Unlock instant, AI-driven research and patent intelligence for your innovation.

Edge-guided recurrent convolutional neural network building change detection method and system

An edge-guided, neural network technology, applied in the field of remote sensing image processing, can solve the problems of not fully utilizing the geometric characteristics of buildings, unclear building outlines, and adhesion of detection results, so as to improve change detection performance, clear building outlines, The effect of detecting performance improvements

Active Publication Date: 2021-07-27
HUNAN UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the shortcomings of the above two methods are that they do not make full use of the geometric characteristics of the building, resulting in unclear outlines of the buildings in the detection results and the presence of adhesion in the detection results in densely distributed areas.

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
  • Edge-guided recurrent convolutional neural network building change detection method and system
  • Edge-guided recurrent convolutional neural network building change detection method and system
  • Edge-guided recurrent convolutional neural network building change detection method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] Such as figure 1 As shown, the edge-guided circular convolutional neural network building change detection method of this embodiment includes:

[0033] 1) Remote sensing images T for different time phases 1 and T 2 Perform multi-level feature extraction to obtain multiple feature pairs f 1i and f 2i ,in f 1i is the remote sensing image T 1 on the i The features extracted by the layer, f 2i is the remote sensing image T 2 on the i The features obtained by layer extraction;

[0034] 2) For each feature pair f 1i and f 2i Calculate difference features and perform difference enhancement to obtain corresponding layer difference analysis results p i ;

[0035] 3) The results of the difference analysis of each layer p i Input to the decoder to obtain the decoding result of the corresponding layer by upsampling the deep layer result layer by layer and fusing the shallow layer result q i ;

[0036] 4) The decoding results of each layer q i Input the p...

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 an edge-guided circular convolution neural network building change detection method and system. The invention includes remote sensing image T 1 and T 2 Perform multi-level feature extraction to obtain feature pairs f 1i and f 2i , calculate the difference features and perform difference enhancement, upsample the deep layer results layer by layer and fuse the shallow layer results, then input the edge-guided probability prediction network to obtain a multi-level change probability map and edge probability map, choose T 1 The change probability map with the same size is binarized to obtain the detection result M . The present invention can amplify the difference in the changing area and suppress the difference in the non-changing area, thereby improving the accuracy of change detection, improving the detection performance, and further improving the prior information of the edge structure of the building through edge guidance The performance of change detection is improved, the outline of buildings in the detection results is clearer, and the situation of building adhesion in the detection results of densely distributed areas is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to an edge-guided circular convolutional neural network building change detection method and system. Background technique [0002] Remote sensing images provide land coverage and utilization information, and the dynamic changes of buildings can be monitored by analyzing multi-temporal images in the same area through change detection technology. Building change detection has been widely used in areas such as urban planning, disaster assessment, and supervision of illegal buildings. The core problem of building change detection is to analyze the correlation between bitemporal images. Due to the phenomenon of easy confusion of ground objects and radiation differences in images of different time phases, change detection is a nonlinear task. Methods based on supervised learning can effectively accomplish this task. [0003] In recent years, deep lear...

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): G06K9/00G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V20/176G06V10/454G06N3/048G06N3/044G06N3/045G06F18/24G06F18/214
Inventor 李树涛白北方卢婷
Owner HUNAN UNIV