CN107766856A - A kind of ore visible images method for separating based on Adaboost machine learning - Google Patents
A kind of ore visible images method for separating based on Adaboost machine learning Download PDFInfo
- Publication number
- CN107766856A CN107766856A CN201610715882.9A CN201610715882A CN107766856A CN 107766856 A CN107766856 A CN 107766856A CN 201610715882 A CN201610715882 A CN 201610715882A CN 107766856 A CN107766856 A CN 107766856A
- Authority
- CN
- China
- Prior art keywords
- ore
- adaboost
- machine learning
- visible images
- image
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Abstract
The invention discloses a kind of ore visible images method for separating based on Adaboost machine learning, belongs to automation ore deposit selecting technology field, comprises the following steps：(1) ore primary election is classified；(2) image preprocessing；(3) machine learning model is trained；(4) target prediction.The present invention utilizes Adaboost machine learning method, ore visible images are learnt, trained, predicted, the more rich technique of preparing personnel of experience can fully be simulated and carry out ore sorting, every machine is set all to possess same, accurate ore sorting experience, it effectively prevent the subjectivity and individual difference of artificial separation ore, people can preferably be substituted to carry out work or complete the work that the mankind can not complete, labor intensity is reduced, improve production quality and labor productivity.
Description
Technical field
The invention belongs to automate ore deposit selecting technology field, more particularly to a kind of ore based on Adaboost machine learning can
See light image method for separating.
Background technology
Machine learning is a multi-field cross discipline, is related to probability theory, statistics, Approximation Theory, convextiry analysis, algorithm complexity
The multi-door subject such as topology degree.Specialize in the learning behavior that the mankind were simulated or realized to computer how, with obtain new knowledge or
Technical ability, reorganize the existing structure of knowledge and be allowed to constantly improve the performance of itself.Machine learning is the core of artificial intelligence, is
Make computer that there is the fundamental way of intelligence, every field of its application throughout artificial intelligence.
Supervised learning in machine learning is to infer the machine learning task of One function from the training data of mark.Instruction
Practicing data includes a set of training example.In supervised learning, each example be by an input object (be usually vector) and
One desired output valve (also referred to as supervisory signals) composition.Supervised learning algorithm is to analyze the training data, and produces one
The function of deduction, it can be used for mapping out new example.One optimal scheme will allow the algorithm correctly to determine that
The class label of a little invisible examples.This require learning algorithm be a kind of mode of " reasonable " from one kind from training data to
Formed in the case of invisible.
Decision tree is on the basis of known various situation probability of happening, and the phase of net present value (NPV) is asked for by forming decision tree
Prestige value is more than or equal to zero probability, assessment item risk, judges the method for decision analysis of its feasibility, is intuitively with probability point
A kind of diagram method of analysis.Because this decision branch is drawn as limb of the figure like one tree, therefore claim decision tree.In machine learning
In, decision tree is a forecast model, and what he represented is a kind of mapping relations between object properties and object value.Decision tree is
A kind of tree structure, wherein each internal node represents the test on an attribute, each branch represents a test output, often
Individual leaf node represents a kind of classification.Decision tree is a kind of very conventional sorting technique, and he is a kind of supervised learning, and classics represent
Boosting models.
In recent years, as China is with perfect, the accumulation of technology, fund of supporting infrastructure, profit in all trades and professions
Manually intelligence solves the problems, such as that the demand of actual production starts to occur extensively, domestic near about institution of higher learning, research institute and enterprise
Positive thinking and bold trial have been carried out in field of artificial intelligence within 2 years, progressively started the application of industry spot.
However, current China mining activities, when ore sorts still using the technical staff with abundant examination experience
Ore deposit choosing is carried out, the drawbacks of it is present is it is clear that experienced technical staff is a small number of after all.Utilize machine learning skill
The powerful superiority of art：The learning mechanic of human neuronal is simulated, is contained much information, function is more, efficiency high.If by machine learning
Technology detects applied to ore primary election, by with the incomparable advantage of artificial detection.Effective line of the ore deposit element in ore
Though the key character of reason distribution and color with uniqueness, feature is fine and smooth, discontinuous, causes very big difficulty for sorting, passes through
It is difficult to hold when testing not good enough technical staff's sorting ore, experienced personnel can fully be simulated by being predicted using machine learning
Study and discriminating power, but also validity feature can be quantitatively described, avoid the separation results to vary with each individual, subtract
Small ore primary election error, improve production efficiency and sharpness of separation.Machine learning is established on the basis of objective analysis and reasoning,
The subjectivity manually screened and individual difference are effectively prevent, improves accuracy of detection.The purpose of a little technical research of engineering
Exactly machine is preferably substituted people come work according to the Reasoning With Learning characteristic of the mankind or complete the mankind from the work completed,
Labor intensity is reduced, improves constantly production quality and labor productivity.Compared with traditional ore dressing mode, machine learning skill
Art has significant advantage and huge economic value, and detection efficiency is high, and detection has continuity and repeatability.
However, machine learning techniques are used for industrial ore primary election, training sample set storehouse is set up in a manner of artificial tape label,
Because researcher is to the difference of ore cognition, by different degrees of introduction human error, this human error is by certain journey
The precision of machine learning model is influenceed on degree；Secondly the number of plies of machine learning model is more, and precision of prediction will obtain certain carry
Height, but predetermined speed is slower, the bad control of equalization point of the model number of plies and precision；Again, the parameter of machine learning model is set
It is improper to put, and will produce over-fitting or the problem of poor fitting.In summary, machine learning techniques simulation mankind's Reasoning With Learning, its
Model also needs to continue to optimize.
The content of the invention
For above-mentioned the deficiencies in the prior art, the invention provides a kind of ore based on Adaboost machine learning is visible
Light image method for separating, this method are learnt to ore visible images, trained, predicted, it is relatively abundant can fully to simulate experience
Technique of preparing personnel carry out ore sorting, every machine is all possessed same, accurate ore sorting experience, can effectively keep away
The subjectivity and individual difference of manpower-free's sorting ore, it can preferably substitute people and carry out work or complete what the mankind can not complete
Work, labor intensity is reduced, improve production quality and labor productivity.
To achieve these goals, the present invention is achieved through the following technical solutions：
A kind of ore visible images method for separating based on Adaboost machine learning, comprises the following steps：
(1) ore primary election is classified：Ore deposit of the content below 0.15% is classified as barren rock, the ore deposit of content more than 0.15% is classified as ore deposit
Stone；
(2) image preprocessing：
A. OTSU segmentations are carried out to the gray level image of the ore visible images of sampling：
B. profile maximum region in image, and its corresponding boundary rectangle are found out after segmentation：
C. the corresponding region in color ore visible images, as required ROI region are determined according to external matrix；
D., the image of the rgb color space of ROI region is converted to the image of HSI color spaces；
E. the gray scale of 0 °, 45 °, 90 °, 135 ° four direction for calculating the image H passages that ROI corresponds to HSI color spaces is total to
Raw matrix；
F. take 0 ° after calculating, 45 °, 90 °, on 135 ° of four directions, the maximum matrix of consequence of entropy is deposited as characteristic vector
Enter characteristic vector storehouse；
(3) machine learning model is trained：Machine learning model uses the AdaBoost models based on decision tree, AdaBoost
For adaptive Boosting versions, i.e., multiple Weak Classifiers are formed into a strong classifier, at the same time AdaBoost can also
Effectively suppressed study；
A. the characteristic vector storehouse with ore and barren rock label information is inputted into Adaboost models, allow its carry out study with
Training；
B. to the end of Adaboost model learnings, final Adaboost models will be generated；
(4) target prediction：
A. ore sampling picture to be predicted is subjected to image preprocessing；
B. the characteristic vector calculated is inputted into the AdaBoost models trained and be predicted；
C. prediction result exports.
Further, in the step (2) when profile maximum region in image after finding out segmentation, area filtering threshold takes
2000 pixels.
Further, gray level co-occurrence matrixes in the step (2), maximum tonal gradation parameter value take 20.
Further, gray level co-occurrence matrixes in the step (2), distance parameter value take 5.
Further, gray level co-occurrence matrixes in the step (2), maximum grey parameter value take 360.
Further, Adaboost models in the step (3), type use Gentle Adaboost.
Further, Adaboost models in the step (3), Weak Classifier number take 450.
Further, Adaboost models in the step (3), the maximum number of plies take 5 layers.
Further, parameter and (2) image preprocessing, the training of (3) machine learning model are pre- used in the step (4)
Survey.
The beneficial effects of the invention are as follows：Using Adaboost machine learning method, to ore visible images
Practise, train, prediction, can fully simulate the more rich technique of preparing personnel of experience and carry out ore sorting, every machine is all possessed
Same, accurate ore sorting experience, effectively prevent the subjectivity and individual difference of artificial separation ore, can be more preferable
Substitute people to carry out work or complete the work that the mankind can not complete, reduce labor intensity, improve production quality and work is given birth to
Produce efficiency.
Embodiment
The present invention is described in detail below in conjunction with specific embodiment, herein illustrative examples and explanation of the invention
For explaining the present invention, but it is not as a limitation of the invention.
A kind of ore visible images method for separating based on Adaboost machine learning, comprises the following steps：
(1) ore primary election is classified：Ore deposit of the content below 0.15% is classified as barren rock, the ore deposit of content more than 0.15% is classified as ore deposit
Stone；
(2) image preprocessing：
A. OTSU segmentations are carried out to the gray level image of the ore visible images of sampling；
B. profile maximum region in image, and its corresponding boundary rectangle are found out after segmentation；
C. the corresponding region in color ore visible images, as required ROI region are determined according to external matrix；
D., the image of the rgb color space of ROI region is converted to the image of HSI color spaces；
E. the gray scale of 0 °, 45 °, 90 °, 135 ° four direction for calculating the image H passages that ROI corresponds to HSI color spaces is total to
Raw matrix；
F. take 0 ° after calculating, 45 °, 90 °, on 135 ° of four directions, the maximum matrix of consequence of entropy is deposited as characteristic vector
Enter characteristic vector storehouse；
(3) machine learning model is trained：Machine learning model uses the AdaBoost models based on decision tree, AdaBoost
For adaptive Boosting versions, i.e., multiple Weak Classifiers are formed into a strong classifier, at the same time AdaBoost can also
Effectively suppressed study；
A. the characteristic vector storehouse with ore and barren rock label information is inputted into Adaboost models, allow its carry out study with
Training；
B. to the end of Adaboost model learnings, final Adaboost models will be generated；
(4) target prediction：
A. ore sampling picture to be predicted is subjected to image preprocessing；
B. the characteristic vector calculated is inputted into the AdaBoost models trained and be predicted；
C. prediction result exports.
Further, in step (2) when profile maximum region in image after finding out segmentation, area filtering threshold takes 2000
Individual pixel.
Further, gray level co-occurrence matrixes in step (2), maximum tonal gradation parameter value take 20.
Further, gray level co-occurrence matrixes in step (2), distance parameter value take 5.
Further, gray level co-occurrence matrixes in step (2), maximum grey parameter value take 360.
Further, Adaboost models in step (3), type use Gentle Adaboost.
Further, Adaboost models in step (3), Weak Classifier number take 450.
Further, Adaboost models in step (3), the maximum number of plies take 5 layers.
Further, parameter is predicted with (2) image preprocessing, the training of (3) machine learning model used in step (4).
Because ore adopts figure light path conditions constant, area filtering threshold takes 2000 pixels, is with several ore sample graphs
The statistical value of ore area is foundation in piece, and as area threshold ＞ 2000, the ROI region being partitioned into is ore effective district
Domain, as area threshold ＜=2000, multiple ROI regions will be partitioned into, i.e., the non-UNICOM of ore profile, lose details, Huo Zheshou
The influence of other background profiles of dropping introduces noise, and interference is brought for the feature extraction of the effective area image of ore.
Gray level co-occurrence matrixes, maximum tonal gradation parameter value take 20, and this parameter takes too senior general to increase computation burden, and redundancy
Information is too many, too small, loses characteristic information, takes 20 as EQUILIBRIUM CALCULATION FOR PROCESS burden, time and characteristic information amount；Distance parameter value can
Between taking 5~10, the fine and smooth changes in distribution distance of the more identical ore gray scale texture of value between 5~10；Maximum grey parameter
Value takes 360, is angle because HSI color space H channel values obtain implication, so for the texture features of analysis color comprehensively, Gu takes
This parameter is 360, extracts the texture color changes in distribution feature of mineral surface comprehensively.
Using the Gentle Adaboost models based on mensurable feature, by effective over-fitting suppressed in machine learning
With poor fitting problem, and there is preferable Generalization Capability；For balance model complexity and efficiency, and within the effective time
To prediction result, while ensure the precision of prediction of model, Weak Classifier number takes 450 to take 5 layers with the maximum number of plies.
Illustrate above with respect to parameter value, actual tests constantly regulate gained preferably parameter when being scheelite sorting,
Both the precision of prediction of model had been ensure that, has balanced complexity and execution efficiency again., need to be special according to ore when sorting other ores
Property, finely tune above parameter.
Claims (9)
1. a kind of ore visible images method for separating based on Adaboost machine learning, it is characterised in that including following step
Suddenly：
(1) ore primary election is classified：Ore deposit of the content below 0.15% is classified as barren rock, the ore deposit of content more than 0.15% is classified as ore；
(2) image preprocessing：
A. OTSU segmentations are carried out to the gray level image of the ore visible images of sampling；
B. profile maximum region in image, and its corresponding boundary rectangle are found out after segmentation；
C. the corresponding region in color ore visible images, as required ROI region are determined according to external matrix；
D., the image of the rgb color space of ROI region is converted to the image of HSI color spaces；
E. calculating ROI corresponds to the gray scale symbiosis square of 0 °, 45 °, 90 °, 135 ° four direction of the image H passages of HSI color spaces
Battle array；
F. take 0 ° after calculating, 45 °, 90 °, on 135 ° of four directions, the maximum matrix of consequence of entropy is used as characteristic vector, deposit spy
Levy vectorial storehouse；
(3) machine learning model is trained：Machine learning model uses the AdaBoost models based on decision tree, and AdaBoost is certainly
The Boosting versions of adaptation, i.e., multiple Weak Classifiers are formed into a strong classifier, at the same time AdaBoost can also be effective
Suppressed study；
A. the characteristic vector storehouse with ore and barren rock label information is inputted into Adaboost models, allows it to be learnt and be instructed
Practice；
B. to the end of Adaboost model learnings, final Adaboost models will be generated；
(4) target prediction：
A. ore sampling picture to be predicted is subjected to image preprocessing；
B. the characteristic vector calculated is inputted into the AdaBoost models trained and be predicted；
C. prediction result exports.
2. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In in the step (2) when profile maximum region in image after finding out segmentation, area filtering threshold takes 2000 pixels.
3. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In gray level co-occurrence matrixes in the step (2), maximum tonal gradation parameter value takes 20.
4. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In gray level co-occurrence matrixes in the step (2), distance parameter value takes 5.
5. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In gray level co-occurrence matrixes in the step (2), maximum grey parameter value takes 360.
6. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In Adaboost models in the step (3), type uses Gentle Adaboost.
7. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In Adaboost models in the step (3), Weak Classifier number takes 450.
8. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In Adaboost models in the step (3), the maximum number of plies takes 5 layers.
9. the ore visible images method for separating based on Adaboost machine learning, its feature exist as claimed in claim 1
In parameter is predicted with (2) image preprocessing, the training of (3) machine learning model used in the step (4).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610715882.9A CN107766856A (en) | 2016-08-20 | 2016-08-20 | A kind of ore visible images method for separating based on Adaboost machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610715882.9A CN107766856A (en) | 2016-08-20 | 2016-08-20 | A kind of ore visible images method for separating based on Adaboost machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107766856A true CN107766856A (en) | 2018-03-06 |
Family
ID=61264304
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610715882.9A Pending CN107766856A (en) | 2016-08-20 | 2016-08-20 | A kind of ore visible images method for separating based on Adaboost machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107766856A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109740656A (en) * | 2018-12-26 | 2019-05-10 | 华侨大学 | A kind of ore method for separating based on convolutional neural networks |
| CN111259972A (en) * | 2020-01-20 | 2020-06-09 | 北矿机电科技有限责任公司 | Flotation bubble identification method based on cascade classifier |
| CN112365484A (en) * | 2020-11-16 | 2021-02-12 | 北京霍里思特科技有限公司 | Classification method, classification system and mineral product classifier |
| CN112580659A (en) * | 2020-11-10 | 2021-03-30 | 湘潭大学 | Ore identification method based on machine vision |
| CN113591583A (en) * | 2021-06-30 | 2021-11-02 | 沈阳科来沃电气技术有限公司 | Intelligent boron ore beneficiation system and method |
| CN117252784A (en) * | 2023-11-16 | 2023-12-19 | 肇庆市大正铝业有限公司 | Automatic sorting system for reclaimed aluminum alloy raw materials |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102855492A (en) * | 2012-07-27 | 2013-01-02 | 中南大学 | Classification method based on mineral flotation foam image |
| CN103235954A (en) * | 2013-04-23 | 2013-08-07 | 南京信息工程大学 | Improved AdaBoost algorithm-based foundation cloud picture identification method |
| CN104850854A (en) * | 2015-05-08 | 2015-08-19 | 广西师范大学 | Talc ore product sorting processing method and talc ore product sorting system |
| CN204719776U (en) * | 2015-05-08 | 2015-10-21 | 广西师范大学 | Talc Ore product sorting disposing system |
| CN105678224A (en) * | 2015-12-31 | 2016-06-15 | 中国电子科技集团公司第二十七研究所 | Image identification system for intelligent selection of ore, and selection method using the same |
-
2016
- 2016-08-20 CN CN201610715882.9A patent/CN107766856A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102855492A (en) * | 2012-07-27 | 2013-01-02 | 中南大学 | Classification method based on mineral flotation foam image |
| CN103235954A (en) * | 2013-04-23 | 2013-08-07 | 南京信息工程大学 | Improved AdaBoost algorithm-based foundation cloud picture identification method |
| CN104850854A (en) * | 2015-05-08 | 2015-08-19 | 广西师范大学 | Talc ore product sorting processing method and talc ore product sorting system |
| CN204719776U (en) * | 2015-05-08 | 2015-10-21 | 广西师范大学 | Talc Ore product sorting disposing system |
| CN105678224A (en) * | 2015-12-31 | 2016-06-15 | 中国电子科技集团公司第二十七研究所 | Image identification system for intelligent selection of ore, and selection method using the same |
Non-Patent Citations (2)
| Title |
|---|
| 王兰莎: "多目标矿业复杂图像特征提取与分类", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 赵守香等著: "《大数据分析与应用》", 31 December 2015, 航空工业出版社 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109740656A (en) * | 2018-12-26 | 2019-05-10 | 华侨大学 | A kind of ore method for separating based on convolutional neural networks |
| CN111259972A (en) * | 2020-01-20 | 2020-06-09 | 北矿机电科技有限责任公司 | Flotation bubble identification method based on cascade classifier |
| CN111259972B (en) * | 2020-01-20 | 2023-08-11 | 北矿机电科技有限责任公司 | Flotation bubble identification method based on cascade classifier |
| CN112580659A (en) * | 2020-11-10 | 2021-03-30 | 湘潭大学 | Ore identification method based on machine vision |
| CN112365484A (en) * | 2020-11-16 | 2021-02-12 | 北京霍里思特科技有限公司 | Classification method, classification system and mineral product classifier |
| CN113591583A (en) * | 2021-06-30 | 2021-11-02 | 沈阳科来沃电气技术有限公司 | Intelligent boron ore beneficiation system and method |
| CN117252784A (en) * | 2023-11-16 | 2023-12-19 | 肇庆市大正铝业有限公司 | Automatic sorting system for reclaimed aluminum alloy raw materials |
| CN117252784B (en) * | 2023-11-16 | 2024-04-02 | 肇庆市大正铝业有限公司 | Automatic sorting system for reclaimed aluminum alloy raw materials |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107766856A (en) | A kind of ore visible images method for separating based on Adaboost machine learning | |
| Gill et al. | Fruit image classification using deep learning | |
| CN108009509A (en) | Vehicle target detection method | |
| CN112819802B (en) | Method for supervising and predicting blast furnace condition abnormality based on tuyere information deep learning | |
| CN108304316B (en) | Software defect prediction method based on collaborative migration | |
| CN107481188A (en) | A kind of image super-resolution reconstructing method | |
| CN109741328A (en) | A kind of automobile apparent mass detection method based on production confrontation network | |
| CN105825511A (en) | Image background definition detection method based on deep learning | |
| JP2021534055A (en) | How to determine the total alumina production index based on a multi-scale deep convolutional neural network | |
| CN105678278A (en) | Scene recognition method based on single-hidden-layer neural network | |
| CN104463199A (en) | Rock fragment size classification method based on multiple features and segmentation recorrection | |
| CN105718959B (en) | A kind of object identification method based on from coding | |
| CN105654505B (en) | A kind of collaboration track algorithm and system based on super-pixel | |
| Jiang et al. | Delving into sample loss curve to embrace noisy and imbalanced data | |
| Zeng et al. | Steel sheet defect detection based on deep learning method | |
| CN103559642A (en) | Financial data mining method based on cloud computing | |
| CN103903009A (en) | Industrial product detection method based on machine vision | |
| CN109214467B (en) | Transformer operator dressing robust identification method considering classifier output sensitivity | |
| Yin et al. | Recognition of grape leaf diseases using MobileNetV3 and deep transfer learning | |
| CN106127813B (en) | The monitor video motion segments dividing method of view-based access control model energy sensing | |
| CN103034868A (en) | Classification method aiming at small sample and high dimensional images | |
| CN107423771A (en) | A kind of two phase method for detecting change of remote sensing image | |
| Ning et al. | Background Modeling and Fuzzy Clustering for Motion Detection from Video. | |
| Nienaber et al. | Spiral concentrator interface monitoring through image processing: A statistical learning approach | |
| Chen et al. | A feature extraction strategy of fire hole video based on VGG16 and migration learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180306 |
|
| RJ01 | Rejection of invention patent application after publication |