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

Region of Interest Extraction Method for Large Format Remote Sensing Imagery Based on Visual Attention Mechanism

A technology of visual attention mechanism and region of interest, applied in the field of remote sensing image processing, can solve the problem of lack of the ability to search for potential targets in large-scale remote sensing images, and achieve the effect of extraction

Active Publication Date: 2019-01-08
HARBIN INST OF TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In order to solve the lack of ability to search for potential targets in large-scale remote sensing images in existing remote sensing image processing technology, the present invention proposes a large-scale remote sensing image interest region extraction method based on visual attention mechanism

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
  • Region of Interest Extraction Method for Large Format Remote Sensing Imagery Based on Visual Attention Mechanism
  • Region of Interest Extraction Method for Large Format Remote Sensing Imagery Based on Visual Attention Mechanism
  • Region of Interest Extraction Method for Large Format Remote Sensing Imagery Based on Visual Attention Mechanism

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0037] Embodiment 1. The method for extracting a region of interest from a large-scale remote sensing image based on a visual attention mechanism described in this embodiment is carried out in the following steps:

[0038] Step 1. Target and ROI characteristic analysis:

[0039] Characteristic analysis is the basis of region of interest extraction (ROI). The characteristics of the target and the region of interest can be used as prior knowledge for the next level of ROI extraction. The distribution density of certain geometric features of the target can reflect the position distribution of the target. , so the characteristic analysis of the target and the region of interest can largely determine the accuracy of ROI extraction.

[0040] Target and area of ​​interest characteristics analysis, respectively analyze the target characteristics of man-made targets with regular shape and regular distribution, and irregular-shaped natural features, and study the distribution of man-mad...

specific Embodiment approach 2

[0058] Specific Embodiment 2. This embodiment is a further description of the method for extracting regions of interest from large-format remote sensing images based on the visual attention mechanism described in Specific Embodiment 1. The method described in Step 1 is aimed at artificial objects with regular shapes and regular distribution. The commonly used geometric parameters when analyzing the target and the target characteristics of irregularly shaped natural features are: ①Circularity C 1

[0059] C 1 =P 2 / A 0 (1)

[0060] where P is the perimeter of the object, A 0 is the area of ​​the object. When the object is circular, the minimum value of circularity is 4π, and the more complicated the shape, the larger the value;

[0061] ②Body ratio C 2

[0062] C 2 =W / L (2)

[0063] Among them, W is the width of the smallest circumscribing rectangle of the object, L is the length of the smallest circumscribing rectangle of the object, and when the object is a square o...

specific Embodiment approach 3

[0073] Specific embodiment three, this embodiment is a further description of the large-scale remote sensing image region of interest extraction method based on the visual attention mechanism described in the specific embodiment one or two, and the Itti model described in the step 2 is to the input image I(x ,y) Gaussian pyramid G(x,y,σ) is used to perform non-uniform sampling at different levels, which is defined as follows:

[0074]

[0075] in, is the convolution operator;

[0076] The brightness of the image and the components on the red, green, blue, and yellow color channels of the original image are expressed as:

[0077]

[0078] R(x)=r(x)-(g(x)+b(x)) / 2

[0079] G(x)=g(x)-(r(x)+b(x)) / 2 (7)

[0080] B(x)=b(x)-(r(x)+g(x)) / 2

[0081] Y(x)=(r+g) / 2-|r(x)-g(x)| / 2-b(x)

[0082] Among them, I(x) is the image brightness, R(x) is the component on the red channel, G(x) is the component on the green channel, B(x) is the component on the blue channel, Y(x) is yellow co...

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

A method for extracting a region of interest from a large format remote sensing image based on a visual attention mechanism relates to the technical field of remote sensing image processing, in particular to a method for extracting a region of interest from a large format remote sensing image based on a visual attention mechanism. The invention aims to solve the problem that the existing remote sensing image processing technology lacks the ability to search for potential targets in large-format remote sensing images. The present invention is carried out according to the following steps: 1. Target and ROI characteristic analysis; 2. ROI extraction based on visual attention mechanism; 3. Level 1 ROI extraction based on bottom-up stimulation driving mechanism; 4. Secondary region-of-interest extraction for a top-down object-driven mechanism. On the basis of analyzing the characteristics of the target and the region of interest, the invention introduces a psychological model and a calculation model of a visual attention mechanism to study a method for extracting a region of interest from a large-scale remote sensing image. The invention can be applied to the technical field of remote sensing image processing.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing, in particular to a method for extracting a region of interest from a large format remote sensing image based on a visual attention mechanism. Background technique [0002] With the improvement of satellite resolution, the data width and data volume of high-resolution remote sensing satellites increase sharply. The increase in image data volume and complexity makes it more difficult to automatically identify targets, and the corresponding data processing technology is also difficult to adapt to real-time Requirements for effective processing; and remote sensing images are often managed by satellite shooting strips, the data width can reach more than ten or even tens of kilometers, and the data volume can reach more than 1GB. The efficiency of manual interpretation is low, which can no longer meet the application requirements; the existing The images used in most research me...

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): G06K9/32G06T7/00
CPCG06T2207/10032G06V10/25
Inventor 张钧萍李彤毛宇
Owner HARBIN INST OF 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