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

Remote sensing image building automatic extraction method and system based on residual network

A remote sensing image and automatic extraction technology, applied in the field of multi-spectral remote sensing images, can solve the problems of low degree of automation, achieve the effect of improving accuracy, reducing interpretation cost, and improving interpretation efficiency

Active Publication Date: 2020-05-05
CHINA SURVEY SURVEYING & MAPPING TECH
View PDF7 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, at present, many remote sensing images still require manual interpretation. Even if there are some interpretation algorithms, most of them are not highly automated and require manual intervention.

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
  • Remote sensing image building automatic extraction method and system based on residual network
  • Remote sensing image building automatic extraction method and system based on residual network
  • Remote sensing image building automatic extraction method and system based on residual network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0030] like figure 1 , in this embodiment, the method for automatically extracting buildings from remote sensing images based on the residual network includes:

[0031] Step 101, constructing a training sample data set.

[0032] In this embodiment, the training sample data set can be constructed in the following manner: obtain multispectral (including four bands of red, green, blue, and infrared) and panchromatic images collected by remote sensing satellite cameras; combine multispectral and panchromatic The images are fused to obtain a fused image; the fused image is segmented to obtain a segmented image; the label corresponding to the segmented image is determined; and the training sample data set is obtained according to the obtained segmented images with labels. Among them, the label is used to indicate the location of the building.

[0033] Preferably, when performing segmentation processing on the fused image, the segmentation size includes but not limited to: 500x500 ...

Embodiment 2

[0051] On the basis of the above embodiments, the following will be described in conjunction with a specific example.

[0052]The embodiment of the invention discloses a method for automatically extracting buildings from remote sensing images based on a residual network, which is used for extracting buildings in a wide range. Under the Caffe framework, first complete the construction of the network, migrate the model trained on ImageNet, train the expanded training samples, and then use the tuned model parameters to extract buildings from the test remote sensing images.

[0053] Step 1: Prepare the training sample data set.

[0054] In this embodiment, 10 scenes of Gaofen No. 1 images and 5 scenes of ZY302 images in Beijing, Tianjin, Zhejiang and other regions were used to make training samples. The training sample in this experiment is a panchromatic and multi-spectral fusion image, with a resolution of 2 meters, including four channels, namely blue, green, red and infrared ...

Embodiment 3

[0072] On the basis of the above embodiments, the present invention also discloses a system for automatically extracting buildings from remote sensing images based on a residual network, including: a training set construction module, used to construct a training sample data set; a training set processing module, used to The samples in the training sample data set are expanded to obtain the expanded training sample data set; the model building module is used to construct a residual network model to be trained for automatic extraction of remote sensing image buildings; the parameter training module is used to The training samples in the final training sample data set are trained to train the residual network model to obtain network parameters; the model update module is used to update the parameters of the residual network model to be trained according to the obtained network parameters to obtain the training residual network The model; the detection module is used to perform bui...

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 remote sensing image building automatic extraction method and a system based on a residual network. The method comprises the following steps: constructing a training sample data set; expanding samples in the training sample data set to obtain an expanded training sample data set; constructing a to-be-trained residual network model for automatic extraction of the remote sensing image building; training a to-be-trained residual error network model according to the training samples in the expanded training sample data set to obtain network parameters; updating parametersof the residual network model to be trained according to the obtained network parameters to obtain a training residual network model; and performing building detection on the test data according to the training residual network model. According to the method, automatic extraction of buildings is realized, the interpretation cost is reduced, and the interpretation efficiency is improved.

Description

technical field [0001] The invention belongs to the technical field of multispectral remote sensing images, in particular to a method and system for automatically extracting buildings from remote sensing images based on a residual network. Background technique [0002] Remote sensing is the technique of detecting and identifying research objects at long distances using modern vehicles and sensors. Based on telemetry instruments on artificial earth satellites, aviation and other platforms, the electromagnetic wave characteristics of target objects are obtained from a long distance, and through the transmission, storage, satellite, correction, and identification of target objects of this information, sensory telemetry and resource management are implemented on the earth's surface. With the development of economy and science and technology, remote sensing earth observation technology has entered a period of rapid development. The launch of various sensors has enriched satellit...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V20/176G06N3/045G06F18/214
Inventor 王海波公雪霜王冰冰齐建超曹琼邵俊
Owner CHINA SURVEY SURVEYING & MAPPING 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