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

A hyperspectral remote sensing image restoration method based on non-convex low-rank sparse constraints

A hyperspectral remote sensing and sparse-constrained technology, which is applied in the field of hyperspectral remote sensing image restoration based on a non-convex low-rank sparse constrained model, can solve the problems affecting the quality of image restoration and the convergence speed cannot meet the requirements, and achieve the preservation of image details and high Image recovery quality, to achieve the effect of recovery

Active Publication Date: 2021-08-24
HARBIN INST OF TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, such algorithms mostly use convex relaxation with low-rank constraints or sparse constraints. 1 norm to construct the constrained optimization model, but l 1 Due to the shrinkage effect, the norm sometimes has an estimation bias, which affects the quality of image restoration; in addition, when the dimension of the image data matrix increases, the convergence speed often cannot meet the requirements

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
  • A hyperspectral remote sensing image restoration method based on non-convex low-rank sparse constraints
  • A hyperspectral remote sensing image restoration method based on non-convex low-rank sparse constraints
  • A hyperspectral remote sensing image restoration method based on non-convex low-rank sparse constraints

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0096] Example 1: This example selects a typical dataset in the field of remote sensing——EO-1Hyperion Australia dataset. The size of the original image is 3858×256×242. Due to space limitations, after removing the overlapping bands in the range of visible light, near-infrared and short-wave infrared , select a size of 400×200×150, subtract the minimum pixel value from each pixel value, and then divide by the difference between the maximum and minimum pixel values ​​to obtain the data normalized to [0, 1]. Use the method proposed by the present invention to carry out image recovery processing, and use the above-mentioned LRMR, LRTV, NAILRMA to carry out comparative experiments. Experimental results such as figure 2 shown, where figure 2 (a) is the original hyperspectral remote sensing image of the 52nd band, figure 2 (b) is the effect of the LRMR method, figure 2 (c) is the effect of LRTV method, figure 2 (d) is the effect of the NAILRMA method, figure 2 (e) is the e...

example 2

[0097] Example 2: This example selects a typical data set in the field of remote sensing—Hyperspectral Digital ImageryCollection Experiment (HYDICE) Washington DC Mall data set. The size of the original image is 1208×307×191. Due to space limitations, the selected image size is 256×256×11 . Same as Example 1, the original image data is normalized, and Gaussian noise with an average value of 25dB is randomly added to all bands artificially. Then use the method proposed by the present invention to restore and compare with other methods. Experimental results such as image 3 shown, where image 3 (a) is one of the clear hyperspectral remote sensing image data, image 3 (b) is the image after adding Gaussian noise and strip noise, image 3 (c) is the effect of LRMR method (SNR=10.8786dB, MSSIM=0.66377), image 3 (d) is the effect of LRTV method (SNR=10.6603dB, MSSIM=0.74259), image 3 (e) is the NAILRMA method effect (SNR=15.5521dB, MSSIM=0.83878), image 3 (f) is the effec...

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 restoration method of hyperspectral remote sensing image based on non-convex low rank sparse constraint belongs to the field of hyperspectral remote sensing image processing in remote sensing image processing. , including the following steps: input the hyperspectral remote sensing image; initialize the weight coefficient matrix, the number of iterations and the convergence threshold, initialize the sub-image size and scan step size, and divide the sub-blocks; establish an image restoration model; Iteratively solve the value minimum algorithm; judge whether the restoration result meets the convergence condition; obtain the required hyperspectral restoration image when the number of iterations is reached, otherwise return to the corresponding step to continue the iterative operation; calculate the weight coefficient matrix and assign appropriate weights to each sub-block; restore Hyperspectral remote sensing images are obtained to obtain the final restored hyperspectral remote sensing images. The denoising effect is obvious and the image details are preserved.

Description

technical field [0001] The invention belongs to the field of hyperspectral image processing in remote sensing image processing, and in particular relates to a hyperspectral remote sensing image restoration method based on a non-convex low-rank sparse constraint model. [0002] Background of the invention [0003] Hyperspectral remote sensing imaging technology combines spectral analysis and optical imaging technology to detect the two-dimensional geometric space and one-dimensional spectral information of the target, and obtain high-resolution continuous and narrow-band image data. At present, hyperspectral imaging technology is developing rapidly, and it is widely used in environmental research, geological exploration, military surveillance and other fields because of its rich spectral information of ground objects. However, due to the physical defects of the sensor, photon effects, transmission loss, and calibration errors, the hyperspectral remote sensing images obtained i...

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): G06T5/00G06T7/11
CPCG06T7/11G06T2207/10032G06T2207/20021G06T2207/20192G06T5/70
Inventor 胡悦李晓迪赵旷世苑鑫
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