Hyperspectral image denoising method based on non-convex low rank matrix decomposition

A technique of hyperspectral image, low rank matrix, applied in the field of hyperspectral image denoising

Inactive Publication Date: 2017-05-24
SHANGHAI ZENGRONG DATA TECH CO LTD
View PDF4 Cites 42 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0013] Purpose of the invention: This invention aims to solve the technical problem of hyperspectral image denoising in the field of remote sensing images, and provides a hyperspectral image denoising method based on non-convex low-rank matrix decomposition, thereby effectively removing the mixed noise existing in hyperspectral images Simultaneously preserve spectral and spatial features in 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
  • Hyperspectral image denoising method based on non-convex low rank matrix decomposition
  • Hyperspectral image denoising method based on non-convex low rank matrix decomposition
  • Hyperspectral image denoising method based on non-convex low rank matrix decomposition

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0147]The experimental hardware environment for this implementation is: Intel-Core4i76700 3.4GHz, 16G memory, graphics card NVIDIAGeForce GTX950. The software environment is MATLAB 2015b. The test images are derived from hyperspectral image datasets published on the Internet: Hyperspectral Digital Imagery Collection Experiment (HYDICE) urbandataset, the Earth Observing-1 (EO-1) Hyperion Australia dataset, and the Airborne Visible / InfraredImaging Spectrometer (AVIRIS) Indian Pines dataset. In order to verify the effectiveness of the method, the present invention has selected an advanced denoising method proposed in recent years: Video Block Matching 3-D filtering method (Video Block Matching 3-D filtering, VBM3D [1] ), low-rank matrix recovery method (Low-RankMatrix Recovery, LRMR [2] ), noise-adaptive iterative low-rank matrix approximation method (Noise-AdjustedIterative Low-Rank Matrix Approximation, NAILRMA [3] ), a weighted p-norm-based low-rank matrix approximation met...

example 1

[0148] Example 1: This example selects a typical data set in the field of remote sensing—HYDICEurban data set, which is an image data polluted by Gaussian noise, impulse noise, stripes, atmosphere, hydrology, etc. It is an ideal data for evaluating hyperspectral denoising methods. The size of the whole image is 307×307, which contains 210 bands in total. It is easy to see that the method in the present invention and other comparison methods can remove the noise in the lightly noise-polluted band, and also preserve the local detail information of the hyperspectral image (such as figure 2 shown, where figure 2 (a) represents the original noise-containing image, (b)-(f) are images after denoising by algorithms VBM3D, LRMR, NAILRMA, WSN-LRMA, and NonLRMA respectively, and the rest of the images are operated in the same way); for moderate For noise-contaminated bands, some denoising algorithms cannot effectively remove noise. On the contrary, this method not only removes various...

example 2

[0149] Example 2: This example selects a typical dataset in the field of remote sensing—EO-1Hyperion Australia dataset. This dataset was shot on December 4, 2012. The original image size is 256×3858 and contains 242 bands in total. Due to space limitations, this method selects 200×400 sub-image blocks. This data set mostly contains vertical stripes, such as Figure 5-Figure 8 shown. It is easy to see that only this method can effectively remove the band noise and preserve the local detail information of the image.

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 present invention discloses a hyperspectral image denoising method based on non-convex low rank matrix decomposition. The method comprises the following steps: 1. dividing a hyperspectral image; 2. generating a matrix; 3. carrying out nonconvex low rank matrix decomposition; 4. restoring the sub-block; and 5. restoring the hyperspectral image, and finally obtaining the denoised hyperspectral image. According to the method disclosed by the present invention, a method for efficiently and quickly removing multiple kinds of noises is provided for the remote sensing hyperspectral image denoising.

Description

technical field [0001] The invention belongs to the fields of computer image, information processing technology, remote sensing technology and the like, and in particular relates to a hyperspectral image denoising method based on non-convex low-rank matrix decomposition. Background technique [0002] With the rapid development of digital remote sensing technology, hyperspectral images are widely used in military reconnaissance, environmental science, geological exploration, biomedical imaging and other fields. Due to the influence of various factors such as mechanical failure of the sensor and image transmission failure, the hyperspectral image will inevitably be polluted by various noises during the acquisition and transmission process, such as Gaussian noise, pulse noise, stripes, etc. Further applications of hyperspectral imagery. At the same time, the rapid increase in the dimensionality of hyperspectral images has led to the "curse of dimensionality". Therefore, desig...

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 Applications(China)
IPC IPC(8): G06T5/00
CPCG06T5/002G06T2207/10036
Inventor 陈勇勇霍兆坤郭延文
Owner SHANGHAI ZENGRONG DATA TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap