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

A Mineral Content Spectral Inversion Method Based on Deep Neural Network

A deep neural network and neural network technology, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve problems such as optimal solution, overfitting, data acquisition difficulties, etc., to avoid falling into local optimal solutions, The effect of reducing the amount of sample data

Active Publication Date: 2021-10-19
BEIJING RES INST OF URANIUM GEOLOGY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] As a nonlinear analysis method, neural network has been widely used in mineral spectral analysis, but the current application still has the following problems: shallow neural network requires a large number of spectral sample data with known components, and it is difficult to obtain data; it is easy to fall into local optimal solution and the phenomenon of overfitting

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 Mineral Content Spectral Inversion Method Based on Deep Neural Network
  • A Mineral Content Spectral Inversion Method Based on Deep Neural Network
  • A Mineral Content Spectral Inversion Method Based on Deep Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0060]A kind of mineral content spectral inversion method based on deep neural network of the present embodiment, comprises the following steps:

[0061] Step S1: Collect the mixed spectra of various minerals as samples of the deep neural network, and identify the mineral name and content corresponding to each spectrum.

[0062] Step S2: Set the first layer of deep neural network architecture, and calculate the input layer, hidden layer (ie, spectral feature layer) and output layer forward, such as figure 1 shown.

[0063] For the input layer: convert each sample spectrum into a column vector X as the input layer, the number of neurons in the input layer is the number of bands of the spectrum, +1 is the offset, and the neurons in the input layer are the reflection of each band Rate.

[0064] For the hidden layer: the number of neurons in the hidden layer is the number of sample mineral types, and the activation value of each neuron in the hidden layer is calculated by using ...

Embodiment 2

[0102] Four minerals including muscovite, calcite, dolomite and feldspar are selected for content inversion, and other minerals can refer to the steps of this embodiment.

[0103] (1) Calculate the single-scattering albedo of the laboratory spectrum of muscovite, calcite, dolomite, and feldspar, randomly set the contents of the four minerals, calculate the mixed single-scattering albedo, and calculate the mixed single-scattering albedo according to the mixed single-scattering albedo rate, and simulated 15,000 spectra of the mixture of four minerals. Among them, 10,000 samples are used as samples of the deep neural network, and the mineral name and content corresponding to each spectrum are identified, and 5,000 samples are used as test data.

[0104] (2) According to the number of samples and mineral types, a three-layer neural network is initialized, the input neurons are 420 bands, the output neurons (mineral types) are 4, the first layer of hidden layer neurons is 100, and ...

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 belongs to the technical field of geological exploration, and in particular relates to a mineral content spectrum inversion method based on a deep neural network. According to the mechanism and inversion characteristics of the rock mineral spectrum, the present invention establishes a neural network model capable of quickly and accurately inverting the mineral content of the surface rock and soil, and can determine, analyze and evaluate the mineral content of the surface rock and soil by using hyperspectral data.

Description

technical field [0001] The invention belongs to the technical field of geological exploration, and in particular relates to a mineral content spectrum inversion method based on a deep neural network. Background technique [0002] The rapid development of hyperspectral technology has made it possible to use spectral information of tens to hundreds of "narrow" bands in succession for mineral type identification, content inversion, and chemical composition identification. The spectral resolution of currently available data sources (mainly aviation and ground hyperspectral equipment) can support the analysis and extraction of mineral information to a certain extent, and on this basis, hyperspectral thematic geoscience data products can be produced (including mineral distribution and abundance maps, hydrothermal activity fault structures, strata, and rock mass interpretation maps), these information, which were difficult to obtain in the past with multi-spectral remote sensing te...

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): G06N3/08G06N3/06G06N3/04G01N21/31
CPCG06N3/084G06N3/061G01N21/31G06N3/047
Inventor 秦凯赵英俊赵宁博杨越超
Owner BEIJING RES INST OF URANIUM GEOLOGY
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