Bilateral depth feature domain heterogeneous remote sensing image change detection method

A remote sensing image and depth feature technology, applied in neural learning methods, instruments, biological neural network models, etc., can solve the problems of inability to learn targeted image features, poor learning ability, etc., achieve good effects, enhance effects, and overcome illumination uneven effect

Pending Publication Date: 2020-06-19
NANJING UNIV OF SCI & TECH
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method cannot learn targeted image features. At the same time, for the convenience of optimization, the network has only one convolutional layer, and the learning ability is poor.

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
  • Bilateral depth feature domain heterogeneous remote sensing image change detection method
  • Bilateral depth feature domain heterogeneous remote sensing image change detection method
  • Bilateral depth feature domain heterogeneous remote sensing image change detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] Such as figure 1 As shown, a method for detecting changes in heterogeneous remote sensing images in bilateral depth feature domains, the steps are as follows:

[0045] Step 1: Input two heterogeneous remote sensing images:

[0046] Separate the optical images with different time domain X 1 And synthetic aperture radar image X 2 Enter into the network from both sides.

[0047] Step 2: Construct a symmetrical convolutional neural network with bilateral input:

[0048] (1) Construct as figure 2 The bilateral convolutional deep neural network shown, the neural network has a symmetrical structure, the neuron structure, the number of network layers and the multi-layer structure are the same, the middle layer is the convolution layer, and the two sides are set to the convolution kernel size of 5×5 5-layer convolutional network, the number of units in each layer is C-20-50-50-50, and the network is initialized;

[0049] (2) Separate the optical images X with different time domains 1 An...

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 bilateral depth feature domain heterogeneous remote sensing image change detection method. The method comprises the following steps: constructing a bilateral input symmetricconvolutional neural network; establishing a feature domain constrained probability model; iteratively optimizing the probability model, and solving a difference probability graph; and thresholding the difference probability graph, and outputting a binary change graph. According to the bilateral symmetric convolutional neural network, unsupervised feature automatic extraction of two heterogeneousimages can be realized, the probability of change is measured in a feature domain, the change detection false alarm rate caused by non-uniform illumination, small particle noise and the like can be overcome, the effect of a heterogeneous image difference graph is enhanced, and the change detection accuracy is improved.

Description

Technical field [0001] The invention relates to a change detection technology in remote sensing image analysis, in particular to a method for detecting changes in a heterogeneous remote sensing image in a bilateral depth feature domain. Background technique [0002] Image change detection is a process of detecting changing areas or objects in the same scene by analyzing multiple time domain images. Change detection plays a key role in many image processing applications, such as video surveillance, medical diagnosis, and remote sensing monitoring and evaluation. There are a variety of remote sensing sensors, such as: visible light sensors, multispectral sensors, hyperspectral sensors and synthetic aperture radar (SAR). Therefore, different types of images can generate different varying attributes. At present, global environmental changes are intensified, cities are developing rapidly, and natural disasters such as floods and earthquakes occur frequently. All of these require tim...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V20/13G06N3/047G06N3/045G06F18/2415G06F18/241Y02T10/40
Inventor 刘嘉张皓程肖亮刘芳
Owner NANJING UNIV OF SCI & TECH
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