CN110348538B - Multispectral spectral information and 1D-CNN coal and gangue identification method - Google Patents
Multispectral spectral information and 1D-CNN coal and gangue identification method Download PDFInfo
- Publication number
- CN110348538B CN110348538B CN201910652387.1A CN201910652387A CN110348538B CN 110348538 B CN110348538 B CN 110348538B CN 201910652387 A CN201910652387 A CN 201910652387A CN 110348538 B CN110348538 B CN 110348538B
- Authority
- CN
- China
- Prior art keywords
- gangue
- coal
- multispectral
- cnn
- spectral information
- 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.)
- Active
Links
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method for identifying coal and gangue by multispectral spectral information and 1D-CNN, which comprises the following steps: acquiring multispectral spectral information of coal and gangue; (2) dividing samples of coal and gangue spectral information; (3) extracting spectral features of the one-dimensional convolution neural network; and (4) constructing a probability neural network coal and gangue identification model. The invention adopts 1D CNN-PNN to construct the identification model of the multispectral spectral information of the coal and the gangue, provides a new one-dimensional convolutional neural network model which can extract more and more effective characteristic information, can effectively avoid the problems of overfitting and the like, and is very suitable for real-time and accurate identification of the coal and the gangue.
Description
Technical Field
The invention relates to the technical field of coal and gangue identification, in particular to a coal and gangue identification method based on multispectral spectral information and 1D-CNN.
Background
The essential characteristics of energy resources of rich coal, poor oil and less gas determine the important position of coal in primary energy. In the coal mining process, a large amount of gangue and coal are mined together. When the gangue is mixed with the coal, the heating value of the coal is influenced, the quality of the coal is influenced, and meanwhile, the environment is seriously polluted in the combustion process. Therefore, the separation of the gangue from the coal is a crucial treatment link before the coal is used. In the coal separation technology at home and abroad at present, besides manual gangue separation, the automatic gangue (coal) separation technology can be divided into wet gangue separation and dry gangue separation according to whether water resources are utilized or not. The wet-method gangue separation needs to consume a large amount of water resources, and the generated coal slime pollution is difficult to treat; certain radiation exists in ray gangue separation such as gamma rays and X-ray gangue separation, and the interference of factors such as light rays on common image identification gangue separation is large.
Multispectral Imaging (MSI) was first applied to the military field, and then gradually applied to various aspects of agriculture as the technology is continuously developed. A plurality of images in different spectral regions are obtained after multispectral imaging, and the multispectral imaging technology can solve the problems that the RGB images are narrow in wave band range and easy to be interfered by environments such as illumination. Meanwhile, compared with the traditional CCD imaging, the multispectral imaging technology can acquire the spectral information of different wave bands besides the image information of different spectral regions. At present, analysis aiming at multispectral spectral information mainly adopts analysis steps of feature extraction and pattern recognition, on one hand, the processing steps are various, and on the other hand, analysis results are greatly influenced by selection of the feature extraction and pattern recognition methods.
Disclosure of Invention
The invention aims to provide a coal and gangue identification method based on multispectral spectral information and 1D-CNN, which aims to overcome the defects of the existing coal and gangue identification method and accurately identify coal and gangue in real time.
In order to achieve the purpose, the invention adopts the technical scheme that: a method for identifying coal and gangue by multispectral spectral information and 1D-CNN comprises the following steps:
(1) Acquiring multispectral spectral information of coal and gangue;
(2) Dividing samples of spectral information of coal and gangue;
(3) Extracting spectral features of the one-dimensional convolutional neural network;
(4) And (5) constructing a probability neural network coal and gangue identification model.
Preferably, in the step (1), the multispectral imaging technology is used to obtain the spectral information of the multispectral data of the coal and the gangue, so as to obtain the multispectral spectral data set of the coal and the gangue.
Preferably, in the step (2), the spectral data of the preprocessed coal and gangue multispectral are divided into an independent training set and an independent testing set according to a certain proportion by adopting a random sampling method.
Preferably, in the step (3), the One-dimensional convolutional neural network (1D-CNN) for extracting spectral information features is a network structure including two One-dimensional convolutional units (1D Conv blocks), and mainly includes a normalization layer, a convolutional layer, a pooling layer, and a full connection layer.
Preferably, in the step (4), a Probabilistic Neural Network (PNN) coal and gangue identification model is constructed on a training set by using the spectral features extracted by the CNN, parameters of the PNN coal and gangue identification model are determined, and then the test set is used to test the identification effect and verify the model performance.
Through the technical scheme, the invention has the beneficial effects that: acquiring multispectral spectral information of coal and gangue by adopting a multispectral imaging technology and identifying the coal and gangue so as to solve the defects of the existing coal and gangue identification method; the new one-dimensional convolutional neural network model can extract more and more effective characteristic information, can effectively avoid the problems of overfitting and the like, and is very suitable for real-time and accurate identification of coal and gangue.
Drawings
FIG. 1 is a flow chart of a coal and gangue identification method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a one-dimensional convolutional neural network for extracting features according to an embodiment of the present invention;
FIG. 3 is a Block diagram of Conv Block Unit 1 according to embodiment 1D of the present invention;
fig. 4 is a structural diagram of the Conv Block unit 2 in embodiment 1D of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to specific embodiments and the accompanying drawings.
The invention is carried out in Windows 10 environment, keras (V2.2.4) is adopted for analysis, tensorFlow (V1.10.0) is used as the rear end of the Keras, and Intel core I7-9700K and Intel Geforce RTX 2070 are used as hardware.
The embodiment of the invention provides a method for identifying multispectral spectral information and 1D-CNN coal gangue to solve all or part of the defects of the prior art, which comprises the following steps:
(1) Acquiring multispectral spectral information of coal and gangue;
(2) Dividing samples of coal and gangue spectral information;
(3) Extracting spectral features of the one-dimensional convolutional neural network;
(4) And (5) constructing a probability neural network coal and gangue identification model.
In order to make the purpose, technical scheme and advantages of the embodiment of the present invention more clearly understood, a multispectral spectral information and a coal and gangue identification method of 1D-CNN provided by the embodiment of the present invention are described in detail with reference to the accompanying drawings, and the multispectral spectral information and coal and gangue identification method of 1D-CNN provided by the embodiment of the present invention includes the steps as shown in fig. 1:
101: the multispectral imaging module is used for collecting multispectral spectral data of a plurality of samples of coal and gangue by using a real-time multispectral Mosaic surface camera of Shanghai five-bell photoelectric technology company Limited to obtain multispectral spectral data of the coal and gangue, wherein the data volume is 25, and the wavelength range is 600-875nm.
102: and (3) dividing samples of spectral information of the coal and the gangue, and dividing multispectral spectral data of the preprocessed coal and gangue into an independent training set and an independent testing set according to the proportion of 80% of the training set to 20% of the testing set by adopting a random sampling method.
103: the spectral feature extraction of the one-dimensional convolutional neural network, the one-dimensional convolutional neural network (1D-CNN) for extracting the spectral feature of the multispectral is a network structure including two one-dimensional convolutional units (1D Conv blocks), and mainly includes a normalization layer, a convolutional layer, a pooling layer, a full-link layer, and the like, and a schematic structural diagram of the network structure is shown in fig. 2, and is specifically described as follows:
201: the standard layer uses batch standardization layer BatchNormalization in Keras, and the output size is consistent with the input size;
202: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 16;
203: the pooling layer uses 1D maximum pooling Max Pooling1D in Keras, and the pooling size is 2;
204: the 1D Conv Block unit 1 is configured using a convolutional layer and a pooling layer in Keras, and its structure is shown in fig. 3, and is specifically described as follows:
301: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 16;
302: maxPooling1D was maximally pooled in the pooling layer using 1D in Keras, with pooling size of 2;
205: the 1D Conv Block unit 2 is configured using a convolutional layer and a pooling layer in Keras, and its structure is shown in fig. 4, and is specifically described as follows:
401: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 32;
402: the pooling layer uses 1D maximum pooling Max Pooling1D in Keras, and the pooling size is 2;
in particular, in the 1D Conv Block unit 1 and the 1D Conv Block unit 2, the padding manner of all the convolution layers is set to same.
206: the full connection layer uses Dense in Keras to carry out flattening and dot multiplication on input;
207: and the multispectral spectral characteristics extracted by the output CNN are used as the input of the PNN and used for constructing a model and testing the model.
104: the method comprises the steps of constructing a probability neural network coal and gangue identification model, constructing a PNN coal and gangue identification model on a training set by utilizing multispectral spectral features extracted by CNN, determining parameters of the PNN coal and gangue identification model, testing an identification effect by utilizing a test set, and verifying the performance of the model.
Through the technical scheme, the invention has the beneficial effects that: acquiring multispectral spectral information of coal and gangue by adopting a multispectral imaging technology and identifying the coal and gangue so as to solve the defects of the existing coal and gangue identification method; the new one-dimensional convolutional neural network model can extract more and more effective characteristic information, can effectively avoid the problems of overfitting and the like, and is very suitable for real-time and accurate identification of coal and gangue.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (5)
1. A method for identifying multispectral spectral information and coal and gangue of 1D-CNN is characterized by comprising the following steps: the method comprises the following steps:
(1) Acquiring multispectral spectral information of coal and gangue;
(2) Dividing samples of coal and gangue spectral information;
(3) Extracting spectral features of a one-dimensional convolutional neural network;
(4) Constructing a probability neural network coal and gangue identification model;
in the step (3), in the one-dimensional convolutional neural network spectral feature extraction, the one-dimensional convolutional neural network (1D-CNN) for extracting multispectral spectral features is a network structure including two one-dimensional convolutional units (1D Conv blocks), and mainly includes a normalization layer, a convolutional layer, a pooling layer, a full-link layer, and the like, and is specifically described as follows:
201: the standard layer uses batch standardization layer BatchNormalization in Keras, and the output size is consistent with the input size;
202: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 16;
203: the pooling layer uses 1D maximum pooling Max Pooling1D in Keras, and the pooling size is 2;
204: the 1D Conv Block unit 1 is constructed using the convolutional and pooling layers in Keras, as follows:
301: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 16;
302: the pooling layer uses 1D maximum pooling Max Pooling1D in Keras, and the pooling size is 2;
205: the 1D Conv Block unit 2 was constructed using the convolutional and pooling layers of Keras, as described in detail below:
401: the convolutional layer uses 1D convolutional layer Conv1D in Keras, the size of a convolution kernel is 3, and the number of the convolution kernels is 32;
402: the pooling layer uses 1D maximum pooling Max Pooling1D in Keras, and the pooling size is 2;
specifically, in the 1D Conv Block unit 1 and the 1D Conv Block unit 2, the padding manner of all the convolution layers is set to same;
206: the full-connection layer uses Dense in Keras to carry out flattening and dot multiplication on input;
207: and the multispectral spectral characteristics extracted by the output CNN are used as the input of the PNN and used for constructing a model and testing the model.
2. The method of claim 1, wherein the multispectral spectral information and 1D-CNN gangue identification method comprises: in the step (1), the multispectral imaging technology is used to obtain the spectral information of the multispectral data of the coal and the gangue, and a multispectral spectral data set of the coal and the gangue is obtained.
3. The method for identifying the gangue of the 1D-CNN based on the multispectral spectral information as claimed in claim 1, wherein the method comprises the following steps: in the step (2), the spectral data of the preprocessed coal and gangue multispectral are divided into an independent training set and an independent testing set according to a certain proportion by adopting a random sampling method.
4. The method of claim 1, wherein the multispectral spectral information and 1D-CNN gangue identification method comprises: in the step (3), the One-dimensional convolutional neural network (1D-CNN) for extracting spectral information features is a network structure including two One-dimensional convolutional units (1D Conv Block), and mainly includes a normalization layer, a convolutional layer, a pooling layer, and a full connection layer.
5. The method of claim 1, wherein the multispectral spectral information and 1D-CNN gangue identification method comprises: in the step (4), a Probabilistic Neural Network (PNN) coal and gangue identification model is constructed on a training set by using spectral features extracted by CNN, parameters of the PNN coal and gangue identification model are determined, and then a test set is used to test an identification effect and verify the performance of the model.
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 CN110348538A (en) | 2019-10-18 |
| CN110348538B true 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) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111062403B (en) * | 2019-12-26 | 2022-11-22 | 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 |
| CN113138178B (en) * | 2021-04-15 | 2023-07-07 | 上海海关工业品与原材料检测技术中心 | Method for identifying imported iron ore brands |
Citations (1)
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108932480B (en) * | 2018-06-08 | 2022-03-15 | 电子科技大学 | Distributed optical fiber sensing signal feature learning and classifying method based on 1D-CNN |
-
2019
- 2019-07-18 CN CN201910652387.1A patent/CN110348538B/en active Active
Patent Citations (1)
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| CNN在煤矿突水水源LIF光谱图像识别的应用;周孟然等;《光谱学与光谱分析》;20180715(第07期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110348538A (en) | 2019-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110348538B (en) | Multispectral spectral information and 1D-CNN coal and gangue identification method | |
| CN110378295B (en) | Coal and gangue identification method based on heterogeneous fusion of multispectral image information and spectral information | |
| CN110245635B (en) | Infrared image recognition method for coal and gangue | |
| CN110321959B (en) | Multispectral image information and CNN coal and gangue identification method | |
| US20220309772A1 (en) | Human activity recognition fusion method and system for ecological conservation redline | |
| CN109446982B (en) | AR glasses-based electric power cabinet pressing plate state identification method and system | |
| CN103226706B (en) | A kind of based on the water, electricity and gas meter reading automatic identifying method rolling coupling | |
| CN109410184B (en) | Live broadcast pornographic image detection method based on dense confrontation network semi-supervised learning | |
| CN111077093A (en) | Method and device for quickly detecting coal gangue based on multispectral technology | |
| CN111105389B (en) | Road surface crack detection method integrating Gabor filter and convolutional neural network | |
| CN110210608A (en) | The enhancement method of low-illumination image merged based on attention mechanism and multi-level features | |
| CN112949460B (en) | Human behavior network model based on video and identification method | |
| CN116363440B (en) | Deep learning-based identification and detection method and system for colored microplastic in soil | |
| CN116167964A (en) | Tumor classification method and system based on tumor hyperspectral image | |
| CN113657472B (en) | Multi-source remote sensing data fusion method based on subspace learning | |
| CN112329791B (en) | Automatic extraction method for hyperspectral image water area | |
| Guo et al. | A deep learning method for the recognition of solar radio burst spectrum | |
| CN103679726A (en) | Method for improving imaging quality of rock debris image | |
| CN113378672A (en) | Multi-target detection method for defects of power transmission line based on improved YOLOv3 | |
| CN111931618A (en) | Hyperspectral classification method based on separable residual three-dimensional dense convolution | |
| CN113780406A (en) | YOLO-based bundled log end face detection method | |
| CN114898137A (en) | Face recognition-oriented black box sample attack resisting method, device, equipment and medium | |
| CN106097293A (en) | A kind of image severability evaluation method for image capturing system optimization design | |
| CN103824288A (en) | Array image registration template for lens array | |
| Mo et al. | Deep-Learning Image Steganalysis Based on Generalized Gaussian Distribution Features Clustering |
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 |
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 |
|
| EE01 | Entry into force of recordation of patent licensing contract |