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

Synthetic aperture radar image classifying method based on stacked automatic coding machines

A synthetic aperture radar, automatic encoder technology, applied in computer parts, instruments, character and pattern recognition, etc., can solve the problems of nonlinear characteristics, low classification accuracy, noise pollution of SAR image data, etc.

Inactive Publication Date: 2015-03-04
HARBIN INST OF TECH
View PDF2 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem of low classification accuracy of SAR image data due to serious noise pollution and nonlinear characteristics, and provides a synthetic aperture radar image classification method based on a stacked autoencoder

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
  • Synthetic aperture radar image classifying method based on stacked automatic coding machines

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0024] Specific implementation mode one: the following combination figure 1 Describe this embodiment, the synthetic aperture radar image classification method based on stacked autoencoder described in this embodiment, it comprises the following steps:

[0025] Step 1: Collect N pieces of original SAR images for data preparation, and obtain SAR image data and SAR supervision data;

[0026] Step 2: Obtain training image data and test image data from SAR image data; obtain training supervision data and test supervision data from SAR supervision data;

[0027] Step 3: Use the training image data to pre-train the autoencoder at each layer of the network based on the stacked autoencoder, learn data features and extract features; on the basis of pre-training, use the training supervision data to Perform supervised learning on the resulting network, fine-tune the network, and obtain a trained network based on a stacked autoencoder;

[0028] Step 4: Input the test image data into the...

specific Embodiment approach 2

[0030] Specific implementation mode two: this implementation mode further explains implementation mode one, and the specific method for obtaining SAR image data and SAR supervision data in step one is:

[0031] Set the size of each original SAR image as M×M, read each original SAR image, and convert each original SAR image into M 2 The row vector of N; N original SAR images are traversed to obtain SAR image data, which is N×M 2 The matrix; Determine the corresponding image category label by the original SAR image, as the SAR supervision data, it is a vector of N × 1, the i-th element marks the category label of the i-th original SAR image in the vector, i=1, 2, 3, ... N.

specific Embodiment approach 3

[0032] Specific embodiment three: this embodiment further explains embodiment two, obtain training image data and test image data by SAR image data in step two; Obtain the specific method of training supervision data and test supervision data by SAR supervision data as follows:

[0033] For the SAR image data, starting from the first row, the row vectors are extracted from top to bottom to form training image data; starting from the second row, the row vectors are extracted from top to bottom to form test image data;

[0034] For SAR supervised data, starting from the first row, row vectors are extracted from top to bottom to form training supervision data; starting from the second row, row vectors are extracted from top to bottom to form test supervision data.

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 synthetic aperture radar (SAR) image classifying method based on stacked automatic coding machines, belongs to the technical field of radar image classification, and solves the problem of low classifying precision of SAR image data due to serious noise pollution and nonlinear property. The method comprises the following steps: firstly, acquiring N original SAR images for data preparation to obtain SAR image data and SAR supervision data; secondly, obtaining training image data and test image data according to the SAR image data; thirdly, obtaining training monitoring data and test monitoring data according to the SAR supervision data; fourthly, pre-training all layers of the automatic coding machines by the training image data, performing supervised learning on a network by the training supervision data on the basis of pre-training, and finely adjusting the network to obtain the trained network based on the stacked automatic coding machines; finally, inputting the test image data into the trained network based on the stacked automatic coding machines to classify the SAR images. The method is used for classifying the SAR images.

Description

technical field [0001] The invention relates to a synthetic aperture radar image classification method based on a stacked automatic encoding machine, and belongs to the technical field of radar image classification. Background technique [0002] Synthetic Aperture Radar (SAR) is an active sensor that uses microwaves for perception. Compared with other sensors, SAR can acquire information all-weather and all-day. With the continuous development of SAR technology, more and more SAR images can be obtained, how to effectively use these images has become an urgent problem to be solved. [0003] Due to serious noise pollution and nonlinear characteristics of SAR image data, its classification accuracy is limited. Contents of the invention [0004] The purpose of the present invention is to solve the problem of low classification accuracy of SAR image data due to serious noise pollution and non-linear characteristics, and provides a synthetic aperture radar image classification ...

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
CPCG06V30/194G06F18/24
Inventor 刘柏森张晔陈雨时
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