DSM local missing repairing method based on deep learning

A repair method and deep learning technology, applied in image data processing, instruments, biological neural network models, etc., can solve the problems of complex process and low accuracy, and achieve the effect of improving repair accuracy and enhancing extraction ability.

Pending Publication Date: 2020-11-24
PLA STRATEGIC SUPPORT FORCE INFORMATION ENG UNIV PLA SSF IEU
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a deep learning-based DSM local missing repair method to solve the problems of low precision and complicated process in the current DSM partial correct repair method

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
  • DSM local missing repairing method based on deep learning
  • DSM local missing repairing method based on deep learning
  • DSM local missing repairing method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0047] The repair method of the present invention applies the deep learning image repair method to DSM repair. The algorithm can effectively reduce the repair error by combining partial convolution and attention modules on the basis of U-Net, and has better robustness sex. Among them, partial convolution can enhance the ability to extract irregular and missing edge features; the attention module can increase the feature weight adaptive learning mechanism in the two dimensions of channel and space.

[0048] Specifically, the repair model of the present invention uses a partially convolutional U-Net network plus an attention module, such as figure 1 As shown, it includes three modules: feature extraction module, channel fusion module and resolution restoration module. The input DSM first extracts features layer by layer through the feature e...

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 relates to a DSM local missing repairing method based on deep learning, and belongs to the technical field of DSM data processing. According to the method, a repair model comprising a feature extraction module, a channel fusion module and a resolution recovery module is constructed, and the feature extraction module performs feature extraction by adopting partial convolution, so thatthe shape of a convolution kernel can be randomly changed according to the shape of a mask, and the extraction capability of irregular missing edge features is enhanced; the channel fusion module adds feature weights in two dimensions of a channel and a space, and important features are selected for restoration; and the resolution recovery module recovers the resolution of the feature map in an up-sampling and partial convolution mode. By means of the restoration model, the restoration precision is higher on the whole, and meanwhile better robustness is achieved in the aspect of missing proportion change.

Description

technical field [0001] The invention relates to a deep learning-based DSM local defect repair method, which belongs to the technical field of DSM data processing. Background technique [0002] Digital Surface Model (DSM) refers to the ground elevation model including the height of surface buildings, bridges and trees. It plays an important role in military, urban planning and other fields. Its acquisition methods are mainly obtained through three methods: airborne lidar, digital photogrammetry workstation, and high-resolution satellite remote sensing image stereo pair. Among them, airborne radar acquisition is a kind of active earth observation, which is relatively less affected by weather, light, etc., and has high data acquisition accuracy and faster speed than the latter two. It is especially suitable for DSM acquisition for military affairs and sudden natural disasters. Especially in recent years, with the continuous development of sensor acquisition technology and the...

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): G06T5/00G06T5/50G06N3/04
CPCG06T5/005G06T5/50G06T2207/10032G06T2207/10044G06T2207/20221G06N3/045
Inventor 金飞刘智官恺韩佳容芮杰王淑香林雨准谢功健
Owner PLA STRATEGIC SUPPORT FORCE INFORMATION ENG UNIV PLA SSF IEU
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap