CN110309807A - CAD diagram paper intelligent identification Method - Google Patents

CAD diagram paper intelligent identification Method Download PDF

Info

Publication number
CN110309807A
CN110309807A CN201910610394.5A CN201910610394A CN110309807A CN 110309807 A CN110309807 A CN 110309807A CN 201910610394 A CN201910610394 A CN 201910610394A CN 110309807 A CN110309807 A CN 110309807A
Authority
CN
China
Prior art keywords
diagram paper
component
cad diagram
text information
cad
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
Application number
CN201910610394.5A
Other languages
Chinese (zh)
Inventor
杨涛
周浩然
石国伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northwestern Polytechnical University
Northwest University of Technology
Original Assignee
Northwest University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Northwest University of Technology filed Critical Northwest University of Technology
Priority to CN201910610394.5A priority Critical patent/CN110309807A/en
Publication of CN110309807A publication Critical patent/CN110309807A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40Document-oriented image-based pattern recognition
    • G06V30/41Analysis of document content
    • G06V30/413Classification of content, e.g. text, photographs or tables
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40Document-oriented image-based pattern recognition
    • G06V30/41Analysis of document content
    • G06V30/418Document matching, e.g. of document images
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40Document-oriented image-based pattern recognition
    • G06V30/42Document-oriented image-based pattern recognition based on the type of document
    • G06V30/422Technical drawings; Geographical maps

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of CAD diagram paper intelligent identification Method, the technical issues of the practicability is poor for solving existing CAD diagram paper recognition method.Technical solution is classified to a large amount of drawings of importing, and whether told automatically using the convolutional neural networks based on region has the equipment such as electrical equipment and component;For the drawing containing electrical equipment or component, component therein and text information are positioned and identified respectively also with the convolutional neural networks method based on region;Each component and its corresponding text information are matched, and it is mono- to generate BOM.The present invention is handled and is analyzed to CAD diagram paper by intelligent recognition and matched method, largely reduces the consumption of manpower, and obtains higher recognition speed and accuracy rate, and practicability is good.

Description

CAD diagram paper intelligent identification Method
Technical field
The present invention relates to a kind of CAD diagram paper recognition method, in particular to a kind of CAD diagram paper intelligent identification Method.
Background technique
The high speed development of computer technology plays an increasingly important role it in many industries, as engineering design, Computer technology bring convenience has all been enjoyed in the fields such as machine-building, and cad technique is exactly a kind of very outstanding computer Technology.Since 21 century, equipment manufacture industry generally carries out engineering practice using CAD software, this sets traditional product Deep variation has occurred in meter method and production model, and the economic development for society brings huge economic benefit.However, with CAD diagram paper file it is cumulative, engineering design field needs a kind of method that can accurately extract CAD diagram paper content, with The Material Takeoff of quotation, design and production process is rapidly completed convenient for enterprise, and then promotes the production efficiency of enterprise.
Document " Chinese invention patent that application publication number is CN102693334A " discloses a kind of based on CAD electronic drawing Dynamic component recognition methods.This method grabs the two-dimemsional number on CAD diagram paper by using interactive devices such as mouse, keyboards manually According to;Candidate primitive information is searched by way of traversal in some near field;It further extracts optimal in ranking results Primitive information feeds back to client.For this method, this process is not only relatively complicated, but also the sequence of candidate primitive information Mode is easy to be influenced by manual operation.
Summary of the invention
In order to overcome the shortcomings of existing CAD diagram paper recognition method, the practicability is poor, and the present invention provides a kind of CAD diagram paper intelligently knowledge Other method.This method classifies to a large amount of drawings of importing, and being told automatically using the convolutional neural networks based on region is It is no to have the equipment such as electrical equipment and component;For the drawing containing electrical equipment or component, also with based on region Convolutional neural networks method is positioned and is identified respectively to component therein and text information;To each component and its right The text information answered matches, and it is mono- to generate BOM.The present invention by intelligent recognition and matched method to CAD diagram paper at Reason and analysis, largely reduce the consumption of manpower, and obtain higher recognition speed and accuracy rate, practicability It is good.
The technical solution adopted by the present invention to solve the technical problems: a kind of CAD diagram paper intelligent identification Method, its main feature is that The following steps are included:
Step 1: the automatic classification of CAD diagram paper.
Drawing content is identified using the convolutional neural networks based on region, found out from candidate CAD diagram paper containing The drawing of component.
Neural network model is trained using the CAD diagram paper of existing tape label, model is enable sufficiently to learn every kind of member The characteristics of image of device;
CAD diagram paper to be sorted is detected using trained neural network model, specified component will be contained Drawing is sorted out, and CAD diagram paper is classified and managed automatically.
Step 2: CAD diagram paper component and text information intelligent recognition based on deep learning.
Include many components and its corresponding model text information in CAD diagram paper, constructs two convolution based on region Neural network model proposes component and text information, using supervised learning method, learns the pixel letter in each CAD image Breath, for new CAD image, accomplishes the identification component and text information of precise and high efficiency.The convolution mind based on region used Various information in CAD diagram paper are grasped by the method depth of supervised learning through network model.
It is generated Step 3: the Auto-matching and BOM of CAD diagram paper component and text information are mono-.
By in CAD diagram paper component and its corresponding text information match, result is stored in the mono- form of BOM And displaying.A convolutional neural networks model is constructed, the component identified before is associated with text information, it will be first Device corresponds with text information.Matched result is converted into BOM simple form formula.
The beneficial effects of the present invention are: this method classifies to a large amount of drawings of importing, the convolution based on region is utilized Whether neural network is told automatically the equipment such as electrical equipment and component;For the figure containing electrical equipment or component Paper is positioned and is known respectively to component therein and text information also with the convolutional neural networks method based on region Not;Each component and its corresponding text information are matched, and it is mono- to generate BOM.The present invention passes through intelligent recognition and matching Method CAD diagram paper is handled and is analyzed, largely reduce the consumption of manpower, and obtain higher identification Speed and accuracy rate, practicability are good.
Specifically, 1. recognition speeds are fast.
Using itself, previously study had been identified and had been divided to new CAD diagram paper to characteristics of image rule deep learning model Class does not need again to handle new drawing, this is directly a process end to end, makes the recognition speed of drawing than existing Some methods are faster.
2. convenient for operation.
Due to having been completed the training and optimization of CAD diagram paper intelligent recognition model before, therefore it may only be necessary to by be identified CAD diagram paper be importing directly into identification model, do not need additional operation, this than existing method save many manpowers, Material resources and time cost.
3. accuracy rate is high.
Deep learning method is compared with other methods, and advantage can exactly extract the deeper feature of object, has been slapped Hold the rule of its essence.CAD diagram paper intelligent recognition based on deep learning shows in accuracy rate compared with existing method more excellent.
4. scalability is strong.
Existing CAD diagram paper content identification method is often directed to some aspect and specific mode is taken to go to be known Not, it needs to model again when identifying object and requiring to change, this is unfavorable for the extension of CAD diagram paper identification work.And CAD diagram paper intelligent identification Method based on deep learning, it is only necessary to enough training datas can solve the problems, such as it is a kind of, and Model itself has good generalization ability, even if certain change has occurred in object, can also guarantee to know to a certain extent Other accuracy rate.
It elaborates with reference to the accompanying drawings and detailed description to the present invention.
Detailed description of the invention
Fig. 1 is the flow chart of CAD diagram paper intelligent identification Method of the present invention.
Fig. 2 is CAD diagram paper automatic classification method flow chart in embodiment of the present invention method.
Fig. 3 is that RPN picture material frame selects schematic diagram in embodiment of the present invention method.
Fig. 4 is CAD diagram paper content detection effect picture in embodiment of the present invention method.
Fig. 5 is CAD diagram paper component and text information identification and matching process flow chart in embodiment of the present invention method.
Specific embodiment
Referring to Fig.1-5.Specific step is as follows for CAD diagram paper intelligent identification Method of the present invention:
Step 1: manually being marked to a certain amount of CAD image, sample is provided for the study of neural network model;Step Rapid two, Faster-RCNN model is constructed, it is trained using the CAD image with label, depth grasps CAD diagram paper Content information;Classify Step 3: CAD diagram paper to be identified is imported into trained neural network model, if certain A CAD diagram paper without specific component (such as high and low pressure bin, distribution box), then continues to import and identify next through detecting CAD diagram paper is opened, if so, then carrying out the identification and matching of component and text information;Step 4: by matched result with BOM Single form is shown and stores.Detailed embodiment is as follows:
One, the automatic classification of CAD diagram paper.
In CAD diagram paper identification process, to do is to the automatic classification of CAD diagram paper first, can not only be chosen by classification The drawing with certain components is selected, and the text information of component and its model can be positioned.
Original feature extraction is carried out firstly, CAD image is input in a convolutional neural networks, obtains original image Characteristic pattern.
Then, the characteristic pattern extracted is input in RPN, the effect of RPN is to generate candidate frame.RPN can be in original graph The different candidate frame of multiple areas, length-width ratio is generated as in and is scored it, and chooses highest frame conduct of wherein scoring The input of next step.Rois (regions of interest) is obtained after the completion of frame choosing, it is the feature of candidate frame corresponding position Value.
Next, rois is input in Fast RCNN, the effect of Fast RCNN is examined to the content of institute's frame choosing It surveys.In order to quickly generate candidate frame, RPN and Fast RCNN share convolutional layer.Training for sharing convolutional layer needs as follows Step:
A. the rois that RPN is generated before utilizing, by Fast RCNN training, one is individually detected network;
B. with the training of detection netinit RPN, shared convolutional layer is fixed at this time, and finely tunes the exclusive layer of RPN;
C. it keeps shared convolutional layer to fix, the full articulamentum of Fast RCNN is finely tuned, in this way, two network shares are identical Convolutional layer, constitute a unified network.
By way of this shared parameter, it can make model that CAD image content is quickly positioned and detected.
Finally, classifying according to testing result to CAD diagram paper, and obtain exact position of the component in figure.
Two, CAD diagram paper component is identified and is matched with the synchronous intelligent of text information.
The identification and matching of CAD diagram paper component and text information are built upon on the automatic basis of classification of drawing, i.e., first Component is positioned and detected with text information.
Firstly, according to the process of CAD diagram paper automatic content classification component and text information are carried out respectively positioning and Identification, but it is different from the automatic classification of CAD diagram paper, the identification of component and text information requires to carry out respectively.Therefore, for member Device and text information train respective neural network model, and between two models be it is independent, be independent of each other, utilize instruction The neural network perfected is positioned and is identified respectively with text information to the component in image.
Then, it needs to position the corresponding model text information of resulting component to match.For accurate perception Corresponding relationship between them constructs a convolutional neural networks model, using component as label, to component and text information Between corresponding relationship learnt.In face of new CAD diagram paper, component and text information accurately can be carried out one by model One identification and matching.
Finally, showing and storing obtained component and its text information of model in BOM mono- form.

Claims (1)

1. a kind of CAD diagram paper intelligent identification Method, it is characterised in that the following steps are included:
Step 1: the automatic classification of CAD diagram paper;
Drawing content is identified using the convolutional neural networks based on region, is found out from candidate CAD diagram paper containing first device The drawing of part;
Neural network model is trained using the CAD diagram paper of existing tape label, model is enable sufficiently to learn every kind of component Characteristics of image;
CAD diagram paper to be sorted is detected using trained neural network model, by the drawing containing specified component Sorted out, CAD diagram paper is classified and managed automatically;
Step 2: CAD diagram paper component and text information intelligent recognition based on deep learning;
Include many components and its corresponding model text information in CAD diagram paper, constructs two convolutional Neurals based on region Network model proposes that component and text information learn the Pixel Information in each CAD image using supervised learning method, right In new CAD image, accomplish the identification component and text information of precise and high efficiency;The convolutional neural networks based on region used Model grasps various information in CAD diagram paper by the method depth of supervised learning;
It is generated Step 3: the Auto-matching and BOM of CAD diagram paper component and text information are mono-;
By in CAD diagram paper component and its corresponding text information match, result is stored and exhibition in the mono- form of BOM Show;A convolutional neural networks model is constructed, the component identified before is associated with text information, by component It corresponds with text information;Matched result is converted into BOM simple form formula.
CN201910610394.5A 2019-07-08 2019-07-08 CAD diagram paper intelligent identification Method Pending CN110309807A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910610394.5A CN110309807A (en) 2019-07-08 2019-07-08 CAD diagram paper intelligent identification Method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910610394.5A CN110309807A (en) 2019-07-08 2019-07-08 CAD diagram paper intelligent identification Method

Publications (1)

Publication Number Publication Date
CN110309807A true CN110309807A (en) 2019-10-08

Family

ID=68078259

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910610394.5A Pending CN110309807A (en) 2019-07-08 2019-07-08 CAD diagram paper intelligent identification Method

Country Status (1)

Country Link
CN (1) CN110309807A (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110765639A (en) * 2019-11-05 2020-02-07 国网重庆市电力公司电力科学研究院 Electrical simulation modeling method and device and readable storage medium
CN110795809A (en) * 2019-11-07 2020-02-14 国网河北省电力有限公司电力科学研究院 Method, system and medium for transformer substation electrical secondary circuit connection relation based on CAD graph automatic pickup
CN110909650A (en) * 2019-11-15 2020-03-24 清华大学 CAD drawing identification method and device based on domain knowledge and target detection
CN111079528A (en) * 2019-11-07 2020-04-28 国网辽宁省电力有限公司电力科学研究院 Primitive drawing checking method and system based on deep learning
CN111160144A (en) * 2019-12-16 2020-05-15 广东施富电气实业有限公司 Method and system for identifying components by combining electric drawing with pictures and texts and storage medium
CN111160018A (en) * 2019-12-13 2020-05-15 广东施富电气实业有限公司 Method and system for recognizing non-component text of electrical drawing and storage medium
CN111242024A (en) * 2020-01-11 2020-06-05 北京中科辅龙科技股份有限公司 Method and system for recognizing legends and characters in drawings based on machine learning
CN111814791A (en) * 2020-07-24 2020-10-23 西门子(中国)有限公司 Method and device for identifying components in system diagram
CN112036814A (en) * 2020-08-18 2020-12-04 苏州加非猫精密制造技术有限公司 Intelligent production management system and management method based on 3D retrieval technology
CN112085791A (en) * 2020-08-26 2020-12-15 广州市纬纶国际建筑设计有限公司 Automatic positioning method, device and equipment for construction design drawing and storage medium
CN112861713A (en) * 2021-02-06 2021-05-28 贵州博汇云技术开发有限公司 Large-scale drawing multi-local amplification comparison analysis system
CN112883801A (en) * 2021-01-20 2021-06-01 上海品览智造科技有限公司 Accurate identification method for household distribution box system diagram subgraph in CAD distribution system diagram
CN113298697A (en) * 2021-03-19 2021-08-24 广州天越电子科技有限公司 Method for converting two-dimensional graphic elements into vector graphic elements based on artificial neural network
CN113792614A (en) * 2021-08-24 2021-12-14 四川渔光物联技术有限公司 Photovoltaic array group string component position matching and serial number identification method
WO2022132040A1 (en) * 2020-12-17 2022-06-23 National University Of Singapore Systems for ai-driven creation of bill of materials
CN114821599A (en) * 2022-04-21 2022-07-29 国网河南省电力公司电力科学研究院 Method for identifying characteristic graphic element in electrical drawing
CN118134398A (en) * 2024-05-06 2024-06-04 安徽省交通规划设计研究总院股份有限公司 Marker quantity table generation system, device and medium based on CAD object characteristics

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102609687A (en) * 2012-01-31 2012-07-25 华中科技大学 Subway construction drawing and engineering parameter automatic identification method
CN105205448A (en) * 2015-08-11 2015-12-30 中国科学院自动化研究所 Character recognition model training method based on deep learning and recognition method thereof
CN109376758A (en) * 2018-09-07 2019-02-22 广州算易软件科技有限公司 A kind of Identify chip method, system, device and storage medium based on figure
CN109446689A (en) * 2018-11-07 2019-03-08 国网江苏省电力有限公司电力科学研究院 DC converter station electrical secondary system drawing recognition methods and system
CN109446885A (en) * 2018-09-07 2019-03-08 广州算易软件科技有限公司 A kind of text based Identify chip method, system, device and storage medium
CN109918746A (en) * 2019-02-19 2019-06-21 西门子电站自动化有限公司 The method and apparatus for generating bill of materials based on CAD drawing file

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102609687A (en) * 2012-01-31 2012-07-25 华中科技大学 Subway construction drawing and engineering parameter automatic identification method
CN105205448A (en) * 2015-08-11 2015-12-30 中国科学院自动化研究所 Character recognition model training method based on deep learning and recognition method thereof
CN109376758A (en) * 2018-09-07 2019-02-22 广州算易软件科技有限公司 A kind of Identify chip method, system, device and storage medium based on figure
CN109446885A (en) * 2018-09-07 2019-03-08 广州算易软件科技有限公司 A kind of text based Identify chip method, system, device and storage medium
CN109446689A (en) * 2018-11-07 2019-03-08 国网江苏省电力有限公司电力科学研究院 DC converter station electrical secondary system drawing recognition methods and system
CN109918746A (en) * 2019-02-19 2019-06-21 西门子电站自动化有限公司 The method and apparatus for generating bill of materials based on CAD drawing file

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
公安部第三研究所: "《多摄像机协同关注目标检测跟踪技术》", 30 June 2017, 东南大学出版社 *

Cited By (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110765639A (en) * 2019-11-05 2020-02-07 国网重庆市电力公司电力科学研究院 Electrical simulation modeling method and device and readable storage medium
CN110765639B (en) * 2019-11-05 2023-08-04 国网重庆市电力公司电力科学研究院 Electrical simulation modeling method and device and readable storage medium
CN110795809A (en) * 2019-11-07 2020-02-14 国网河北省电力有限公司电力科学研究院 Method, system and medium for transformer substation electrical secondary circuit connection relation based on CAD graph automatic pickup
CN111079528A (en) * 2019-11-07 2020-04-28 国网辽宁省电力有限公司电力科学研究院 Primitive drawing checking method and system based on deep learning
CN110795809B (en) * 2019-11-07 2023-08-29 国网河北省电力有限公司电力科学研究院 Method, system and medium for automatically picking up connection relation of electric secondary circuit of transformer substation based on CAD (computer aided design) drawing
CN110909650A (en) * 2019-11-15 2020-03-24 清华大学 CAD drawing identification method and device based on domain knowledge and target detection
CN111160018B (en) * 2019-12-13 2022-11-01 广东施富电气实业有限公司 Method and system for recognizing non-component text of electrical drawing and storage medium
CN111160018A (en) * 2019-12-13 2020-05-15 广东施富电气实业有限公司 Method and system for recognizing non-component text of electrical drawing and storage medium
CN111160144A (en) * 2019-12-16 2020-05-15 广东施富电气实业有限公司 Method and system for identifying components by combining electric drawing with pictures and texts and storage medium
CN111160144B (en) * 2019-12-16 2023-04-07 广东施富电气实业有限公司 Method and system for identifying components by combining electric drawing with pictures and texts and storage medium
CN111242024A (en) * 2020-01-11 2020-06-05 北京中科辅龙科技股份有限公司 Method and system for recognizing legends and characters in drawings based on machine learning
CN111814791B (en) * 2020-07-24 2024-03-19 西门子(中国)有限公司 Method and device for identifying components in system graph
CN111814791A (en) * 2020-07-24 2020-10-23 西门子(中国)有限公司 Method and device for identifying components in system diagram
CN112036814A (en) * 2020-08-18 2020-12-04 苏州加非猫精密制造技术有限公司 Intelligent production management system and management method based on 3D retrieval technology
CN112085791A (en) * 2020-08-26 2020-12-15 广州市纬纶国际建筑设计有限公司 Automatic positioning method, device and equipment for construction design drawing and storage medium
WO2022132040A1 (en) * 2020-12-17 2022-06-23 National University Of Singapore Systems for ai-driven creation of bill of materials
CN112883801A (en) * 2021-01-20 2021-06-01 上海品览智造科技有限公司 Accurate identification method for household distribution box system diagram subgraph in CAD distribution system diagram
CN112883801B (en) * 2021-01-20 2024-05-24 上海品览智造科技有限公司 Accurate identification method for resident distribution box system diagram sub-graph in CAD distribution system diagram
CN112861713A (en) * 2021-02-06 2021-05-28 贵州博汇云技术开发有限公司 Large-scale drawing multi-local amplification comparison analysis system
CN113298697A (en) * 2021-03-19 2021-08-24 广州天越电子科技有限公司 Method for converting two-dimensional graphic elements into vector graphic elements based on artificial neural network
CN113298697B (en) * 2021-03-19 2024-04-26 广州天越电子科技有限公司 Method for converting two-dimensional graphic elements into vector graphic elements based on artificial neural network
CN113792614A (en) * 2021-08-24 2021-12-14 四川渔光物联技术有限公司 Photovoltaic array group string component position matching and serial number identification method
CN114821599A (en) * 2022-04-21 2022-07-29 国网河南省电力公司电力科学研究院 Method for identifying characteristic graphic element in electrical drawing
CN118134398A (en) * 2024-05-06 2024-06-04 安徽省交通规划设计研究总院股份有限公司 Marker quantity table generation system, device and medium based on CAD object characteristics

Similar Documents

Publication Publication Date Title
CN110309807A (en) CAD diagram paper intelligent identification Method
CN107368700A (en) Based on the microbial diversity interaction analysis system and method for calculating cloud platform
CN108564115A (en) Semi-supervised polarization SAR terrain classification method based on full convolution GAN
CN107545000A (en) The information-pushing method and device of knowledge based collection of illustrative plates
CN108268527B (en) A method of detection land use pattern variation
CN109919106A (en) Gradual target finely identifies and description method
Cetinic et al. Automated painter recognition based on image feature extraction
US20210056466A1 (en) System and methodology for data classification, learning and transfer
CN114066848B (en) FPCA appearance defect visual detection system
CN110377727A (en) A kind of multi-tag file classification method and device based on multi-task learning
CN111932639B (en) Detection method of unbalanced defect sample based on convolutional neural network
CN111415131A (en) Big data talent resume analysis method based on natural language processing technology
CN105868269A (en) Precise image searching method based on region convolutional neural network
CN107305555A (en) Data processing method and device
CN104809468A (en) Multi-view classification method based on indefinite kernels
CN112434173B (en) Search content output method and device, computer equipment and readable storage medium
CN113191320A (en) Food safety index monitoring method based on deep learning
CN110110129B (en) Fabric image retrieval method based on fine texture characterization
CN110377790A (en) A kind of video automatic marking method based on multi-modal privately owned feature
CN113344031B (en) Text classification method
Spehr et al. Image statistics for clustering paintings according to their visual appearance
CN109543512A (en) The evaluation method of picture and text abstract
CN115631374A (en) Control operation method, control detection model training method, device and equipment
Deac et al. Feature selection for paintings classification by optimal tree pruning
CN108108721A (en) A kind of method that road extraction is carried out using EO-1 hyperion

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20191008