Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A new building recognition method in remote sensing images based on deep learning

A remote sensing image and deep learning technology, applied in the field of image processing, can solve the problems of low clarity and integrity of remote sensing images

Active Publication Date: 2022-03-25
SOUTH CHINA UNIV OF TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Due to external interference such as shooting weather, altitude, and cloud cover, the clarity and integrity of the available remote sensing images are lower than those of other types of 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
  • A new building recognition method in remote sensing images based on deep learning
  • A new building recognition method in remote sensing images based on deep learning
  • A new building recognition method in remote sensing images based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0051] like figure 1 As shown, a method for identifying new buildings in remote sensing images based on deep learning includes the following steps:

[0052] S1. First, manually collect two remote sensing images of a specific area in two different time periods, and mark some of the newly added buildings to obtain the corresponding new building background (label) images;

[0053] For the two remote sensing images of two time periods, according to the prior knowledge, some new buildings are manually marked, and the pixel value of the new building is set to 1, and the rest is set to 0.

[0054] S2, such as figure 2 (a) and figure 2 (b) shows the original remote sensing images of two time periods. Using the designed image sliding cutting algorithm, the two remote sensing images with the original size and the corresponding new building background images are cut and cropped to obtain several small-size images; The size of the small-size image is 256×256; the idea of ​​the image ...

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 method for recognizing newly added buildings in remote sensing images based on deep learning, comprising the following steps: obtaining sample images: remote sensing images in two time periods and a background image of newly added buildings; and cutting out the sample images with original sizes respectively. Cut to obtain small-size images; perform data enhancement processing on all small-size images; then perform centralization and global comparison and normalization of the enhanced small-size remote sensing images of two time periods, and then subtract one by one after completion to obtain remote sensing images Difference map; input the remote sensing difference map and the small-sized new building background image into the modified two deep neural networks for network parameter training; then input the remote sensing image to be tested into the trained two deep neural networks, and output them in the softmax of the network Model fusion is performed at the layer, and then the initial output results are modified to obtain the final new building recognition image. The method of the invention has the advantages of high recognition accuracy of newly added buildings in the remote sensing image and wide applicability.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method for recognizing newly added buildings in remote sensing images based on deep learning. Background technique [0002] The recognition of new buildings in remote sensing images based on deep learning is completed by image segmentation technology. [0003] Existing traditional image segmentation algorithms include: segmentation methods based on color and brightness, segmentation methods based on regions, segmentation methods based on graph theory, and segmentation methods based on energy functionals. The segmentation method based on color and brightness is to divide each pixel by the color or brightness of the image. For example, the K-Means algorithm regards the image as a point set composed of RGB three-dimensional features, and performs all pixel points in the image. Clustering achieves the purpose of segmentation; region-based segmentation methods, including r...

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 Patents(China)
IPC IPC(8): G06V20/10G06V10/34G06V10/82G06V10/774G06V10/764G06V10/26G06N3/04
CPCG06V20/176G06V10/30G06V10/44G06N3/045G06F18/214
Inventor 陈佳余卫宇王珂尧
Owner SOUTH CHINA UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com