CN109444141A - A kind of method and device of grain seed detection and counting based on deep learning - Google Patents
A kind of method and device of grain seed detection and counting based on deep learning Download PDFInfo
- Publication number
- CN109444141A CN109444141A CN201811583256.4A CN201811583256A CN109444141A CN 109444141 A CN109444141 A CN 109444141A CN 201811583256 A CN201811583256 A CN 201811583256A CN 109444141 A CN109444141 A CN 109444141A
- Authority
- CN
- China
- Prior art keywords
- grain seed
- transport mechanism
- detection
- grain
- high definition
- 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
- 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
- G01N21/84—Systems specially adapted for particular applications
- G01N21/85—Investigating moving fluids or granular solids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
- G06M7/00—Counting of objects carried by a conveyor
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of method and devices of grain seed detection and counting based on deep learning in grain seed counting technology field, including cabinet, high definition camera, bright field illumination lamp, grain seed support board, transport mechanism, dark field illuminator, grain seed feeding inlet, human-computer interaction touch display screen and electromechanical controlling device, grain seed feeding inlet is located above cabinet, the surface of transport mechanism is fixed with grain seed support board, electromechanical controlling device is electrically connected with transport mechanism, between transport mechanism is fixed on before and after cabinet wall, high definition camera is located in the box body top of chamber, high definition camera is connect with human-computer interaction touch display screen signal, pass through depth convolution learning network modelling and parametric solution, it can be very good to detect intensive and adhesion grain seed target, by test, the recall rate of grain seed detection and correct Rate is very high, and calculating process is simple, easy to operate.
Description
Technical field
The present invention relates to grain seed counting technology field, it is specially a kind of based on deep learning grain seed detection with
The method and device of counting.
Background technique
The detection of grain seed counts the image processing techniques that most methods at present are namely based on machine vision.Machine
The most technologies used in vision are image segmentations, and the method for image segmentation has Otsu threshold method, edge method, watershed method
And range conversion method.These methods can all have some defects or problem, they can not overcome, the inspection of grain seed target
It surveys unlike other target detections, they are smaller with some distinctive attributes and state, such as kernels particles, in a picture
Middle seed number is more and more intensive, and will appear the states such as more adhesions.
Currently, there are also researchs to mention using some special image collecting devices, as patent CN105430350A is mentioned
A kind of grain seed image capturing system arrived is that each grain seed is sucked aperture by inhaling grain device, then cloth
Onto relevant loading surface.There are some problems for this method: the wheat in such as grain has a different kinds, different geographical and not
Same wheat breed, seed size is different, even if there are also broken kernel or scab grain etc. are not perfect with a batch Wheat Species
Grain seed is less than normal, and normal grain is just fuller, and volume and weight is relatively large, if sprouted kernel sprouts, seed also can be bigger than normal.
The grain seed of these different volumes and surface area can not suck on same standard hole type disk, once using this device,
It will suck inside smaller particle to suction ventilator, or even suction ventilator can be blocked, cause the short circuit of suction ventilator.
The detection of grain seed and counting technology are the bases of grain seed identification, each grain seed can be partitioned into
Come, so as to be identified to individual grain seed.Deep learning is being applied in various fields at present, it can be seen that
Depth learning technology is a kind of effective technological means, can be used for current grain seed detection and counts, is based on this, this hair
The bright method and device for devising a kind of grain seed detection and counting based on deep learning, to solve the above problems.
Summary of the invention
The method and device of the purpose of the present invention is to provide a kind of grain seed detection and counting based on deep learning,
To solve the problems, such as that the prior art mentioned above in the background art is unable to satisfy the detection and counting of different grain seeds.
To achieve the above object, the invention provides the following technical scheme: a kind of grain seed detection based on deep learning
With method of counting, comprising the following steps:
S1, depth convolution learning network modelling;
S2, depth convolution learning network model parameter solve.
Preferably, depth convolution learning network model construction needs to solve the detection of intensive Small object in the step S1,
Intensive Small object characteristic extracting module must be designed, dense feature extracts evolution module and module of target detection.
Preferably, the dense feature is extracted evolution module and is connected each convolutional layer by the way of intensive connection, should
Intensive connection is primarily referred to as each layer and is all attached between all layers in front, and adds in interlayer connection design deletion and duplication
Add strategy, in the training process, every certain training the number of iterations, carries out certain connection validation verification, random erasure one
It contributes lesser layer to connect, and will contribute on biggish Connection-copy to deletion connection.
Preferably, the module of target detection is used to generate target candidate frame and provides the probability value of each target frame,
Its core is internal dimensionality reduction and rises dimension process, reduces characteristic pattern dimension using the convolution kernel of 1x1, main purpose be in order to
Accelerate speed;Using bilinear interpolation method expand characteristic pattern, main purpose be in order to obtain more accurate edge and up and down
Literary information.
A kind of detection of grain seed and counting device based on deep learning, including cabinet, high definition camera, bright field illumination
Lamp, grain seed support board, transport mechanism, dark field illuminator, grain seed feeding inlet, human-computer interaction touch display screen and electromechanics
Control device, the human-computer interaction touch display screen and electromechanical controlling device are fixed on tank surface, the grain seed pan feeding
Mouth is located above cabinet, and the surface of the transport mechanism is fixed with grain seed support board, and transport mechanism is located at grain seed
Below feeding inlet, the electromechanical controlling device is electrically connected with transport mechanism, and the dark field illuminator is located at below transport mechanism, institute
It states transport mechanism to be fixed between cabinet wall front and back, the high definition camera is located in the box body top of chamber, the bright field illumination lamp
Below high definition camera, the high definition camera, bright field illumination lamp, grain seed support board, transport mechanism and dark field illuminator
On same vertical line, the high definition camera is connect with human-computer interaction touch display screen signal, and the transport mechanism is two groups
The concordant horizontally disposed belt transmission agency in front and back, and the flexible connection of grain seed support board is between two groups of conveyer belts.
Compared with prior art, the beneficial effects of the present invention are: the present invention passes through the learning network modelling of depth convolution
And parametric solution, it can be very good detection intensively and the grain seed target of adhesion, by test, the detection of grain seed is called together
The rate of returning and accuracy are very high, and calculating process is simple, easy to operate.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, will be described below to embodiment required
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for ability
For the those of ordinary skill of domain, without creative efforts, it can also be obtained according to these attached drawings other attached
Figure.
Fig. 1 is flow diagram of the present invention;
Fig. 2 is deep learning overall network structure chart of the present invention;
Fig. 3 is that dense feature of the present invention extracts evolution flow chart;
Fig. 4 is detection module network structure of the present invention;
Fig. 5 is the content of present invention perceptual structure figure;
Fig. 6 is schematic structural view of the invention.
In attached drawing, parts list represented by the reference numerals are as follows:
The intensive Small object characteristic extracting module of 1-, 2- dense feature extract evolution module, 3- module of target detection, 100- case
Body, 200- high definition camera, 300- bright field illumination lamp, 400- grain seed support board, 500- transport mechanism, 600- dark-ground illumination
Lamp, 700- grain seed feeding inlet, 800- human-computer interaction touch display screen, 900- electromechanical controlling device.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts all other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1-6 is please referred to, the present invention provides a kind of technical solution: a kind of to detect with based on by the grain seed of deep learning
Counting method, comprising the following steps:
S1, depth convolution learning network modelling;
S2, depth convolution learning network model parameter solve.
Depth convolution learning network model construction needs to solve the detection of intensive Small object in step S1, it is necessary to which design is intensive
Small object characteristic extracting module 1, dense feature extract evolution module 2 and module of target detection 3, intensive Small object feature extraction mould
The triple channel image of input color in block 1.
Dense feature is extracted evolution module 2 and is connected each convolutional layer by the way of intensive connection, the intensive connection master
Refer to that each layer is all attached between all layers in front, and connects design in interlayer and delete and replicate addition strategy,
In training process, every certain training the number of iterations, certain connection validation verification is carried out, some contributions of random erasure are smaller
Layer connection, and will contribute biggish Connection-copy to delete connection on, this effect be similar to genetic algorithm in evolution effect
Fruit.
For module of target detection 3 for generating target candidate frame and providing the probability value of each target frame, core is interior
The dimensionality reduction and liter dimension process in portion, characteristic pattern dimension are reduced using the convolution kernel of 1x1, main purpose is to accelerate speed;
Using bilinear interpolation method expand characteristic pattern, main purpose be in order to obtain more accurate edge and contextual information, it is close
Collect in feature extraction evolution module 2, designing each layer, all layers are connected with adjacent layer and before, while designing a kind of mechanism to delete
Connection is removed and increased, after repeatedly trained and iteration, when the number of iterations error change is smaller or increases, starts this machine
System.When starting this mechanism, 20% in overall connection is randomly choosed every time, finds test verifying precision when deleting some connection
There are many decline, are 1 by the linkage flag, and record accuracy decline value, illustrate that the connection is critically important, cannot delete;If smart
Degree decline is seldom or does not decline, then the linkage flag is 0, illustrates that the connection is not essential;If precision goes up not down, explanation
The connection must delete, which is -1, and records precision rising value.After a wheel connection is deleted, with those shadows
It rings overall precision and declines the connection that biggish those needs of connection replacement are deleted.So as to obtain a kind of feature extraction of evolution
Network.
32,34,35 are primarily referred to as the convolution in deep learning in module of target detection 3, can be the volume of 3x3 or 1x1
It accumulates, the direct-connected effect for being equivalent to 5x5 convolution of the convolution of two 3x3, the connection in 33 mainly increases the number of characteristic pattern, also has
Do not increasing feature map number, the same position on the same characteristic pattern carries out relevant arithmetical operation, makes in this programme
It is to increase feature map number.36,37 be target frame and score that module of target detection 3 obtains grain seed, in 31 modules
Primarily to increasing contextual information, intensive adhesion target is solved the problems, such as.Dimensionality reduction in 312 and 314 is the volume by 1x1
The long-pending number to reduce feature, so as to reduce operation.
After installing device and designing deep learning model, the training of Boot Model, the accuracy to model reaches very
When high or whole network loss very little, illustrate that network is housebroken very well.By trained model, to the figure of acquisition
Detection as carrying out grain seed.
The data detected by verification experimental verification grain seed are as follows:
Grain kernal number 9033 in figure, actually detected number are 9068, wherein correct seed is 9024, missing inspection
9, erroneous detection 44;
The recall rate is defined asWherein TP is correct detection
Number, FN are leakage kernal number, and FP is erroneous detection kernal number.
According to the definition of recall rate, it is 99.9% that actual measurement, which goes out recall rate, accuracy 99.51%;
The definition correctly detected is that degree of overlapping IOU is greater than 50% between detection seed frame bbox and practical callout box gtbox;
The IOU is defined as
A kind of detection of grain seed and counting device based on deep learning, including cabinet 100, high definition camera 200, light field
It is headlamp 300, grain seed support board 400, transport mechanism 500, dark field illuminator 600, grain seed feeding inlet 700, man-machine
Interaction touch display screen 800 and electromechanical controlling device 900, human-computer interaction touch display screen 800 and electromechanical controlling device 900 are fixed
On 100 surface of cabinet, grain seed feeding inlet 700 is located at 100 top of cabinet, and the surface of transport mechanism 500 is fixed with grain seed
Grain support board 400, and transport mechanism 500 is located at 700 lower section of grain seed feeding inlet, electromechanical controlling device 900 and transport mechanism
500 electrical connections, dark field illuminator 600 are located at 500 lower section of transport mechanism, and transport mechanism 500 is fixed on 100 inner wall front and back of cabinet
Between, high definition camera 200 is located at 100 inner cavity top of cabinet, and bright field illumination lamp 300 is located at 200 lower section of high definition camera, high definition camera
200, bright field illumination lamp 300, grain seed support board 400, transport mechanism 500 and dark field illuminator 600 are located at same vertical line
On, high definition camera 200 is connect with 800 signal of human-computer interaction touch display screen, and transport mechanism 500 is that the concordant level in two groups of front and backs is set
The belt transmission agency set, and grain seed support board 400 is flexibly connected between two groups of conveyer belts,
Grain seed enters the device by grain seed feeding inlet 700, and is carried grain seed by transport mechanism 500
Object plate 400 is sent in dark field illuminator 600, is located at immediately below bright field illumination lamp 300 and high definition camera 200, high definition camera 200
After collecting grain seed image, image is sent to electromechanical controlling device 900 by connecting line, in electromechanical controlling device 900
Include belt transmission agency control logic, Image Acquisition and storage control, deep learning algorithm, interact UI and software logic,
It finally will test result to be shown in human-computer interaction touch display screen 800, can also be carried out in human-computer interaction touch display screen 800
Grain variety selection and sensitivity selection control.
In the description of this specification, the description of reference term " one embodiment ", " example ", " specific example " etc. means
Particular features, structures, materials, or characteristics described in conjunction with this embodiment or example are contained at least one implementation of the invention
In example or example.In the present specification, schematic expression of the above terms may not refer to the same embodiment or example.
Moreover, particular features, structures, materials, or characteristics described can be in any one or more of the embodiments or examples to close
Suitable mode combines.
Present invention disclosed above preferred embodiment is only intended to help to illustrate the present invention.There is no detailed for preferred embodiment
All details are described, are not limited the invention to the specific embodiments described.Obviously, according to the content of this specification,
It can make many modifications and variations.These embodiments are chosen and specifically described to this specification, is in order to better explain the present invention
Principle and practical application, so that skilled artisan be enable to better understand and utilize the present invention.The present invention is only
It is limited by claims and its full scope and equivalent.
Claims (5)
1. a kind of detection of grain seed and method of counting based on deep learning, it is characterised in that: the following steps are included:
S1, depth convolution learning network modelling;
S2, depth convolution learning network model parameter solve.
2. a kind of method of grain seed detection and counting based on deep learning according to claim 1, feature exist
In: depth convolution learning network model construction needs to solve the detection of intensive Small object in the step S1, it is necessary to which design is intensive
Small object characteristic extracting module (1), dense feature extract evolution module (2) and module of target detection (3).
3. a kind of method of grain seed detection and counting based on deep learning according to claim 2, feature exist
In: the dense feature is extracted evolution module (2) and is connected each convolutional layer by the way of intensive connection, the intensive connection master
Refer to that each layer is all attached between all layers in front, and connects design in interlayer and delete and replicate addition strategy,
In training process, every certain training the number of iterations, certain connection validation verification is carried out, some contributions of random erasure are smaller
Layer connection, and will contribute biggish Connection-copy to delete connection on.
4. a kind of method of grain seed detection and counting based on deep learning according to claim 2, feature exist
In: for the module of target detection (3) for generating target candidate frame and providing the probability value of each target frame, core is interior
The dimensionality reduction and liter dimension process in portion, characteristic pattern dimension are reduced using the convolution kernel of 1x1, main purpose is to accelerate speed;
Characteristic pattern is expanded using the method for bilinear interpolation, main purpose is to obtain more accurate edge and contextual information.
5. a kind of detection of grain seed and counting device based on deep learning, it is characterised in that: including cabinet (100), high definition
Camera (200), bright field illumination lamp (300), grain seed support board (400), transport mechanism (500), dark field illuminator (600),
Grain seed feeding inlet (700), human-computer interaction touch display screen (800) and electromechanical controlling device (900), the human-computer interaction touching
It touches display screen (800) and electromechanical controlling device (900) is fixed on cabinet (100) surface, grain seed feeding inlet (700) position
Above cabinet (100), the surface of the transport mechanism (500) is fixed with grain seed support board (400), and transport mechanism
(500) it is located at below grain seed feeding inlet (700), the electromechanical controlling device (900) is electrically connected with transport mechanism (500),
The dark field illuminator (600) is located at below transport mechanism (500), and the transport mechanism (500) is fixed on cabinet (100) inner wall
Between front and back, the high definition camera (200) is located at cabinet (100) inner cavity top, and the bright field illumination lamp (300) is located at high definition phase
Below machine (200), the high definition camera (200), bright field illumination lamp (300), grain seed support board (400), transport mechanism
(500) it is located on same vertical line with dark field illuminator (600), the high definition camera (200) and human-computer interaction touch display screen
(800) signal connects, and the transport mechanism (500) is two groups of front and back concordantly horizontally disposed belt transmission agencies, and grain seed
Grain support board (400) is flexibly connected between two groups of conveyer belts.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811583256.4A CN109444141B (en) | 2018-12-24 | 2018-12-24 | Grain kernel detection and counting method and device based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811583256.4A CN109444141B (en) | 2018-12-24 | 2018-12-24 | Grain kernel detection and counting method and device based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109444141A true CN109444141A (en) | 2019-03-08 |
CN109444141B CN109444141B (en) | 2021-07-02 |
Family
ID=65537836
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811583256.4A Active CN109444141B (en) | 2018-12-24 | 2018-12-24 | Grain kernel detection and counting method and device based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109444141B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111982788A (en) * | 2020-07-27 | 2020-11-24 | 江苏大学 | High-speed seed counting sensor and detection method |
CN113791008A (en) * | 2021-08-25 | 2021-12-14 | 安徽高哲信息技术有限公司 | Grain imperfect grain detection equipment and detection method |
CN114030907A (en) * | 2022-01-10 | 2022-02-11 | 安徽高哲信息技术有限公司 | Feeding system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102279976A (en) * | 2011-09-22 | 2011-12-14 | 河南工业大学 | Method for constructing and identifying BP (Back Propagation) neural network for identifying different coarse rice seeds |
CN107122798A (en) * | 2017-04-17 | 2017-09-01 | 深圳市淘米科技有限公司 | Chin-up count detection method and device based on depth convolutional network |
WO2018104897A1 (en) * | 2016-12-08 | 2018-06-14 | Sigtuple Technologies Private Limited | A method and system for determining quality of semen sample |
CN108615046A (en) * | 2018-03-16 | 2018-10-02 | 北京邮电大学 | A kind of stored-grain pests detection recognition methods and device |
CN108776807A (en) * | 2018-05-18 | 2018-11-09 | 复旦大学 | It is a kind of based on can the double branch neural networks of skip floor image thickness grain-size classification method |
-
2018
- 2018-12-24 CN CN201811583256.4A patent/CN109444141B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102279976A (en) * | 2011-09-22 | 2011-12-14 | 河南工业大学 | Method for constructing and identifying BP (Back Propagation) neural network for identifying different coarse rice seeds |
WO2018104897A1 (en) * | 2016-12-08 | 2018-06-14 | Sigtuple Technologies Private Limited | A method and system for determining quality of semen sample |
CN107122798A (en) * | 2017-04-17 | 2017-09-01 | 深圳市淘米科技有限公司 | Chin-up count detection method and device based on depth convolutional network |
CN108615046A (en) * | 2018-03-16 | 2018-10-02 | 北京邮电大学 | A kind of stored-grain pests detection recognition methods and device |
CN108776807A (en) * | 2018-05-18 | 2018-11-09 | 复旦大学 | It is a kind of based on can the double branch neural networks of skip floor image thickness grain-size classification method |
Non-Patent Citations (3)
Title |
---|
GAO HUANG等: "Densely Connected Convolutional Networks", 《COMPUTER SOCIETY》 * |
胡安翔 等: "基于Faster R-CNN改进的数粒机系统", 《包装工程》 * |
郭杰荣 等: "《光电信息技术实验教程》", 30 September 2015 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111982788A (en) * | 2020-07-27 | 2020-11-24 | 江苏大学 | High-speed seed counting sensor and detection method |
CN111982788B (en) * | 2020-07-27 | 2022-07-22 | 江苏大学 | High-speed seed counting sensor and detection method |
CN113791008A (en) * | 2021-08-25 | 2021-12-14 | 安徽高哲信息技术有限公司 | Grain imperfect grain detection equipment and detection method |
CN113791008B (en) * | 2021-08-25 | 2024-03-15 | 安徽高哲信息技术有限公司 | Grain imperfect grain detection equipment and detection method |
CN114030907A (en) * | 2022-01-10 | 2022-02-11 | 安徽高哲信息技术有限公司 | Feeding system |
Also Published As
Publication number | Publication date |
---|---|
CN109444141B (en) | 2021-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110060237B (en) | Fault detection method, device, equipment and system | |
CN111178197B (en) | Mass R-CNN and Soft-NMS fusion based group-fed adherent pig example segmentation method | |
CN109444141A (en) | A kind of method and device of grain seed detection and counting based on deep learning | |
CN108596166B (en) | Container number identification method based on convolutional neural network classification | |
CN111696128B (en) | High-speed multi-target detection tracking and target image optimization method and storage medium | |
CN110245697B (en) | Surface contamination detection method, terminal device and storage medium | |
CN112153483B (en) | Information implantation area detection method and device and electronic equipment | |
CN107545271B (en) | Image recognition method, device and system | |
CN110723432A (en) | Garbage classification method and augmented reality equipment | |
CN110378313A (en) | Cell mass recognition methods, device and electronic equipment | |
CN113781485B (en) | Intelligent detection method and device for PCB defect types, electronic equipment and medium | |
CN113033558B (en) | Text detection method and device for natural scene and storage medium | |
CN111914902B (en) | Traditional Chinese medicine identification and surface defect detection method based on deep neural network | |
CN110533654A (en) | The method for detecting abnormality and device of components | |
CN109636801B (en) | Electronic device and medium for automatically acquiring HER2 fluorescence in situ hybridization result | |
US11600088B2 (en) | Utilizing machine learning and image filtering techniques to detect and analyze handwritten text | |
CN111767908A (en) | Character detection method, device, detection equipment and storage medium | |
CN110287767A (en) | Can attack protection biopsy method, device, computer equipment and storage medium | |
CN111369317B (en) | Order generation method, order generation device, electronic equipment and storage medium | |
CN111753690B (en) | Dish tray identification method and dish identification method based on same | |
Peng et al. | Automatic monitoring system for seed germination test based on deep learning | |
CN111163332A (en) | Video pornography detection method, terminal and medium | |
CN114972947B (en) | Depth scene text detection method and device based on fuzzy semantic modeling | |
CN111932515B (en) | Short circuit detection method and system for product residual defects and defect classification system | |
CN104391981A (en) | Text classification method and device |
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 |