Image change detection method based on deep neural network structure optimization

A technology of image change detection and network structure, which is applied in the field of image processing, can solve problems such as reduced flexibility, time and effort, and poor generalization performance, and achieve the effects of saving time and overhead, improving generalization performance and flexibility

Active Publication Date: 2019-09-06
XIDIAN UNIV
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method uses a hand-designed dual-channel convolutional neural network to extract the representation vector of the image, which improves the accuracy of change detection for SAR images to a certain extent, but the disadvantage is that the network can be used for other images. When the dete

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
  • Image change detection method based on deep neural network structure optimization
  • Image change detection method based on deep neural network structure optimization
  • Image change detection method based on deep neural network structure optimization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] Below in conjunction with accompanying drawing and specific embodiment, the present invention is described in further detail:

[0033] refer to figure 1 , the implementation steps of this example are as follows:

[0034] Step 1 reads in the image and performs normalization processing.

[0035] First, read in the two registered and corrected images taken at the same location in different phases I x and I y , the purpose of registration and correction is to geometrically match images of different time phases, different bands or different types, so that the images have uniform geographic coordinates and pixel spatial resolution;

[0036] Then, for the two images I x and I y Perform normalization processing respectively to obtain two normalized images I 1 and I 2 , respectively as follows:

[0037]

[0038]

[0039] Step 2 obtains the training sample set.

[0040] For the normalized two images I 1 and I 2 10% of the image area in the image is marked, and th...

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 image change detection method based on deep neural network structure optimization, and mainly solves the problems of low detection generalization performance and low flexibility in an existing image change detection method. According to the implementation scheme, the method comprises the following steps: reading an image, performing normalization processing, constructinga training sample set, and encoding a deep neural network structure; initializing a population according to a network structure coding mode; based on the training sample set, searching an optimal network structure by using an evolutionary algorithm; and outputting an image change detection result graph by using the optimal network structure. The optimal deep neural network structure of a given task can be efficiently and automatically searched based on a search strategy of an evolutionary algorithm, good effects and robustness are achieved for various complex conditions, the flexibility of change detection is improved, and the method can be used for land cover detection, disaster assessment and video monitoring.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an image change detection method, which can be used for land cover detection, disaster assessment and video monitoring. Background technique [0002] Image change detection refers to the use of multi-temporal images covering the same surface area and other auxiliary data to determine and analyze surface changes. It uses a computer image processing system to identify and analyze the changes of objects or phenomena in different periods of time, and can determine the changes of objects or phenomena within a certain time interval, and provide qualitative and quantitative information on the spatial distribution of objects and their changes. The key to image change detection is that under the given data set, appropriate multi-level image features can be extracted through algorithmic modeling, so as to enhance the change information of the image. [0003] The curre...

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): G06T7/00G06T7/30G06T5/00G06N3/04
CPCG06T7/0002G06T7/30G06T5/006G06T2207/10004G06T2207/20081G06N3/045
Inventor 刘若辰张浪浪张豪刘静李阳阳慕彩红
Owner XIDIAN UNIV
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