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

SAR Image Segmentation Method Based on Wavelet Pooling Convolutional Neural Network

A convolutional neural network and image segmentation technology, applied in the field of image processing, can solve problems such as structural damage, unfavorable SAR image segmentation, etc., and achieve the effect of maintaining consistency

Active Publication Date: 2018-03-06
XIDIAN UNIV
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the traditional pooling methods in convolutional neural networks, that is, maximum pooling and average pooling, simply take the maximum value or average value in the neighborhood window of the feature map, which will affect the structure of the learned features. Certain damage is not conducive to the segmentation of SAR images

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 Segmentation Method Based on Wavelet Pooling Convolutional Neural Network
  • SAR Image Segmentation Method Based on Wavelet Pooling Convolutional Neural Network
  • SAR Image Segmentation Method Based on Wavelet Pooling Convolutional Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] Reference figure 1 , The implementation steps of the present invention are as follows:

[0023] Step 1. Construct a wavelet pooling layer and form a wavelet pooling convolutional neural network.

[0024] (1.1) According to the following formula, the characteristic map CF k Perform wavelet transform,

[0025]

[0026] Where k=1,...,N, N is the number of feature maps: CF k η(x,y) Is feature map CF k In the neighborhood block centered at the point (x, y), * represents the convolution operation, ψ is the wavelet basis function, and down(·) represents the down 2 sampling operation, Yes The points in each subband obtained by wavelet transform, a, h, v, d represent approximate subband, horizontal subband, vertical subband and diagonal subband respectively;

[0027] (1.2) Use approximate subband feature map SF k a Form the wavelet pooling layer, where k=1,...,N, N is the number of feature maps;

[0028] (1.3) Use the wavelet pooling layer to form a wavelet pooling convolutional neura...

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 SAR image segmentation method based on wavelet pooling convolutional neural network. The scheme is: 1. Construct the wavelet pooling layer and form a wavelet pooling convolutional neural network; 2. Select image blocks and input them into the wavelet pooling convolutional neural network for training; 3. Input all image blocks into the training Test in a good network to get the first category of SAR image; 4. Segment the SAR image with superpixels, and fuse the result with the first category of SAR image to get the second category of SAR image; 5. According to Markov random The airport model obtains the third category of SAR image, and fuses it with the result of superpixel segmentation to obtain the fourth category of SAR image; 6. According to the SAR image gradient map, the second category and the fourth category of SAR image fusion to get the final segmentation result. The invention improves the segmentation effect of the SAR image and can be used for target detection and recognition.

Description

Technical field [0001] The invention belongs to the technical field of image processing, relates to a SAR image segmentation method, and can be used for target detection and recognition. Background technique [0002] With the rapid development of remote sensing technology, SAR image interpretation is playing an increasingly important role in energy, environment, and archaeology. SAR image segmentation is a basic and important part of SAR image interpretation. It can provide help for subsequent classification, detection, recognition and tracking. The main goal of SAR image segmentation is to divide the SAR image into connected regions without intersection. To accomplish this goal, each pixel of the SAR image needs to be labeled, so it is a very difficult task. [0003] The existing SAR image segmentation methods are mainly divided into feature-based methods and statistical model-based methods. Feature-based methods are mainly to extract some SAR image features for segmentation, su...

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): G06T7/10
Inventor 刘芳段一平郝红侠焦李成李玲玲尚荣华马文萍杨淑媛马晶晶
Owner XIDIAN 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