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

High-resolution image weak supervision building extraction method combining pixel semantic association and boundary attention

A semantic association and building technology, applied in the field of remote sensing images, can solve problems such as ignoring boundary information

Pending Publication Date: 2020-12-08
CENT SOUTH UNIV
View PDF0 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Such methods rely too much on the information in the superpixel prior, while ignoring the boundary information from the image

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
  • High-resolution image weak supervision building extraction method combining pixel semantic association and boundary attention
  • High-resolution image weak supervision building extraction method combining pixel semantic association and boundary attention
  • High-resolution image weak supervision building extraction method combining pixel semantic association and boundary attention

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0078] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention can be implemented in various ways defined and covered by the claims.

[0079] see Figure 1 to Figure 4 , a weakly supervised extraction method for high-resolution images that combines pixel semantic association and boundary attention, including the following steps:

[0080] Step A, training data preparation, including generating superpixel maps of all training high-score remote sensing images, generating building category heat maps, building background heat maps, and initial pixel semantic correlation labels;

[0081] a1. Generation of superpixel maps. The obtained high-resolution remote sensing images are used as training input, input into the superpixel segmentation model, and the number of superpixels in each superpixel map is set to generate superpixels corresponding to all high-resolution remote sensing images. Figure ...

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 provides a high-resolution image weak supervision building extraction method combining pixel semantic association and boundary attention. The method comprises the steps of training datapreparation, deep feature extraction, boundary feature fusion, pixel semantic association degree learning, loss function calculation and building pseudo annotation generation. By designing a boundaryattention module, superpixel prior information and boundary information extracted by a network are combined, building boundary features are enhanced; semantic information between pixels is effectivelyspread in an image by learning semantic relevance between the pixels, and a pseudo label which is more complete, dense and clearer in boundary is generated. Meanwhile, a full convolutional network model is adopted for training in cooperation with a high-resolution remote sensing image, and building feature automatic extraction is achieved.

Description

technical field [0001] The invention relates to the field of remote sensing images, and more specifically, relates to a method for extracting high-resolution images and weakly supervised buildings by combining pixel semantic association and boundary attention. Background technique [0002] As one of the applications of semantic segmentation of remote sensing images, building extraction has important practical value in many fields such as the establishment and update of urban geographic databases, urban population estimation, and land cover change. In recent years, with the rapid development of remote sensing imaging technology, the spatial resolution and spectral resolution of satellite images have been greatly improved, making it possible to accurately identify and locate buildings. There are two traditional methods for extracting buildings from remote sensing images: pixel-based and object-oriented. The pixel-based method takes a single pixel as the basic unit, and mainly...

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
IPC IPC(8): G06K9/00G06K9/34G06K9/62G06N3/04G06T3/40G06T7/194
CPCG06T7/194G06T3/4007G06T2207/10032G06V20/176G06V10/267G06N3/045G06F18/214
Inventor 陈杰何玢李建辉郭亚孙庚邓敏
Owner CENT SOUTH UNIV
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