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

Convolutional neural network hyperspectral image classification method based on edge preserving

A convolutional neural network and hyperspectral image technology, applied in the field of hyperspectral image classification based on edge-preserving convolutional neural network, can solve the problem of less samples for training the network, difficulty in obtaining hyperspectral images, and insufficient training of deep networks, etc. problem, to achieve the effect of good classification and good classification performance

Inactive Publication Date: 2019-11-15
TIANJIN UNIV
View PDF6 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The existing methods are poor in coping with the noise and redundancy existing in hyperspectral images, which makes it difficult for the network to achieve better classification performance; it is very difficult to obtain hyperspectral images, resulting in fewer samples that can be used to train the network, and the depth The network cannot be fully trained

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
  • Convolutional neural network hyperspectral image classification method based on edge preserving
  • Convolutional neural network hyperspectral image classification method based on edge preserving
  • Convolutional neural network hyperspectral image classification method based on edge preserving

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] The embodiment of the present invention proposes a convolutional neural network hyperspectral image classification method based on edge preservation. The method first performs edge preservation filtering on the original hyperspectral image, and then preprocesses the original hyperspectral image and the filtered image separately And spatial enhancement, and then use the convolutional network to extract spatial-spectral features, the specific implementation steps are as follows:

[0034] 1. Perform edge-preserving filtering on the original hyperspectral image

[0035] Edge preserving filtering is a non-linear filter, which uses different filtering cores to process pixels in different positions, so as to smooth the image, filter out image noise, while maintaining the shape of the edge, reducing the edges that appear after filtering Blurred situation. Commonly used edge preserving filters include bilateral filtering, guided filtering, and domain transform recursive filtering. T...

Embodiment 2

[0058] Combine below figure 2 The feasibility verification of the scheme in Example 1 is carried out, as detailed in the following description:

[0059] figure 2 The classification results of the data set of the University of Pavia are given. The first column is a pseudo-color composite map, the second column is a true value map, the third column is a classification map obtained by an embodiment of the present invention, and the fourth column is a different color The corresponding category.

[0060] It can be seen from the results that the hyperspectral image classification method in the embodiment of the present invention obtains a classification map with less noise, and can also maintain a good classification effect for edge pixels, and the edge contour is clear.

[0061] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of a preferred embodiment, and the serial numbers of the above-mentioned embodiments of the present invention ar...

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 convolutional neural network hyperspectral image classification method based on edge preserving. The convolutional neural network hyperspectral image classification method comprises the following steps: carrying out the edge preserving filtering of an original hyperspectral image, and obtaining a filtered image; respectively selecting pixels in central pixel windows of the original hyperspectral image and the filtered image as pixel blocks; rotating and converting the pixel block to obtain a pixel block after data enhancement; and taking the pixel block after data enhancement as the input of a convolutional neural network, and extracting the spatial-spectral features through the convolutional neural network. According to the invention, edge filtering is carried out on the hyperspectral images to realize classification of the hyperspectral images.

Description

Technical field [0001] The invention relates to the fields of image processing and hyperspectral image processing, and in particular to a convolutional neural network hyperspectral image classification method based on edge preservation. Background technique [0002] Hyperspectral remote sensing technology is a comprehensive earth observation technology that emerged in the early 1980s and has received great attention from the industry. Hyperspectral remote sensing uses an imaging spectrometer that covers a certain spectral range and uses a large number of narrow electromagnetic wavebands to obtain ground object information. With the continuous improvement of the performance of hyperspectral remote sensing sensors, the spectrum coverage is getting wider and the spectral resolution is getting higher and higher, which can provide very rich information in the identification and detection of ground objects. Therefore, hyperspectral remote sensing technology is used in military, Agricu...

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): G06K9/00G06K9/46G06K9/62G06T3/60G06T5/20
CPCG06T3/602G06T5/20G06T2207/20024G06T2207/10032G06T2207/20081G06T2207/20084G06T2207/20192G06V20/13G06V20/194G06V10/443G06F18/2414
Inventor 雷建军李鑫宇韩梦芯李奕石雅南杨博兰
Owner TIANJIN 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