CN110348538A - A kind of coal rock detection method of multispectral spectral information and 1D-CNN - Google Patents
A kind of coal rock detection method of multispectral spectral information and 1D-CNN Download PDFInfo
- Publication number
- CN110348538A CN110348538A CN201910652387.1A CN201910652387A CN110348538A CN 110348538 A CN110348538 A CN 110348538A CN 201910652387 A CN201910652387 A CN 201910652387A CN 110348538 A CN110348538 A CN 110348538A
- Authority
- CN
- China
- Prior art keywords
- coal
- spectral information
- multispectral
- cnn
- rock detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Abstract
The invention discloses a kind of coal rock detection methods of multispectral spectral information and 1D-CNN, comprising the following steps: (1) coal and the multispectral spectral information of spoil obtain;(2) sample of coal and spoil spectral information divides;(3) one-dimensional convolutional neural networks Spectra feature extraction;(4) probabilistic neural network coal rock detection model construction.The present invention carries out the identification model building of coal and the multispectral spectral information of spoil using 1D CNN-PNN, it is proposed that a kind of new one-dimensional convolutional neural networks model can extract more, more effective characteristic information, and it is possible to prevente effectively from it is highly suitable for real-time, the accurate identification of coal and spoil the problems such as over-fitting.
Description
Technical field
The present invention relates to coal rock detection technical field, the coal rock detection of specifically a kind of multispectral spectral information and 1D-CNN
Method.
Background technique
The basic characteristics of " rich coal, oil-poor, few gas " energy resources determine critical role of the coal in non-renewable energy.?
It, can be with a large amount of spoil and coal by mined together in progress of coal mining.After spoil is mixed with coal, influence will affect
The calorific value of coal influences the quality of coal, while can cause seriously to pollute in combustion to environment.Therefore, by spoil from
Sort out in coal is that coal uses previous vital processing links.In the outer coal preparation technology of Current Domestic, in addition to people
Work is selected except cash, selects cash (coal) technology that can be divided into wet process according to whether using water resource and cash and dry method is selected to select cash automatically.Wet process choosing
Cash needs to consume a large amount of water resource, while the coal slime pollution generated is difficult to handle;Gamma ray and X-ray select the rays such as cash to select
There are certain radiation for cash, and common image recognition selects cash to be interfered greatly by factors such as light.
Multi-optical spectrum imaging technology (Multispectral Imaging, MSI) earliest be applied to military field, later with
Technology continues to develop, and is gradually applied to agriculture various aspects.Multiple and different SPECTRAL REGION images, mostly light are obtained after multispectral imaging
It is narrow and the problem of vulnerable to environmental disturbances such as illumination that spectral imaging technology can solve RGB image wavelength band.Meanwhile with it is traditional
CCD imaging is compared, and multi-optical spectrum imaging technology is outside the image information for obtaining different SPECTRAL REGIONs, moreover it is possible to get different-waveband
Spectral information.Currently, the analytical procedure of feature extraction and pattern-recognition is mainly used for the analysis of multispectral spectral information,
One side processing step is various, on the other hand analyzes result and is affected by feature extraction and mode identification method selection.
Summary of the invention
It is existing to solve the object of the present invention is to provide a kind of coal rock detection method of multispectral spectral information and 1D-CNN
Deficiency existing for coal rock detection method, in real time, accurate identification is produced coal and spoil.
To achieve the goals above, the technical solution adopted by the present invention are as follows: a kind of multispectral spectral information and 1D-CNN's
Coal rock detection method, comprising the following steps:
(1) coal and the multispectral spectral information of spoil obtain;
(2) sample of coal and spoil spectral information divides;
(3) one-dimensional convolutional neural networks Spectra feature extraction;
(4) probabilistic neural network coal rock detection model construction.
Preferably, obtaining the light of the multispectral data of coal and spoil using multi-optical spectrum imaging technology in the step (1)
Spectrum information obtains the multispectral spectroscopic data collection of coal and spoil.
Preferably, in the step (2), using method of random sampling by pretreated coal and the multispectral spectrum of spoil
Data divide be independent training set and test set according to a certain percentage.
Preferably, in the step (3), for extracting the one-dimensional convolutional neural networks (One- of spectral information characteristics
Dimensional convolutional neural network, 1D-CNN) it is a kind of comprising two one-dimensional convolution unit (1D
Conv Block) network structure, mainly include normalization layer, convolutional layer, pond layer, full articulamentum.
Preferably, constructing probabilistic neural net on training set using the spectral signature that CNN is extracted in the step (4)
Network (Probabilistic neural network, PNN) coal rock detection model, determines the parameter of PNN coal rock detection model, so
The effect of identification is tested using test set afterwards, verifies model performance.
Through the above technical solutions, the beneficial effects of the present invention are: obtaining coal and spoil using multi-optical spectrum imaging technology
Multispectral spectral information simultaneously carries out deficiency existing for the existing coal rock detection method of coal rock detection solution;It proposes a kind of new one-dimensional
Convolutional neural networks model can extract more, more effective characteristic information, and it is possible to prevente effectively from the problems such as over-fitting, very
Real-time, accurate identification suitable for coal and spoil.
Detailed description of the invention
Fig. 1 is the flow chart of case study on implementation coal rock detection method of the present invention;
Fig. 2 is the structure diagram for the one-dimensional convolutional neural networks that case study on implementation of the present invention is used to extract feature;
Fig. 3 is the structure chart of case study on implementation 1D Conv Block unit 1 of the present invention;
Fig. 4 is the structure chart of case study on implementation 1D Conv Block unit 2 of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
The present invention carries out under 10 environment of Windows, is analyzed using Keras (V 2.2.4), and will
TensorFlow (V1.10.0) be used as its rear end, hardware using Intel's Duo I7-9700K and it is tall and handsome reach Geforce RTX
2070。
The embodiment of the present invention provides a kind of multispectral spectrum letter to solve all or part of deficiency of present technology
The coal rock detection method of breath and 1D-CNN, the recognition methods include the following steps:
(1) coal and the multispectral spectral information of spoil obtain;
(2) sample of coal and spoil spectral information divides;
(3) one-dimensional convolutional neural networks Spectra feature extraction;
(4) probabilistic neural network coal rock detection model construction.
Understand in order to make the object, technical scheme and advantages of the embodiment of the invention clearer, in conjunction with attached drawing, to of the invention real
A kind of coal rock detection method of the multispectral spectral information and 1D-CNN of applying example offer elaborates, and the embodiment of the present invention provides
A kind of multispectral spectral information and 1D-CNN coal rock detection method, the recognition methods includes the steps that as shown in Figure 1:
101: coal and the multispectral spectral information of spoil obtain, and multispectral imaging module selects five bell Optoelectronics Technology of Shanghai limited
The face the real-time multi-spectral Mosaic type camera of company carries out multispectral spectrum data gathering to multiple samples of coal and spoil,
The multispectral spectroscopic data of coal and spoil is obtained, data volume is 25, wave-length coverage 600-875nm.
102: the sample of coal and spoil spectral information divides, using method of random sampling by pretreated coal and spoil mostly light
The spectroscopic data of spectrum is independent training set and test set according to the ratio cut partition of training set 80% and test set 20%.
103: one-dimensional convolutional neural networks Spectra feature extraction, for extracting the one-dimensional convolution mind of multispectral spectral signature
It is a kind of network structure comprising two one-dimensional convolution units (1D Conv Block) through network (1D-CNN), it is main to include mark
Standardization layer, convolutional layer, pond layer, full articulamentum etc., structure diagram is as shown in Fig. 2, be described as follows:
201: index bed using the batch normalization layer BatchNormalization in Keras, Output Size with it is defeated
Enter consistent;
202: convolutional layer is using the 1D convolutional layer Conv1D in Keras, and for convolution kernel having a size of 3, convolution nuclear volume is 16;
203: pond layer is using the 1D maximum pond MaxPooling1D in Keras, and pond is having a size of 2;
204:1D Conv Block unit 1 uses the convolutional layer and pond layer composition in Keras, structure such as Fig. 3 institute
Show, be described as follows:
301: convolutional layer is using the 1D convolutional layer Conv1D in Keras, and for convolution kernel having a size of 3, convolution nuclear volume is 16;
302: pond layer is using the 1D maximum pond MaxPooling1D in Keras, and pond is having a size of 2;
205:1D Conv Block unit 2 uses the convolutional layer and pond layer composition in Keras, structure such as Fig. 4 institute
Show, be described as follows:
401: convolutional layer is using the 1D convolutional layer Conv1D in Keras, and for convolution kernel having a size of 3, convolution nuclear volume is 32;
402: pond layer is using the 1D maximum pond MaxPooling1D in Keras, and pond is having a size of 2;
Particularly, in 1D Conv Block unit 1 and 1D Conv Block unit 2, the padding of all convolutional layers
Mode is set as same.
206: full articulamentum flattens input using the Dense in Keras, dot product operates;
207: the CNN of the output multispectral spectral signature extracted by the input as PNN, for construct model and
Test model.
104: probabilistic neural network coal rock detection model construction, the multispectral spectral signature extracted using CNN is in training
PNN coal rock detection model is constructed on collection, is determined the parameter of PNN coal rock detection model, is then tested identification using test set
Effect verifies model performance.
Through the above technical solutions, the beneficial effects of the present invention are: obtaining coal and spoil using multi-optical spectrum imaging technology
Multispectral spectral information simultaneously carries out deficiency existing for the existing coal rock detection method of coal rock detection solution;It proposes a kind of new one-dimensional
Convolutional neural networks model can extract more, more effective characteristic information, and it is possible to prevente effectively from the problems such as over-fitting, very
Real-time, accurate identification suitable for coal and spoil.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped
Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should
It considers the specification as a whole, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
The other embodiments being understood that.
Claims (5)
1. a kind of coal rock detection method of multispectral spectral information and 1D-CNN, it is characterised in that: the following steps are included:
(1) coal and the multispectral spectral information of spoil obtain;
(2) sample of coal and spoil spectral information divides;
(3) one-dimensional convolutional neural networks Spectra feature extraction;
(4) probabilistic neural network coal rock detection model construction.
2. a kind of coal rock detection method of multispectral spectral information and 1D-CNN according to claim 1, it is characterised in that:
In the step (1), the spectral information of the multispectral data of coal and spoil is obtained using multi-optical spectrum imaging technology, obtains coal and cash
The multispectral spectroscopic data collection of stone.
3. a kind of coal rock detection method of multispectral spectral information and 1D-CNN according to claim 1, it is characterised in that:
In the step (2), using method of random sampling by pretreated coal and the multispectral spectroscopic data of spoil according to certain ratio
Example divides and is independent training set and test set.
4. a kind of coal rock detection method of multispectral spectral information and 1D-CNN according to claim 1, it is characterised in that:
In the step (3), for extracting the one-dimensional convolutional neural networks (One-dimensional of spectral information characteristics
Convolutional neural network, 1D-CNN) it is a kind of comprising two one-dimensional convolution units (1D Conv Block)
Network structure, mainly include normalization layer, convolutional layer, pond layer, full articulamentum.
5. a kind of coal rock detection method of multispectral spectral information and 1D-CNN according to claim 1, it is characterised in that:
In the step (4), probabilistic neural network (Probabilistic is constructed on training set using the spectral signature that CNN is extracted
Neural network, PNN) coal rock detection model, it determines the parameter of PNN coal rock detection model, is then surveyed using test set
The effect of identification is tried, model performance is verified.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910652387.1A CN110348538B (en) | 2019-07-18 | 2019-07-18 | Multispectral spectral information and 1D-CNN coal and gangue identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910652387.1A CN110348538B (en) | 2019-07-18 | 2019-07-18 | Multispectral spectral information and 1D-CNN coal and gangue identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110348538A true CN110348538A (en) | 2019-10-18 |
CN110348538B CN110348538B (en) | 2023-01-03 |
Family
ID=68179320
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910652387.1A Active CN110348538B (en) | 2019-07-18 | 2019-07-18 | Multispectral spectral information and 1D-CNN coal and gangue identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110348538B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111062403A (en) * | 2019-12-26 | 2020-04-24 | 哈尔滨工业大学 | Hyperspectral remote sensing data depth spectral feature extraction method based on one-dimensional group convolution neural network |
CN111077093A (en) * | 2020-01-10 | 2020-04-28 | 安徽理工大学 | Method and device for quickly detecting coal gangue based on multispectral technology |
CN112345511A (en) * | 2020-11-20 | 2021-02-09 | 安徽师范大学 | Method for detecting organic chlorine pesticide residue of astragalus membranaceus |
CN113138178A (en) * | 2021-04-15 | 2021-07-20 | 上海海关工业品与原材料检测技术中心 | Method for identifying imported iron ore brand |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107909015A (en) * | 2017-10-27 | 2018-04-13 | 广东省智能制造研究所 | Hyperspectral image classification method based on convolutional neural networks and empty spectrum information fusion |
US20180357542A1 (en) * | 2018-06-08 | 2018-12-13 | University Of Electronic Science And Technology Of China | 1D-CNN-Based Distributed Optical Fiber Sensing Signal Feature Learning and Classification Method |
-
2019
- 2019-07-18 CN CN201910652387.1A patent/CN110348538B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107909015A (en) * | 2017-10-27 | 2018-04-13 | 广东省智能制造研究所 | Hyperspectral image classification method based on convolutional neural networks and empty spectrum information fusion |
US20180357542A1 (en) * | 2018-06-08 | 2018-12-13 | University Of Electronic Science And Technology Of China | 1D-CNN-Based Distributed Optical Fiber Sensing Signal Feature Learning and Classification Method |
Non-Patent Citations (1)
Title |
---|
周孟然等: "CNN在煤矿突水水源LIF光谱图像识别的应用", 《光谱学与光谱分析》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111062403A (en) * | 2019-12-26 | 2020-04-24 | 哈尔滨工业大学 | Hyperspectral remote sensing data depth spectral feature extraction method based on one-dimensional group convolution neural network |
CN111077093A (en) * | 2020-01-10 | 2020-04-28 | 安徽理工大学 | Method and device for quickly detecting coal gangue based on multispectral technology |
CN112345511A (en) * | 2020-11-20 | 2021-02-09 | 安徽师范大学 | Method for detecting organic chlorine pesticide residue of astragalus membranaceus |
CN113138178A (en) * | 2021-04-15 | 2021-07-20 | 上海海关工业品与原材料检测技术中心 | Method for identifying imported iron ore brand |
Also Published As
Publication number | Publication date |
---|---|
CN110348538B (en) | 2023-01-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110348538A (en) | A kind of coal rock detection method of multispectral spectral information and 1D-CNN | |
CN110378295A (en) | A kind of coal rock detection method of multispectral image information and the fusion of spectral information isomery | |
CN102879099B (en) | Wall painting information extraction method based on hyperspectral imaging | |
CN110321959A (en) | A kind of coal rock detection method of multispectral image information and CNN | |
CN109145992A (en) | Cooperation generates confrontation network and sky composes united hyperspectral image classification method | |
CN106503739A (en) | The target in hyperspectral remotely sensed image svm classifier method and system of combined spectral and textural characteristics | |
CN109283148A (en) | A method of based on spectral information automatic identification rock forming mineral | |
CN109975273A (en) | A kind of petrographic classification method based on laser induced breakdown spectroscopy | |
CN104251839B (en) | Spectrum separation detection method of compositions of south-Xinjiang red date sample for south-Xinjiang red date modeling | |
CN108764138A (en) | A kind of highlands cloud snow sorting technique cascading forest based on the more granularities of multidimensional | |
Adjorlolo et al. | Predicting C3 and C4 grass nutrient variability using in situ canopy reflectance and partial least squares regression | |
CN109858557A (en) | A kind of new hyperspectral image data semisupervised classification algorithm | |
CN105825227B (en) | A kind of high spectrum image sparse solution mixing method based on MFOCUSS and low-rank representation | |
CN107679569A (en) | Raman spectrum substance automatic identifying method based on adaptive hypergraph algorithm | |
CN110245635A (en) | A kind of Infrared image recognition of coal and spoil | |
CN112069955B (en) | Typhoon intensity remote sensing inversion method based on deep learning | |
CN110458201A (en) | A kind of remote sensing image object-oriented classification method and sorter | |
CN106960221A (en) | A kind of hyperspectral image classification method merged based on spectral signature and space characteristics and system | |
CN107194423A (en) | The hyperspectral image classification method of the integrated learning machine that transfinites of feature based random sampling | |
CN108960404A (en) | A kind of people counting method and equipment based on image | |
CN107560735A (en) | High-resolution infrared imaging system and method based on compressed sensing | |
CN109883967A (en) | A kind of Eriocheir sinensis quality grade method of discrimination based on information fusion | |
Zheng et al. | Using continous wavelet analysis for monitoring wheat yellow rust in different infestation stages based on unmanned aerial vehicle hyperspectral images | |
CN103868865A (en) | Optical material classification and recognition method based on hyperspectral data information maximization | |
CN106568730B (en) | A kind of rice yin-yang leaf fringe recognition methods based on Hyperspectral imaging near the ground |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
EE01 | Entry into force of recordation of patent licensing contract | ||
EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20191018 Assignee: HEFEI KEDA INDUSTRIAL EQUIPMENT Co.,Ltd. Assignor: Anhui University of Science and Technology Contract record no.: X2023980034434 Denomination of invention: A Method for Identifying Coal Gangue Based on Multispectral Spectral Information and 1D-CNN Granted publication date: 20230103 License type: Common License Record date: 20230406 |