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

Method for automatically marking category of remote sensing image based on author-genre theme model

A remote sensing image and topic model technology, applied in the field of digital images, can solve problems such as the decline in classification accuracy and achieve the effect of improving accuracy

Inactive Publication Date: 2012-12-12
UNIV OF ELECTRONIC SCI & TECH OF CHINA
View PDF2 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in the case of many regional categories, due to the existence of different types of regions with similar characteristics, the classification accuracy of the ATM-based remote sensing image category labeling method is severely reduced.

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
  • Method for automatically marking category of remote sensing image based on author-genre theme model
  • Method for automatically marking category of remote sensing image based on author-genre theme model
  • Method for automatically marking category of remote sensing image based on author-genre theme model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026] The present invention realizes on Matlab R2010a experimental platform, mainly comprises three steps, specifically as follows:

[0027] 1. Training and author-genre topic model generation steps:

[0028] 1) Use the image representation method to represent the training images as visual words, where each training image has a corresponding category label and scene label, and set the visual word w i and author x a , theme z t and genre c g The total number of , and the visual word w i respectively with author x a and genre c g matching relationship. sight word w i ∈{1,2,…,k}, where k is the number of different visual words obtained according to the training results. The visual words are used to represent the quantized affine invariant regions (some similar blocks) of each image block divided by the training image. Each image patch can be mapped to a visual word through the process of clustering. author x a Represents the category label of the image block, genre c ...

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 a method for automatically marking a category of a remote sensing image, for increasing category marking precision under the conditions of more area categories and characteristic similarities of different areas. The method for automatically marking the category of the remote sensing image based on an author-genre theme model comprises the following steps: 1) training and generating the author-genre theme model; 2) calculating visual words of the remote sensing image; and 3) judging and marking the category of the remote sensing image: adding the information about genre of the author so as to endow a same image block with author and genre two marking information, estimating so as to obtain the genre information of the image block, and then correcting the estimated author information, thereby effectively increasing the category marking precision.

Description

technical field [0001] The invention relates to digital image technology, in particular to remote sensing image intelligent processing technology. Background technique [0002] Remote sensing image annotation is one of the important contents of remote sensing image analysis and understanding. Remote sensing image annotation refers to assigning data to remote sensing images in the form of annotations, for example, automatically marking commercial areas, residential areas, farmland and other categories in the image. Remote sensing image annotation plays an important role in remote sensing image retrieval system. [0003] The process of automatically classifying remote sensing images is the process of classifying multiple types of targets in the images according to the predefined remote sensing image target classes. [0004] The existing category labeling methods of remote sensing images are mainly divided into three categories: (1) methods based on machine learning; (2) meth...

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): G06K9/62
Inventor 李宏亮罗旺
Owner UNIV OF ELECTRONIC SCI & TECH OF CHINA