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

SAR image monitoring and classifying method based on conditional random field model

A conditional random field, supervised classification technology, applied in the field of image processing, can solve the problems of too many basic units, insufficient description of complex scene models, affecting the efficiency of image classification, etc., to achieve the effect of fast training and inference.

Inactive Publication Date: 2012-10-31
WUHAN UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Such information is difficult to be mined and utilized in pixel-based processing
In addition, pixel-level processing often results in too many basic units to be processed, which greatly affects the efficiency of image classification.
Although traditional statistical models are suitable for describing homogeneous areas, such as water and cultivated land, they are not sufficient for describing complex scene models

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
  • SAR image monitoring and classifying method based on conditional random field model
  • SAR image monitoring and classifying method based on conditional random field model
  • SAR image monitoring and classifying method based on conditional random field model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] The technical solution of the present invention is described in detail below in conjunction with accompanying drawing and embodiment, and embodiment comprises steps as follows:

[0016] Step 1. Establish the conditional random field model of the regional connection graph:

[0017] Firstly, the image is over-segmented into multiple regions, which can be realized by using the existing MeanShift algorithm during specific implementation, which will not be described in detail in the present invention.

[0018] Suppose the image is over-segmented into Q regions, recorded as S={S 1 , S 2 ..., S Q}, the corresponding label set is: T={X i , i∈(1,2,…,Q)}, where X i The corresponding value is L={1, 2, . . . , K}, where L is a discrete symbol set. Then all possible labeled states (ie solution space) of T have L Q indivual. The region adjacency graph G=(S, E) is built on these over-segmented regions, each region S is regarded as a node, and E represents the edge connecting 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 provides an SAR image monitoring and classifying method based on a conditional random field model, comprising: cutting an image into a plurality of regions, describing the plurality of regions into a region adjacency graph (RAG), and building the conditional random field (CRF) model. A Max-margin algorithm is adopted for parameter learning, and a GraphCut algorithm is used for optimizing and reasoning. When the method of the invention is used for SAR image classification, more complicated SAR image characteristics and suitable relationship between image contexts are combined to obtain more Luban classification results. The establishment of the region adjacency graph causes the model to have big speed advantages on training and test, so that the invention is suitable for classifying large-scale SAR images.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to SAR image supervised classification, and is a SAR image supervised classification method based on a conditional random field model. Background technique [0002] With the increase of the number of SAR images and the image resolution, the traditional pixel-based processing is no longer able to meet the needs of high-level interpretation of SAR images. This is mainly because as the resolution increases, SAR images already contain rich structure and texture information, and these structure and texture information carry strong context information, such as streets and houses often appear together, bridges and rivers Also often appear together etc. Such information is difficult to be mined and utilized in pixel-based processing. In addition, pixel-level processing often leads to too many basic units to be processed, which greatly affects the efficiency of image classification. Althou...

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): G06K9/62
Inventor 杨文代登信
Owner WUHAN UNIV
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