmenting river and lake remote sensing images based on convolutional neural networks and Transformers. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] The "four chaos" of rivers and lakes refers to the phenomenon of "random construction, random occupation, random mining, and random piles" around rivers and lakes. The areas around rivers and lakes have been banned repeatedly, seriously affecting the ecological environment of urban rivers and lakes and the quality of life of nearby residents. The rectification of the "four chaos" in rivers and lakes has become the main content of the supervision of urban rivers and lakes. [0004] At present, the illegal extraction method...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/38G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V20/13G06V10/28G06V10/449G06N3/047G06N3/045G06F18/2415G06F18/241
Inventor 杨公平刘一锟孙启玉邓青李红超郭伟
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products