CN111797930B - Fabric material near infrared spectrum identification and identification method based on twin network - Google Patents
Fabric material near infrared spectrum identification and identification method based on twin network Download PDFInfo
- Publication number
- CN111797930B CN111797930B CN202010647025.6A CN202010647025A CN111797930B CN 111797930 B CN111797930 B CN 111797930B CN 202010647025 A CN202010647025 A CN 202010647025A CN 111797930 B CN111797930 B CN 111797930B
- Authority
- CN
- China
- Prior art keywords
- near infrared
- infrared spectrum
- sample
- fabric material
- identification
- 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
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 105
- 239000004744 fabric Substances 0.000 title claims abstract description 80
- 239000000463 material Substances 0.000 title claims abstract description 75
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000001228 spectrum Methods 0.000 claims abstract description 23
- 230000003595 spectral effect Effects 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 16
- 239000013598 vector Substances 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000012512 characterization method Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims 1
- 239000000523 sample Substances 0.000 description 40
- 238000005406 washing Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 2
- 244000025254 Cannabis sativa Species 0.000 description 1
- 235000012766 Cannabis sativa ssp. sativa var. sativa Nutrition 0.000 description 1
- 235000012765 Cannabis sativa ssp. sativa var. spontanea Nutrition 0.000 description 1
- 229920000742 Cotton Polymers 0.000 description 1
- 229920004933 Terylene® Polymers 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 235000009120 camo Nutrition 0.000 description 1
- 235000005607 chanvre indien Nutrition 0.000 description 1
- 239000012468 concentrated sample Substances 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000011487 hemp Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000005020 polyethylene terephthalate Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- -1 silk Polymers 0.000 description 1
- 238000002791 soaking Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/259—Fusion by voting
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a twin network-based fabric material near infrared spectrum identification and identification method, which comprises the following steps: s100: collecting fabric samples made of different materials, and marking the types; s102: the acquisition of near infrared spectrum data of different types of samples is completed, and one or more near infrared spectrum data of each type of sample is ensured; preprocessing near infrared spectrum data to construct a near infrared spectrum sample database; s104: constructing a twin network, randomly extracting partial characteristics of spectrum sample data of the fabric material in a near infrared spectrum sample database as a near infrared spectrum sample subset, and training to obtain a twin network model; s106: acquiring spectral data of a fabric to be identified and identified, and randomly extracting a plurality of sample subsets; s108: and (4) inputting the plurality of sample subsets extracted randomly in the step (S106) into the twin network model trained in the step (S104), judging the allocation degree of the input sample subsets according to the output result, and judging the final prediction result of different subsets by voting.
Description
Technical Field
The invention belongs to the field of near infrared spectrum substance identification and identification, and particularly relates to a twin network-based fabric material near infrared spectrum identification and identification method.
Background
Currently, intelligent household appliances relate to various household appliance categories, and besides televisions, refrigerators, air conditioners and washing machines are increasingly integrated with intelligent attributes. But for many common consumers, the more intelligent the product, the more functions and the more tedious the operation. The intelligent washing machine also faces such a problem that the control panel becomes more and more complex. According to the traditional full-automatic washing machine, a certain program is preset on a computer control panel of the washing machine, different washing program options are provided for a user to select according to the material of clothes to be washed, when a certain program is selected during washing, the washing machine can automatically complete a series of operations such as soaking, washing, rinsing and dewatering, the washing is automatically stopped, and a buzzer sounds to inform airing. However, the general user does not have the ability to accurately distinguish the material of the clothes to be washed, may give out an inappropriate choice, and may directly select a standard mode more often, thereby reducing the washing effect. The patent publication No. CN201710047604.5, CN201910389465.3, material identification device, method, washing machine and control method thereof, and the like, automatically identify the material of the laundry, recommend the washing program, and optimize the washing mode to a certain extent. However, the clothes are made of various materials, and cotton, silk, terylene, hemp and the like have few samples and different contents, so that accurate identification and identification are difficult to achieve.
Based on the above current situation, how to provide a rapid identification and identification method for the situations of multiple fabric material types, unfixed types and few samples of each type is still a technical problem to be solved urgently.
Disclosure of Invention
The invention aims to solve the problems and provides a twin network-based fabric material near infrared spectrum identification and identification method, which can be used for constructing a rapid detection identification and identification method aiming at the special data format of near infrared spectrum, such as more fabric types, less samples and the like.
In order to achieve the purpose, the invention adopts the following technical scheme:
a near infrared spectrum identification and identification method for fabric materials based on twin networks comprises the following steps:
s100: collecting fabric samples made of different materials, and marking the types;
s102: the acquisition of near infrared spectrum data of different types of samples is completed, and one or more near infrared spectrum data of each type of sample is ensured; preprocessing near infrared spectrum data to construct a near infrared spectrum sample database;
s104: constructing a twin network, randomly extracting partial characteristics of spectrum sample data of the fabric material in a near infrared spectrum sample database as a near infrared spectrum sample subset, and training to obtain a twin network model;
s106: acquiring spectral data of a fabric to be identified and identified, and randomly extracting a plurality of sample subsets;
s108: and (4) inputting the plurality of sample subsets extracted randomly in the step (S106) into the twin network model trained in the step (S104), judging the allocation degree of the input sample subsets according to the output result, and judging the final prediction result of different subsets by voting.
Further, the step S100: collecting fabric samples of different materials, and marking categories, comprising: collecting fabric samples of different materials, collecting fabrics of various materials, classifying the materials with different content components independently, and collecting a small amount of samples in each category to form a fabric sample library; scanning samples in a sample library to obtain corresponding near infrared spectrum data Xi ═ { Xi1, Xi2.. xin }, and marking corresponding sample types Ynum, wherein i is the number of the samples, n is the dimensionality of the near infrared spectrum, and num is the number of the types of the samples.
Further, in S104, the constructing of the twin network includes a near infrared spectrum feature extraction subnet and a near infrared spectrum feature distance calculation decision subnet of the fabric material, the subnet is extracted by near infrared spectrum features, and near infrared spectrum data of a plurality of high latitudes are mapped to a low latitudes spectral feature vector; and (3) calculating the characteristic distance of the near infrared spectrum from the decision subnet, so that the characteristic spectrum distance of the same type of fabric material class is closer, otherwise, the distance of different types is farther, and judging the matching degree of the near infrared spectrum by the distance calculation decision subnet.
Further, the sub-network for extracting the near infrared spectrum characteristics of the fabric material comprises the following steps: the near infrared spectrum data is a one-dimensional sequence, two completely consistent one-dimensional convolution neural networks are adopted in the feature extraction subnet, the feature extraction subnet data is n-dimensional data of the near infrared spectrum, and m-dimensional spectrum feature vectors are output.
Further, the calculating of the near infrared spectrum characteristic distance of the fabric material decision subnet comprises: and the input of the near infrared spectrum characteristic distance calculation decision subnet is the output of m-dimensional spectrum characteristic vectors of the two one-dimensional convolutional neural networks, and the similarity of the two m-dimensional spectrum characteristic vectors is calculated and judged through the distance calculation decision subnet.
Compared with the prior art, the invention has the beneficial effects that:
the method aims at the condition that the near infrared spectrum identification and identification of the fabric material have more and unfixed targets and few samples of each type. Meanwhile, a subnet is extracted according to the characteristics of the specially designed twin network according to the condition of the one-dimensional sequence of the spectral data. The method is characterized in that a one-dimensional convolutional neural network with strong feature extraction capability is adopted to extract the near infrared spectrum data features of the fabric material, the twin network is used for comparing the similarity degree of random feature subsets of the near infrared spectrum data of the two fabric materials, and the output result is judged by voting. The method has the advantages of meeting the requirement of rapid identification of small samples and multi-class near infrared spectrum data. Meanwhile, the characteristic subset is adopted for comprehensive voting judgment, so that the model has higher generalization and is obviously creative in principle and implementation scheme.
Drawings
FIG. 1 is a twin network structure diagram of a fabric material near infrared spectrum identification and identification method in the fabric material near infrared spectrum identification and identification method based on the twin network of the present invention;
FIG. 2 is a schematic flow chart of a twin network-based fabric material near infrared spectrum identification and identification method of the invention;
FIG. 3 is a feature extraction subnet of a twin network based fabric material near infrared spectrum identification and authentication method of the present invention;
FIG. 4 is a complete network of a twin network based fabric material near infrared spectrum identification and characterization method of the present invention;
FIG. 5 shows the result of the near infrared spectrum identification and identification method of fabric material based on twin network training prediction.
Fig. 6 is two spectra of the same fabric at different locations.
Detailed Description
The present invention will be further described with reference to the following examples, which are intended to illustrate only some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, other embodiments used by those skilled in the art without any creative effort belong to the protection scope of the present invention.
Example 1:
as shown in figure 1, the twin network for identifying and identifying the near infrared spectrum of the fabric material mainly comprises two parts, namely a near infrared spectrum characteristic extraction subnet of the fabric material and a near infrared spectrum characteristic distance calculation decision subnet. The sub-network for extracting the near infrared spectrum characteristics of the fabric material consists of two completely consistent networks. According to the characteristics of the input near infrared spectrum data, the two networks are designed into one-dimensional convolution neural networks, and the function is to extract the characteristics of the near infrared spectrum data.
And the input of the near infrared spectrum characteristic distance calculation decision subnet is the output of two near infrared spectrum characteristic extraction subnets, and the similarity of two characteristic vectors is calculated and judged through the distance calculation decision subnet.
As shown in FIG. 2, the method for identifying and identifying the near infrared spectrum of the fabric material based on the twin network comprises the following steps:
step one, collecting fabric samples made of different materials, collecting fabrics made of various materials, classifying the materials with different content components independently, and collecting a small amount of samples in each category to form a fabric sample library.
And step two, scanning the samples in the sample library to obtain corresponding near infrared spectrum data Xi ═ Xi1, Xi2. Wherein i is the number of samples, n is the dimensionality of the near infrared spectrum, and num is the number of classes of the samples.
And step three, preprocessing the near infrared spectrum data, and constructing a fabric material training near infrared sample library.
Step four, building a twin network, and training a random feature subset pair of a near-infrared sample library by using the spectral data of the fabric material to obtain a fabric material near-infrared spectrum recognition twin network model;
and fifthly, scanning the near infrared spectrum data of the fabric sample to be detected for the classification problem, sequentially grouping the near infrared spectrum data with the near infrared spectrum data in the training sample library, randomly extracting a feature subset pair, inputting the feature subset pair into the twin network model, voting and judging the class with the highest similarity as the final class to be output, and completing the prediction of the class of the fabric material. For the identification problem, the near infrared spectrum data of the fabric sample to be identified is paired with the near infrared spectrum data of the class to be identified in the training sample library, the feature subset pairs are randomly extracted and input into the twin network model, and the similar fabric sample is judged to be identified by voting.
The method for constructing and training the spectrum data of the fabric material to obtain the twin network model specifically comprises the following steps:
the constructed twin network consists of two important parts, namely a near infrared spectrum characteristic extraction subnet and a near infrared spectrum characteristic distance calculation decision subnet of the fabric material. Through the feature extraction subnet, hundreds of thousands of dimensions of high latitude near infrared spectrum data are mapped into low latitude spectral feature vectors. And (3) calculating the characteristic distance of the near infrared spectrum from the decision subnet, so that the characteristic spectrum distance of the same type of fabric material class is closer, otherwise, the distance of different types is farther, and in short, the matching degree of the near infrared spectrum is judged by the distance calculation decision subnet.
The sub-network for extracting the near infrared spectrum characteristics of the fabric material is as follows:
because the near infrared spectrum data is a one-dimensional sequence, two completely consistent one-dimensional convolutional neural networks (1D CNN) are adopted by the feature extraction subnet, the feature extraction subnet data is n-dimensional data of the near infrared spectrum, and m is output as a spectrum feature vector.
The sub-network for calculating and deciding the near infrared spectrum characteristic distance is as follows:
the sub-network for calculating the near infrared spectrum characteristic distance of the fabric material comprises the following steps: and the input of the near infrared spectrum characteristic distance calculation decision subnet is the output of m-dimensional spectrum characteristic vectors of the two one-dimensional convolutional neural networks, and the similarity of the two m-dimensional spectrum characteristic vectors is calculated and judged through the distance calculation decision subnet.
The fabric material twin network training data is paired fabric material near infrared spectrum data. For the fabric material spectrum training sample data set of N samples, N (N-1) pairs of fabric material twin network inputs are input. Meanwhile, a feature subset pair which is randomly extracted into paired fabric material near infrared spectrum data is designed to serve as final training data, a small number of samples can provide sufficient training sets, when each pair of two groups of spectrum data come from the same class, the two groups of spectrum data are marked as the same class, otherwise, the two groups of spectrum data are different and serve as the output of a twin network.
After the twin network training is completed, the method can be used for identifying and identifying the fabric material sample to be tested. And for the identification problem, sequentially combining the fabric sample to be identified with the training sample library to form a sample set to be identified, sequentially inputting the sample set into a twin network model, wherein each pair of the sample set comprises a plurality of groups of random feature subsets, and the corresponding training sample set with the minimum voting output result and the highest matching degree is taken as an identification category. For the identification problem, only the sample to be identified and the class concentrated sample to be identified corresponding to the training sample set form a sample set to be tested, each pair has a plurality of groups of random feature subsets, and the size of the voting output result is the matching degree of the sample to be tested and the sample library.
As shown in fig. 3, a sub-network for extracting near infrared spectrum features of an embodiment is formed by combining a convolutional layer, a pooling layer, a full connection layer, and the like by using a one-dimensional convolutional neural network. The network structure can extract near infrared spectrum characteristics of different fabric materials.
As shown in fig. 4, a complete network is a specific embodiment of a twin network-based fabric material near infrared spectrum identification and identification method. And the output of the two near infrared spectrum characteristic extraction subnetworks is used as the input of a near infrared spectrum characteristic distance calculation decision subnet, and the near infrared spectrum characteristic distance calculation decision subnet adopts the Euclidean distance. The final output is the matching degree of the two input samples, and the smaller the distance, the higher the matching degree.
As shown in fig. 5, for the prediction result of the training in this embodiment, for the multi-class problem, the to-be-verified samples may be sequentially combined with the samples in the training sample set to form an input to-be-verified sample pair, and the corresponding training sample set with the minimum output is the class attribution of the to-be-verified samples. For the problem of fabric material identification, only the sample to be identified and the corresponding training sample set are required to form an input sample pair to be verified, and the matching degree of the fabric material is judged according to the output. From the result, the twin network model training convergence speed is high, the prediction can reach high accuracy, and the accuracy is obviously improved on the near infrared spectrum identification of the fabric material compared with the traditional machine learning method on the same sample set.
For example, fig. 6 shows two spectra of the same fabric at different positions, it can be seen that most positions of the two spectra are relatively similar, so that a higher accuracy can be obtained by using a pair of randomly extracted sub-features as training and prediction, and finally voting to determine a final result.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (5)
1. A fabric material near infrared spectrum identification and identification method based on a twin network is characterized by comprising the following steps:
s100: collecting fabric samples made of different materials, and marking the types;
s102: the acquisition of near infrared spectrum data of different types of samples is completed, and one or more near infrared spectrum data of each type of sample is ensured; preprocessing near infrared spectrum data to construct a near infrared spectrum sample database;
s104: constructing a twin network, randomly extracting partial characteristics of spectrum sample data of the fabric material in a near infrared spectrum sample database as a near infrared spectrum sample subset, and training to obtain a twin network model;
s106: acquiring spectral data of a fabric to be identified and identified, and randomly extracting a plurality of sample subsets;
s108: and (4) inputting the plurality of sample subsets extracted randomly in the step (S106) into the twin network model trained in the step (S104), judging the allocation degree of the input sample subsets according to the output result, and judging the final prediction result of different subsets by voting.
2. The method for near infrared spectrum identification and identification of fabric materials based on twin networks as claimed in claim 1, wherein the step S100: collecting fabric samples of different materials, and marking categories, comprising: collecting fabric samples of different materials, collecting fabrics of various materials, classifying the materials with different content components independently, and collecting a small amount of samples in each category to form a fabric sample library; scanning samples in a sample library to obtain corresponding near infrared spectrum data Xi ═ { Xi1, Xi2.. xin }, and marking corresponding sample types Ynum, wherein i is the number of the samples, n is the dimensionality of the near infrared spectrum, and num is the number of the types of the samples.
3. The twin network-based fabric material near infrared spectrum identification and identification method as claimed in claim 1, wherein in S104, the construction of the twin network comprises a near infrared spectrum feature extraction subnet and a near infrared spectrum feature distance calculation decision subnet of the fabric material, the subnet is extracted by near infrared spectrum features, and near infrared spectrum data of a plurality of high latitudes are mapped to a low latitudes spectral feature vector; and (3) calculating the characteristic distance of the near infrared spectrum from the decision subnet, so that the characteristic spectrum distance of the same type of fabric material class is closer, otherwise, the distance of different types is farther, and judging the matching degree of the near infrared spectrum by the distance calculation decision subnet.
4. The twin network based fabric material near infrared spectrum identification and characterization method according to claim 3, wherein the fabric material near infrared spectrum feature extraction sub-network comprises: the near infrared spectrum data is a one-dimensional sequence, two completely consistent one-dimensional convolution neural networks are adopted in the feature extraction subnet, the feature extraction subnet data is n-dimensional data of the near infrared spectrum, and m-dimensional spectrum feature vectors are output.
5. The twin network based fabric material near infrared spectrum identification and characterization method according to claim 3, wherein the near infrared spectrum characteristic distance calculation decision subnet of the fabric material comprises: and the input of the near infrared spectrum characteristic distance calculation decision subnet is the output of m-dimensional spectrum characteristic vectors of the two one-dimensional convolutional neural networks, and the similarity of the two m-dimensional spectrum characteristic vectors is calculated and judged through the distance calculation decision subnet.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010647025.6A CN111797930B (en) | 2020-07-07 | 2020-07-07 | Fabric material near infrared spectrum identification and identification method based on twin network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010647025.6A CN111797930B (en) | 2020-07-07 | 2020-07-07 | Fabric material near infrared spectrum identification and identification method based on twin network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111797930A CN111797930A (en) | 2020-10-20 |
CN111797930B true CN111797930B (en) | 2021-12-17 |
Family
ID=72810134
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010647025.6A Active CN111797930B (en) | 2020-07-07 | 2020-07-07 | Fabric material near infrared spectrum identification and identification method based on twin network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111797930B (en) |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112831976A (en) * | 2021-01-07 | 2021-05-25 | 格力电器(武汉)有限公司 | Clothes material identification and washing scheme determination method, washing machine and server |
CN112801986A (en) * | 2021-02-01 | 2021-05-14 | 深圳市创景数据科技有限责任公司 | Textile material matching method, device, equipment, storage medium and system |
CN113255838A (en) * | 2021-06-29 | 2021-08-13 | 成都数之联科技有限公司 | Image classification model training method, system and device, medium and classification method |
CN114119984A (en) * | 2021-09-06 | 2022-03-01 | 池明旻 | Fiber component classification method based on one-dimensional convolutional neural network |
CN114062305B (en) * | 2021-10-15 | 2024-01-26 | 中国科学院合肥物质科学研究院 | Single grain variety identification method and system based on near infrared spectrum and 1D-In-Resnet network |
CN114062307A (en) * | 2021-10-25 | 2022-02-18 | 池明旻 | Data acquisition specification for near-infrared fabric fiber component nondestructive cleaning analysis |
CN117171566B (en) * | 2022-08-17 | 2024-05-10 | 无锡迅杰光远科技有限公司 | Sample content identification method, device and storage medium |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102006007509A1 (en) * | 2006-02-16 | 2007-08-30 | Contitech Antriebssysteme Gmbh | V-ribbed belt with improved noise behavior |
CN105224961A (en) * | 2015-11-04 | 2016-01-06 | 中国电子科技集团公司第四十一研究所 | A kind of diffuse reflectance infrared spectroscopy of high resolution extracts and matching process |
CN205609936U (en) * | 2016-04-29 | 2016-09-28 | 广州市槿泓电子有限公司 | Net twine drawing -in machine |
CN106676822A (en) * | 2017-01-22 | 2017-05-17 | 无锡小天鹅股份有限公司 | Washing machine, washing control method and device of washing machine based on clothes natures |
CN107478598A (en) * | 2017-09-01 | 2017-12-15 | 广东省智能制造研究所 | A kind of near-infrared spectral analytical method based on one-dimensional convolutional neural networks |
CN207672165U (en) * | 2017-12-27 | 2018-07-31 | 苏州扬昇纺织科技有限公司 | A kind of drawing and setting structure of FDY filament |
CN108490204A (en) * | 2011-09-25 | 2018-09-04 | 赛拉诺斯知识产权有限责任公司 | System and method for multiple analysis |
CN109726756A (en) * | 2018-12-25 | 2019-05-07 | 北京旷视科技有限公司 | Image processing method, device, electronic equipment and storage medium |
CN110222773A (en) * | 2019-06-10 | 2019-09-10 | 西北工业大学 | Based on the asymmetric high spectrum image small sample classification method for decomposing convolutional network |
CN110438746A (en) * | 2019-05-10 | 2019-11-12 | 青岛海尔智能技术研发有限公司 | Material Identification device, method, washing machine and its control method |
CN110717368A (en) * | 2018-07-13 | 2020-01-21 | 北京服装学院 | Qualitative classification method for textiles |
CN111126563A (en) * | 2019-11-25 | 2020-05-08 | 中国科学院计算技术研究所 | Twin network-based space-time data target identification method and system |
CN111247535A (en) * | 2017-10-13 | 2020-06-05 | Sualab有限公司 | Depth-based image comparison apparatus, method, and computer program stored on computer-readable medium |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9507980B2 (en) * | 2014-04-02 | 2016-11-29 | Trimble Navigation Limited | Intelligent container |
-
2020
- 2020-07-07 CN CN202010647025.6A patent/CN111797930B/en active Active
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102006007509A1 (en) * | 2006-02-16 | 2007-08-30 | Contitech Antriebssysteme Gmbh | V-ribbed belt with improved noise behavior |
CN108490204A (en) * | 2011-09-25 | 2018-09-04 | 赛拉诺斯知识产权有限责任公司 | System and method for multiple analysis |
CN105224961A (en) * | 2015-11-04 | 2016-01-06 | 中国电子科技集团公司第四十一研究所 | A kind of diffuse reflectance infrared spectroscopy of high resolution extracts and matching process |
CN205609936U (en) * | 2016-04-29 | 2016-09-28 | 广州市槿泓电子有限公司 | Net twine drawing -in machine |
CN106676822A (en) * | 2017-01-22 | 2017-05-17 | 无锡小天鹅股份有限公司 | Washing machine, washing control method and device of washing machine based on clothes natures |
CN107478598A (en) * | 2017-09-01 | 2017-12-15 | 广东省智能制造研究所 | A kind of near-infrared spectral analytical method based on one-dimensional convolutional neural networks |
CN111247535A (en) * | 2017-10-13 | 2020-06-05 | Sualab有限公司 | Depth-based image comparison apparatus, method, and computer program stored on computer-readable medium |
CN207672165U (en) * | 2017-12-27 | 2018-07-31 | 苏州扬昇纺织科技有限公司 | A kind of drawing and setting structure of FDY filament |
CN110717368A (en) * | 2018-07-13 | 2020-01-21 | 北京服装学院 | Qualitative classification method for textiles |
CN109726756A (en) * | 2018-12-25 | 2019-05-07 | 北京旷视科技有限公司 | Image processing method, device, electronic equipment and storage medium |
CN110438746A (en) * | 2019-05-10 | 2019-11-12 | 青岛海尔智能技术研发有限公司 | Material Identification device, method, washing machine and its control method |
CN110222773A (en) * | 2019-06-10 | 2019-09-10 | 西北工业大学 | Based on the asymmetric high spectrum image small sample classification method for decomposing convolutional network |
CN111126563A (en) * | 2019-11-25 | 2020-05-08 | 中国科学院计算技术研究所 | Twin network-based space-time data target identification method and system |
Non-Patent Citations (7)
Title |
---|
"Deep Heterogeneous Network Embedding based on Siamese Neural Networks";Chen Zhang 等;《Neurocomputing》;20200507;第388卷;1-11 * |
"Siamese neural networks for one-shot image recognition";Koch, Gregory 等;《ICML deep learning workshop》;20150710;第2卷;1-30 * |
"基于深度学习的织物图像疵点区域定位算法研究";刘祥惠;《中国优秀硕士学位论文全文数据库 (工程科技Ⅰ辑)》;20180715(第2018/07期);B024-21 * |
"基于深度学习的非平衡小样本异物检测识别技术研究";余亚萍;《中国优秀硕士学位论文全文数据库 (社会科学Ⅰ辑)》;20200215(第2020/02期);G108-386 * |
"绿泥石矿物近红外光谱吸收谱带的位移机理与控制机制研究";杨敏;《中国博士学位论文全文数据库 (基础科学辑)》;20200115(第2020/01期);A011-8 * |
"颜料和丝织品类文物材料光致变色的分子光谱评价研究";武敬青;《中国优秀硕士学位论文全文数据库 (哲学与人文科学辑)》;20120615(第2012/06期);F097-2 * |
B. Bäuml 等."Deep n-Shot Transfer Learning for Tactile Material Classification with a Flexible Pressure-Sensitive Skin".《IEEE》.2019,4262-4268. * |
Also Published As
Publication number | Publication date |
---|---|
CN111797930A (en) | 2020-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111797930B (en) | Fabric material near infrared spectrum identification and identification method based on twin network | |
CN108914489B (en) | Control method and device of washing machine, washing machine and storage medium | |
Wu et al. | Robust latent factor analysis for precise representation of high-dimensional and sparse data | |
Inoue et al. | Multi-label fashion image classification with minimal human supervision | |
Dong et al. | Non-intrusive signature extraction for major residential loads | |
CN103714077B (en) | Method, the method and device of retrieval verification of object retrieval | |
CN103902591B (en) | Construct the method and device of decision tree classifier | |
CN107682109B (en) | A kind of interference signal classifying identification method suitable for UAV Communication system | |
US11048976B2 (en) | Method and system for controlling machines based on object recognition | |
CN110210567A (en) | A kind of image of clothing classification and search method and system based on convolutional neural networks | |
CN105653171A (en) | Fingerprint identification based terminal control method, terminal control apparatus and terminal | |
Bilski et al. | Generalized algorithm for the non-intrusive identification of electrical appliances in the household | |
CN111639586A (en) | Non-invasive load identification model construction method, load identification method and system | |
CN108628721A (en) | Method for detecting abnormality, device, storage medium and the electronic device of user data value | |
CN115151952A (en) | High-precision identification method and system for power transformation equipment | |
CN115346082A (en) | Image generation method, image generation device, storage medium and electronic device | |
CN106407281B (en) | Image retrieval method and device | |
Yuan et al. | A novel current signal feature and its application in noninvasive load monitoring | |
CN111191659B (en) | Multi-shape clothes hanger identification method for clothing production system | |
Ho et al. | Evaluation of attribute selection measures in decision tree induction | |
Gay et al. | Should we reload time series classification performance evaluation?(a position paper) | |
Bianchi et al. | An interpretable graph-based image classifier | |
CN110004640B (en) | Clothes treatment method of washing machine | |
CN109840336B (en) | Clothing design sample recommendation method and device | |
Pau et al. | Electric current classification with tiny machine learning for home appliances |
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 |