CN109002841A - A kind of building element extracting method based on Faster-RCNN model - Google Patents
A kind of building element extracting method based on Faster-RCNN model Download PDFInfo
- Publication number
- CN109002841A CN109002841A CN201810677252.6A CN201810677252A CN109002841A CN 109002841 A CN109002841 A CN 109002841A CN 201810677252 A CN201810677252 A CN 201810677252A CN 109002841 A CN109002841 A CN 109002841A
- Authority
- CN
- China
- Prior art keywords
- building element
- block
- xml
- file
- sum
- 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
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000013480 data collection Methods 0.000 claims abstract description 15
- 230000011218 segmentation Effects 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 14
- 238000010200 validation analysis Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 230000001629 suppression Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 8
- 238000001514 detection method Methods 0.000 description 16
- 238000000605 extraction Methods 0.000 description 9
- 238000013527 convolutional neural network Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 241000894007 species Species 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- BQCADISMDOOEFD-UHFFFAOYSA-N Silver Chemical compound [Ag] BQCADISMDOOEFD-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 229910052709 silver Inorganic materials 0.000 description 2
- 239000004332 silver Substances 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 101001050287 Mus musculus Transcription factor Jun Proteins 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000003925 brain function Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000013016 learning Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000004377 microelectronic Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of building element extracting methods based on Faster-RCNN model, firstly, the pretreatment such as binaryzation, segmentation is carried out to architectural engineering drawing image, to obtain image block data collection；Then, the building element in the data set is labeled by LabelImg tool, building element identification model is obtained using Faster-RCNN training, and the building element information that model is extracted in drawing is stored in the form of structuring.The method of the present invention effectively extracts building element information in manual architectural engineering drawing, so that manual architectural engineering drawing utilization rate improves, and increases the use value of manual architectural engineering drawing.
Description
Technical field
The invention belongs to deep learnings and image object detection technique field, in particular to a kind of to be based on Faster-RCNN
The building element extracting method of model.
Background technique
Building element extracting method in the present invention has important role and meaning to the utilization of manual architectural drawings paper.
When facing image object test problems, researchers can select traditional Feature Extraction Technology or based on deep learning
Target detection technique carries out image object detection by above-mentioned technology.Manual architectural drawing is carried out using image object detection technique
Paper building element extracts work, to improve the utilization rate of manual architectural drawing, the utilization for manual architectural drawing provides solution
Certainly scheme.
The existing Research foundation of Feng Wanli, Zhu Quanyin et al. includes: Wanli Feng.Research oftheme
statement extraction for chinese literature based on lexical
chain.International Journal of Multimedia and Ubiquitous Engineering,Vol.11,
No.6(2016),pp.379-388；Wanli Feng,Ying Li,Shangbing Gao,Yunyang Yan,Jianxun
Xue.A novel flame edge detection algorithm via a novel active contour
model.International Journal ofHybrid Information Technology,Vol.9,No.9(2016),
pp.275-282；Method for mode matching [J] microelectronics and computer of Liu Jinling, the Feng Wanli based on Feature Dependence relationship,
2011,28(12):167-170；Liu Jinling, Feng Wanli, Zhang Yahong initialize cluster class center and the text of reconstruct scaling function is poly-
Class [J] computer application research, 2011,28 (11): 4115-4117；Liu Jinling, Feng Wanli, Zhang Yahong are based on scale again
Chinese short message Text Clustering Method [J] computer engineering and application, 2012,48 (21): 146-150.；Zhu Quanyin, Pan Lu,
Liu Wenru waits .Web science and technology news classification extraction algorithm [J] Huaiyingong College journal, 2015,24 (5): 18-24；Li Xiang, Zhu
Collaborative filtering recommending [J] the computer science and explore, 2014,8 (6): 751- that full silver joint cluster and rating matrix are shared
759；Quanyin Zhu,Sunqun Cao.ANovel Classifier-independent Feature Selection
Algorithm for Imbalanced Datasets.2009,p:77-82；Quanyin Zhu,Yunyang Yan,Jin
Ding,Jin Qian.The Case Study for Price Extracting of Mobile Phone Sell
Online.2011,p:282-285；Quanyin Zhu,Suqun Cao,Pei Zhou,Yunyang Yan,Hong
Zhou.Integrated Price Forecast based on Dichotomy Backfilling and Disturbance
Factor Algorithm.International Review on Computers and Software,2011,Vol.6
(6):1089-1093；Zhu Quanyin, Feng Wanli et al. application, openly with the related patents of authorization: Feng Wanli, Shao Heshuai, Zhuan Jun
A kind of intelligent refrigerated car state monitoring wireless network terminal installation: CN203616634U [P] .2014；Zhu Quanyin, Hu Rongjing, what
Su Qun, a kind of price forecasting of commodity method Chinese patent based on linear interpolation Yu Adaptive windowing mouth of such as week training: ZL
201110423015.5,2015.07.01；Zhu Quanyin, Cao Suqun, Yan Yunyang, Hu Rong wait quietly, and one kind is repaired based on two divided datas
With the price forecasting of commodity method Chinese patent of disturbing factors: ZL 201110422274.6,2013.01.02；Li Xiang, Zhu Quan
Silver, Hu Ronglin, a kind of all deep Cold Chain Logistics prestowage intelligent recommendation method China Patent Publication No. based on spectral clustering of:
CN105654267A,2016.06.08。
The detection of field image object:
The purpose of image object detection is the target object that setting type is picked out from the background of differing complexity, and
Separating background, to complete the follow-up works such as tracking, identification.Present invention is primarily based on Faster-RCNN models to carry out building structure
Part extracts, and extracts building element information in manual architectural engineering drawing.
Depth convolutional neural networks:
Deep Convolutional Neural Networks (DCNN) i.e. depth convolutional neural networks.With depth
The it is proposed of habit, deep learning are that depth convolutional neural networks are logical the main reason for field of image recognition makes great progress
Cross the automatic learning characteristic of training data.A kind of this deep learning method for attempting to simulate human brain function of DCNN, by Hinton
Raw Krizhevsky etc. has been defeated with the advantage close to 10% based on traditional artificial in 2012 using 8 layers of DCNN (AlexNet)
The method of feature mainly contains 4 kinds of basic operations: convolution, pond (Pooling), full connection and nonlinear transformation.
Faster-RCNN:
Present invention research using Faster-RCNN model, Faster-RCNN be after RCNN, Fast RCNN,
The leader Ross Girshick team newest fruits of image recognition, the RCNN from RCNN to Fast, newest proposition
Faster RCNN uniformly arrives four basic steps of object candidate area generation, feature extraction, classification, position refine detection
Within one depth network frame.All calculating do not repeat, and candidate region is extracted on GPU using RPN, to improve fortune
Scanning frequency degree.In an experiment, Faster R-CNN reaches 17fps to the detection speed of simple network target, on PASCAL VOC
Accuracy rate is 59.9%；Complex network reaches 5fps, and accuracy rate reaches 78.8%.
Faster-RCNN is that the algorithm of target detection based on two-stage is divided into two steps, the first step for target detection
It is using selective search formula method (Selective Search, referred to as heuristic search) or to utilize convolutional Neural net
Network --- Region Proposal Network (target area generates network, RPN) generates target acquisition prediction block, is divided
Class returns and prediction work.
The Faster RCNN first step generates anchor point and prediction alternative area using RPN, carries out feature extraction using DCNN,
Classify next for feature extraction result, object boundary frame regression forecasting finally is carried out to target acquisition region
(Bounding-box regression)。
Faster RCNN carries out the recurrence of target prediction bounding box, and principle is to find a kind of function f, mesh according to anchor point coordinate
Make:
The module of recurrence is Intersection over Union (IoU), it is defined as follows, wherein
Detection Result is testing result, and Ground Truth is legitimate reading:
For building element information in manual architectural engineering drawing, traditional mode is the side for passing through eye recognition and counting
Formula, there is building element classification identification mistake, omission building element and a large amount of manpowers of needs by way of eye recognition
Problem.
Summary of the invention
Goal of the invention: in view of the above problems, the present invention provides one kind by comprehensive analysis, extracts to automate
The building element extracting method based on Faster-RCNN model of building element information in architectural drawing.
Technical solution: the present invention proposes a kind of building element extracting method based on Faster-RCNN model, including as follows
Step:
(1) gray processing, binaryzation, segmentation pretreatment are carried out to architectural engineering pretreatment drawing data collection Drawing, obtained
Pre-process line drawing block data set DrawingBlock；
(2) using building element identification model of the image block data collection DrawingBlock training based on Faster-RCNN
DrawingModel；
(3) building element information is extracted based on building element identification model DrawingModel, obtains building element data
Collect Result；
(4) structured storage building element data set Result.
Further, the specific side of pretreatment line drawing block data set DrawingBlock is obtained in the step (1)
Method are as follows:
(1.1) definition pretreatment drawing data collection Drawing, pre-processes line drawing block data set DrawingBlock,
Drawing number DrawingNumber, designation of drawing DrawingName, drawing length Width, drawing height Height, drawing are total
Number Sum, image block Block, image block number BlockNumber, image block sum BlockSum, define drawing DrawingInfo,
Wherein DrawingInfo={ DrawingNumber, DrawingName, Width, Height }, Drawing=
{DrawingInfo1,DrawingInfo2,…,DrawingInfoSum, Block=DrawingNumber,
BlockNumber }, DrawingBlock={ Block1,Block2,…,BlockBlockSum, counter S1 is defined, S1=is enabled
1, drawing divides number B, and wherein S1 is for traversing Drawing；
(1.2) drawing sum is calculated, Sum is assigned to；
(1.3) if S1≤Sum in step (1.1), (1.4) are thened follow the steps, otherwise, are executed step (1.14), wherein
S1∈[1,Sum]；
(1.4) using OpenCV Runtime Library to DrawingInfoiImage gray processing, binary conversion treatment are carried out, wherein
DrawingInfoi∈Drawing；
(1.5) linear smoothing filtering, smooth DrawingInfo are carried out based on Sobel operatoriImage；
(1.6) defined variable w, h, j, k obtains DrawingInfoiWidth, Height, DrawingNumber, if w
=Width/B, h=Height/B, j=1, wherein w is image block length, and h is image block width, and j, k is based on segmented image
Number, j, k ∈ [1, B]；
(1.7) it if j≤B, thens follow the steps (1.8), it is no to then follow the steps (1.11)；
(1.8) k=1 in step (1.6) is enabled；
(1.9) it if k≤B, thens follow the steps (1.10), it is no to then follow the steps (1.12)；
(1.10) it with coordinate { w*j, w* (j+1), h*k, h* (k+1) } for four vertex segmented images of rectangular shaped rim, enables
BlockNumber=(j-1) * B+k obtains image block Block, Block={ DrawingNumber, BlockNumber },
DrawingBlock=DrawingBlock ∪ Block；
(1.11) S1=S1+1 is enabled；
(1.12) j=j+1 is enabled；
(1.13) k=k+1 is enabled；
(1.14) pretreatment line drawing block data set DrawingBlock is obtained.
Further, Faster-RCNN is based on using image block data collection DrawingBlock training in the step (2)
Building element identification model DrawingModel method particularly includes:
(2.1) definition XML file data set DrawingXML, variable temp, variable Tr, variable V a, variable Te, training number
According to collection TrainBlcok, validation data set ValidationBlcok, test data set TestBlcok, XML file TrainXML,
XML file ValidationXML, XML file TestXML, csv file TrainCSV, csv file ValidationCSV, CSV
File TestCSV, counter S2, enables S2=1, for traversing DrawingBlock；
(2.2) it calculates image block sum and is assigned to BlockSum；
(2.3) if S2≤BlockSum, then follow the steps (2.4), it is no to then follow the steps (2.6), wherein S2 ∈ [1,
BlockSum]；
(2.4) to BlockiBuilding element mark is carried out using LabelImg tool, obtains markup information file xmli, enable
DrawingXML=DrawingXML ∪ xmli, xmliStorage form is XML file；
(2.5) S2=S2+1 is enabled；
(2.6) temp=BlockSum/ (train+validation+test), Tr=train*temp, Va=are set
Validation*temp, Te=test*temp, according to Tr, Va, Te is right respectively by DrawingBlock, DrawingXML at random
It should divide, obtain TrainBlcok={ Block1,Block2,…,BlockTr, ValidationBlcok={ Block1,
Block2,…,BlockVa, TestBlcok={ Block1,Block2,…,BlockTe, TrainXML={ xml1,
xml2,…,xmlTr, ValidationXML={ xml1,xml2,…,xmlVa, TestXML={ xml1,xml2,…,xmlTe}；
(2.7) file TrainXML, ValidationXML, TestXML are converted to data file TrainCSV respectively,
ValidationCSV, TestCSV, file layout CSV；
(2.8) using TensorFlow data file tool by architectural drawing tile data collection TrainBlcok and
TrainCSV is fabricated to data file Train.record, and ValidationBlcok and ValidationCSV are fabricated to data text
Part Validation.record, TestBlcok and TestCSV are fabricated to data file Test.record；
(2.9) training Faster-RCNN model, being provided with model parameter non-maxima suppression threshold value is 0.7, candidate frame
Maximum quantity is 300, and initial alignment loses weight 2.0, and Classification Loss weight is 1.0, and maximum number of iterations 200000 loses letter
Number selection softmax loss；
(2.10) the building element identification model DrawingModel based on Faster-RCNN model is obtained.
Further, obtain building element data set Result's in the step (3) method particularly includes:
(3.1) architectural drawing sum DSum, drawing set D={ drawing to be identified are defined1,drawing2,…,
drawingDSum, building element number of species ClassSum, building element type set Classes={ class1,
class2,…,classClassSum, locating rectangle frame coordinate xmin, xmax, ymin, ymax, building element posting sum
BoxSum, building element posting coordinate set DetectionBoxes={ { xmin1,xmax1,ymin1,ymax1},{xmin2,
xmax2,ymin2,ymax2},..,{xminBoxSum,xmaxBoxSum,yminBoxSum,ymaxBoxSum, building element posting is built
Build Component Category DetectionClasses, building element posting category score DetectionScores, building element identification
Results set Result={ { DetectionBoxes1,DetectionClasses1,DetectionScores1},
{DetectionBoxes2,DetectionClasses2,DetectionScores2},…,
{DetectionBoxesDrawingSum,DetectionClassesDrawingSum,DetectionScoresDrawingSum, counter
S3 enables S3=1；
(3.2) it calculates architectural drawing sum to be identified and is assigned to Dsum, building element kind is obtained by " standard for architectural drawing "
Class set Classes, building element number of species ClassSum；
(3.3) no to then follow the steps (3.6) if S3≤Dsum is thened follow the steps (3.4), wherein [1, Dsum] S3 ∈；
(3.4) using building element identification model DrawingModel to drawing drawingiBuilding element identification is carried out,
Obtain recognition result { DetectionBoxesi,DetectionClassesi,DetectionScoresi, Result=
Result∪{DetectionBoxesi,DetectionClassesi,DetectionScoresi}；
(3.5) S3=S3+1；
(3.6) building element data set Result is obtained.
The present invention is mainly according to " standard for architectural drawing " (GB/T 50104-2010) and to be based on Faster-RCNN model
Carry out building element extraction.Building element information in manual architectural engineering drawing is extracted, so that manual architectural engineering drawing utilizes
Rate improves, and increases the use value of manual architectural engineering drawing
The present invention by adopting the above technical scheme, has the advantages that
The method of the present invention carries out building element identification to manual architectural drawing using Faster-RCNN model, improves and builds
The utilization efficiency of drawing is built, it is specific: firstly, the present invention handles architectural engineering drawing, and to use LabelImg tool
Building element in drawing is labeled, building element identification model is obtained using Faster-RCNN training, then according to model
It extracts the building in drawing and constructs information, and structured storage building element extracts result.The present invention is built by image preprocessing
Component target detection is built, is mainly solved two problems, one of them is the excessive problem of architectural engineering craft blueprint pixel value, separately
One be building element target detection research.By Faster RCNN model, building element is preferably identified, saving is built
Drawing memory space is built, architectural engineering drawing is preferably managed, obtains architectural drawing information.A building element in this way
Extracting method reduces the operating time of architectural drawing building element information extraction significantly, preferably serves manual architectural drawing
Mechanism is managed and used, the utilization efficiency of manual architectural drawing is improved.
Detailed description of the invention
Fig. 1 is overall flow figure of the invention；
Fig. 2 is the flow chart of architectural engineering drawing preprocess method in Fig. 1；
Fig. 3 is the flow chart of building element identification model method of the training based on Faster-RCNN in Fig. 1；
Fig. 4 is that the building element identification model in Fig. 1 based on Faster-RCNN carries out building element information extracting method
Flow chart.
Specific embodiment
Combined with specific embodiments below, the present invention is furture elucidated, it should be understood that these embodiments are merely to illustrate the present invention
Rather than limit the scope of the invention, after the present invention has been read, those skilled in the art are to various equivalences of the invention
The modification of form falls within the application range as defined in the appended claims.
As shown in Figs 1-4, a kind of building element extracting method based on Faster-RCNN model of the present invention, packet
Include following steps:
Step 1: gray processing, binaryzation, segmentation pretreatment are carried out to architectural engineering pretreatment drawing data collection Drawing,
Pretreatment line drawing block data set DrawingBlock is obtained, specific as shown in Figure 2:
Step 1.1: definition pretreatment drawing data collection Drawing pre-processes line drawing block data set
DrawingBlock, drawing number DrawingNumber, designation of drawing DrawingName, drawing length Width, drawing height
Height, drawing sum Sum, image block Block, image block number BlockNumber, image block sum BlockSum, definition figure
Paper DrawingInfo, wherein DrawingInfo={ DrawingNumber, DrawingName, Width, Height },
Drawing={ DrawingInfo1,DrawingInfo2,…,DrawingInfoSum, Block=DrawingNumber,
BlockNumber }, DrawingBlock={ Block1,Block2,…,BlockBlockSum, counter S1 is defined, S1=is enabled
1, drawing divides number B, and wherein S1 is for traversing Drawing；
Step 1.2: calculating drawing sum, be assigned to Sum；
Step 1.3: if S1≤Sum in step 1.1, thening follow the steps 1.4, otherwise, execute step 1.14, wherein S1 ∈
[1,Sum]；
Step 1.4: using OpenCV Runtime Library to DrawingInfoiImage gray processing, binary conversion treatment are carried out, wherein
DrawingInfoi∈Drawing；
Step 1.5: linear smoothing filtering, smooth DrawingInfo are carried out based on Sobel operatoriImage；
Step 1.6: defined variable w, h, j, k obtain DrawingInfoiWidth, Height, DrawingNumber,
If w=Width/B, h=Height/B, j=1, wherein w is image block length, and h is image block width, and j, k are used for segmented image
It counts, j, k ∈ [1, B]；
Step 1.7: if j≤B, then follow the steps 1.8, it is no to then follow the steps 1.11；
Step 1.8: enabling the k=1 in step 1.6；
Step 1.9: if k≤B, then follow the steps 1.10, it is no to then follow the steps 1.12；
Step 1.10: with coordinate { w*j, w* (j+1), h*k, h* (k+1) } for four vertex segmented images of rectangular shaped rim, enabling
BlockNumber=(j-1) * B+k obtains image block Block, Block={ DrawingNumber, BlockNumber },
DrawingBlock=DrawingBlock ∪ Block；
Step 1.11: enabling S1=S1+1；
Step 1.12: enabling j=j+1；
Step 1.13: enabling k=k+1；
Step 1.14: obtaining pretreatment line drawing block data set DrawingBlock.
Step 2: identifying mould using building element of the image block data collection DrawingBlock training based on Faster-RCNN
Type DrawingModel, specific as shown in Figure 3:
Step 2.1: definition XML file data set DrawingXML, variable temp, variable Tr, variable V a, variable Te, instruction
Practice data set TrainBlcok, validation data set ValidationBlcok, test data set TestBlcok, XML file
TrainXML, XML file ValidationXML, XML file TestXML, csv file TrainCSV, csv file
ValidationCSV, csv file TestCSV, counter S2 enable S2=1, for traversing DrawingBlock；
Step 2.2: calculating image block sum and be assigned to BlockSum；
Step 2.3: if S2≤BlockSum, then follow the steps 2.4, it is no to then follow the steps 2.6, wherein S2 ∈ [1,
BlockSum]；
Step 2.4: to BlockiBuilding element mark is carried out using LabelImg tool, obtains markup information file xmli,
Enable DrawingXML=DrawingXML ∪ xmli, xmliStorage form is XML file；
Step 2.5: enabling S2=S2+1；
Step 2.6: setting temp=BlockSum/ (train+validation+test), Tr=train*temp, Va=
Validation*temp, Te=test*temp, according to Tr, Va, Te is right respectively by DrawingBlock, DrawingXML at random
It should divide, obtain TrainBlcok={ Block1,Block2,…,BlockTr, ValidationBlcok={ Block1,
Block2,…,BlockVa, TestBlcok={ Block1,Block2,…,BlockTe, TrainXML={ xml1,
xml2,…,xmlTr, ValidationXML={ xml1,xml2,…,xmlVa, TestXML={ xml1,xml2,…,xmlTe}；
Step 2.7: file TrainXML, ValidationXML, TestXML are converted to data file respectively
TrainCSV, ValidationCSV, TestCSV, file layout CSV；
Step 2.8: using TensorFlow data file tool by architectural drawing tile data collection TrainBlcok and
TrainCSV is fabricated to data file Train.record, and ValidationBlcok and ValidationCSV are fabricated to data text
Part Validation.record, TestBlcok and TestCSV are fabricated to data file Test.record；
Step 2.9: training Faster-RCNN model, being provided with model parameter non-maxima suppression threshold value is 0.7, is waited
Selecting frame maximum quantity is 300, and initial alignment loses weight 2.0, and Classification Loss weight is 1.0, maximum number of iterations 200000, damage
It loses function and selects softmax loss；
Step 2.10: obtaining the building element identification model DrawingModel based on Faster-RCNN model.
Step 3: building element information being extracted based on building element identification model DrawingModel, obtains building element number
It is specific as shown in Figure 4 according to collection Result:
Step 3.1: defining architectural drawing sum DSum to be identified, drawing set D={ drawing1,drawing2,…,
drawingDSum, building element number of species ClassSum, building element type set Classes={ class1,
class2,…,classClassSum, locating rectangle frame coordinate xmin, xmax, ymin, ymax, building element posting sum
BoxSum, building element posting coordinate set DetectionBoxes={ { xmin1,xmax1,ymin1,ymax1},{xmin2,
xmax2,ymin2,ymax2},..,{xminBoxSum,xmaxBoxSum,yminBoxSum,ymaxBoxSum, building element posting is built
Build Component Category DetectionClasses, building element posting category score DetectionScores, building element identification
Results set Result={ { DetectionBoxes1,DetectionClasses1,DetectionScores1},
{DetectionBoxes2,DetectionClasses2,DetectionScores2},…,
{DetectionBoxesDrawingSum,DetectionClassesDrawingSum,DetectionScoresDrawingSum, counter
S3 enables S3=1；
Step 3.2: calculating architectural drawing sum to be identified and be assigned to Dsum, building element is obtained by " standard for architectural drawing "
Type set Classes, building element number of species ClassSum；
Step 3.3: no to then follow the steps 3.6 if S3≤Dsum thens follow the steps 3.4, wherein [1, Dsum] S3 ∈；
Step 3.4: using building element identification model DrawingModel to drawing drawingiCarry out building element knowledge
Not, recognition result { DetectionBoxes is obtainedi,DetectionClassesi,DetectionScoresi, Result=
Result∪{DetectionBoxesi,DetectionClassesi,DetectionScoresi}；
Step 3.5:S3=S3+1；
Step 3.6: obtaining building element data set Result.
Wherein, Faster-RCNN model carries out building building identification model training, obtains building building identification model,
In for Faster-RCNN, it is 0.7 that model parameter non-maxima suppression threshold value, which is arranged, and candidate frame maximum quantity is 300, initially
Positioning loss weight 2.0, Classification Loss weight are 1.0, and maximum number of iterations 200000, loss function selects softmax
loss。
By handling 11257 architectural plans and architectural working drawing, architectural engineering drawing is pre-processed,
Building element in drawing is labeled using Open-Source Tools first, then makes building element data set, uses Faster-RCNN
Training obtains building element identification model, then constructs information according to the building in model extraction drawing, and structured storage mentions
Take result.In the architectural drawing for only containing single building element, building element target detection accuracy rate is 99%, is built a variety of
It builds in the architectural drawing of component, building element target detection accuracy rate reaches 87%.And building element proposed by the present invention extracts
Method is generally applicable to architectural drawing target detection problems.
Claims (4)
1. a kind of building element extracting method based on Faster-RCNN model, which comprises the steps of:
(1) gray processing, binaryzation, segmentation pretreatment are carried out to architectural engineering pretreatment drawing data collection Drawing, obtains pre- place
Manage line drawing block data set DrawingBlock；
(2) using building element identification model of the image block data collection DrawingBlock training based on Faster-RCNN
DrawingModel；
(3) building element information is extracted based on building element identification model DrawingModel, obtains building element data set
Result；
(4) structured storage building element data set Result.
2. a kind of building element extracting method based on Faster-RCNN model according to claim 1, feature exist
In obtaining pretreatment line drawing block data set DrawingBlock in the step (1) method particularly includes:
(1.1) definition pretreatment drawing data collection Drawing, pre-processes line drawing block data set DrawingBlock, drawing
Number DrawingNumber, designation of drawing DrawingName, drawing length Width, drawing height Height, drawing sum
Sum, image block Block, image block number BlockNumber, image block sum BlockSum define drawing DrawingInfo,
Middle DrawingInfo={ DrawingNumber, DrawingName, Width, Height }, Drawing=
{DrawingInfo1,DrawingInfo2,…,DrawingInfoSum, Block=DrawingNumber,
BlockNumber }, DrawingBlock={ Block1,Block2,…,BlockBlockSum, counter S1 is defined, S1=is enabled
1, drawing divides number B, and wherein S1 is for traversing Drawing；
(1.2) drawing sum is calculated, Sum is assigned to；
(1.3) if S1≤Sum in step (1.1), (1.4) are thened follow the steps, otherwise, are executed step (1.14), wherein S1 ∈
[1,Sum]；
(1.4) using OpenCV Runtime Library to DrawingInfoiImage gray processing, binary conversion treatment are carried out, wherein
DrawingInfoi∈Drawing；
(1.5) linear smoothing filtering, smooth DrawingInfo are carried out based on Sobel operatoriImage；
(1.6) defined variable w, h, j, k obtains DrawingInfoiWidth, Height, DrawingNumber, if w=
Width/B, h=Height/B, j=1, wherein w is image block length, and h is image block width, and j, k are counted for segmented image,
j,k∈[1,B]；
(1.7) it if j≤B, thens follow the steps (1.8), it is no to then follow the steps (1.11)；
(1.8) k=1 in step (1.6) is enabled；
(1.9) it if k≤B, thens follow the steps (1.10), it is no to then follow the steps (1.12)；
(1.10) it with coordinate { w*j, w* (j+1), h*k, h* (k+1) } for four vertex segmented images of rectangular shaped rim, enables
BlockNumber=(j-1) * B+k obtains image block Block, Block={ DrawingNumber, BlockNumber },
DrawingBlock=DrawingBlock ∪ Block；
(1.11) S1=S1+1 is enabled；
(1.12) j=j+1 is enabled；
(1.13) k=k+1 is enabled；
(1.14) pretreatment line drawing block data set DrawingBlock is obtained.
3. a kind of building element extracting method based on Faster-RCNN model according to claim 1, feature exist
In the step (2) is middle to be identified using building element of the image block data collection DrawingBlock training based on Faster-RCNN
Model DrawingModel's method particularly includes:
(2.1) definition XML file data set DrawingXML, variable temp, variable Tr, variable V a, variable Te, training dataset
TrainBlcok, validation data set ValidationBlcok, test data set TestBlcok, XML file TrainXML, XML
File ValidationXML, XML file TestXML, csv file TrainCSV, csv file ValidationCSV, csv file
TestCSV, counter S2, enables S2=1, for traversing DrawingBlock；
(2.2) it calculates image block sum and is assigned to BlockSum；
(2.3) if S2≤BlockSum, then follow the steps (2.4), it is no to then follow the steps (2.6), wherein S2 ∈ [1,
BlockSum]；
(2.4) to BlockiBuilding element mark is carried out using LabelImg tool, obtains markup information file xmli, enable
DrawingXML=DrawingXML ∪ xmli, xmliStorage form is XML file；
(2.5) S2=S2+1 is enabled；
(2.6) temp=BlockSum/ (train+validation+test), Tr=train*temp, Va=are set
Validation*temp, Te=test*temp, according to Tr, Va, Te is right respectively by DrawingBlock, DrawingXML at random
It should divide, obtain TrainBlcok={ Block1,Block2,…,BlockTr, ValidationBlcok={ Block1,
Block2,…,BlockVa, TestBlcok={ Block1,Block2,…,BlockTe, TrainXML={ xml1,
xml2,…,xmlTr, ValidationXML={ xml1,xml2,…,xmlVa, TestXML={ xml1,xml2,…,xmlTe}；
(2.7) file TrainXML, ValidationXML, TestXML are converted to data file TrainCSV respectively,
ValidationCSV, TestCSV, file layout CSV；
(2.8) use TensorFlow data file tool by architectural drawing tile data collection TrainBlcok and TrainCSV system
It is made data file Train.record, ValidationBlcok and ValidationCSV are fabricated to data file
Validation.record, TestBlcok and TestCSV are fabricated to data file Test.record；
(2.9) training Faster-RCNN model, being provided with model parameter non-maxima suppression threshold value is 0.7, and candidate frame is maximum
Quantity is 300, and initial alignment loses weight 2.0, and Classification Loss weight is 1.0, maximum number of iterations 200000, loss function choosing
Select softmax loss；
(2.10) the building element identification model DrawingModel based on Faster-RCNN model is obtained.
4. a kind of building element extracting method based on Faster-RCNN model according to claim 1, feature exist
In obtaining building element data set Result's in the step (3) method particularly includes:
(3.1) architectural drawing sum DSum, drawing set D={ drawing to be identified are defined1,drawing2,…,
drawingDSum, building element number of species ClassSum, building element type set Classes={ class1,
class2,…,classClassSum, locating rectangle frame coordinate xmin, xmax, ymin, ymax, building element posting sum
BoxSum, building element posting coordinate set DetectionBoxes={ { xmin1,xmax1,ymin1,ymax1},{xmin2,
xmax2,ymin2,ymax2},..,{xminBoxSum,xmaxBoxSum,yminBoxSum,ymaxBoxSum, building element posting is built
Build Component Category DetectionClasses, building element posting category score DetectionScores, building element identification
Results set Result={ { DetectionBoxes1,DetectionClasses1,DetectionScores1},
{DetectionBoxes2,DetectionClasses2,DetectionScores2},…,
{DetectionBoxesDrawingSum,DetectionClassesDrawingSum,DetectionScoresDrawingSum, counter
S3 enables S3=1；
(3.2) it calculates architectural drawing sum to be identified and is assigned to Dsum, building element kind class set is obtained by " standard for architectural drawing "
Close Classes, building element number of species ClassSum；
(3.3) no to then follow the steps (3.6) if S3≤Dsum is thened follow the steps (3.4), wherein [1, Dsum] S3 ∈；
(3.4) using building element identification model DrawingModel to drawing drawingiBuilding element identification is carried out, is known
Other result { DetectionBoxesi,DetectionClassesi,DetectionScoresi, Result=Result ∪
{DetectionBoxesi,DetectionClassesi,DetectionScoresi}；
(3.5) S3=S3+1；
(3.6) building element data set Result is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810677252.6A CN109002841B (en) | 2018-06-27 | 2018-06-27 | Building component extraction method based on fast-RCNN model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810677252.6A CN109002841B (en) | 2018-06-27 | 2018-06-27 | Building component extraction method based on fast-RCNN model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109002841A true CN109002841A (en) | 2018-12-14 |
| CN109002841B CN109002841B (en) | 2021-11-12 |
Family
ID=64601269
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810677252.6A Active CN109002841B (en) | 2018-06-27 | 2018-06-27 | Building component extraction method based on fast-RCNN model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109002841B (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119718A (en) * | 2019-05-15 | 2019-08-13 | 燕山大学 | A kind of overboard detection and Survivable Control System based on deep learning |
| CN110414551A (en) * | 2019-06-14 | 2019-11-05 | 田洪涛 | A kind of method and system classified automatically based on RCNN network to medical instrument |
| CN110598634A (en) * | 2019-09-12 | 2019-12-20 | 山东文多网络科技有限公司 | Machine room sketch identification method and device based on graph example library |
| CN110781888A (en) * | 2019-10-25 | 2020-02-11 | 北京字节跳动网络技术有限公司 | Method and device for regressing screen in video picture, readable medium and electronic equipment |
| 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 |
| CN111476801A (en) * | 2020-03-31 | 2020-07-31 | 万翼科技有限公司 | Image segmentation method, electronic equipment and related product |
| CN111652251A (en) * | 2020-06-09 | 2020-09-11 | 星际空间（天津）科技发展有限公司 | Method and device for building remote sensing image building feature extraction model and storage medium |
| CN111652250A (en) * | 2020-06-09 | 2020-09-11 | 星际空间（天津）科技发展有限公司 | Remote sensing image building extraction method and device based on polygon and storage medium |
| CN111832447A (en) * | 2020-06-30 | 2020-10-27 | 万翼科技有限公司 | Building drawing component identification method, electronic equipment and related product |
| CN111832437A (en) * | 2020-06-24 | 2020-10-27 | 万翼科技有限公司 | Building drawing identification method, electronic equipment and related product |
| CN111914612A (en) * | 2020-05-21 | 2020-11-10 | 淮阴工学院 | Construction graph primitive self-adaptive identification method based on improved convolutional neural network |
| CN112036268A (en) * | 2020-08-14 | 2020-12-04 | 万翼科技有限公司 | Component identification method and related device |
| CN112241565A (en) * | 2020-10-27 | 2021-01-19 | 万翼科技有限公司 | Modeling method and related device |
| CN112733735A (en) * | 2021-01-13 | 2021-04-30 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by machine learning |
| CN113392761A (en) * | 2021-06-15 | 2021-09-14 | 万翼科技有限公司 | Component identification method, device, equipment and storage medium |
| CN116109899A (en) * | 2022-12-14 | 2023-05-12 | 内蒙古建筑职业技术学院(内蒙古自治区建筑职工培训中心) | Ancient architecture repairing method, system, computer equipment and storage medium |
| JP2023076396A (en) * | 2021-11-22 | 2023-06-01 | コリア プラットフォーム サービステクノロジー株式会社 | Deep learning frame work application database server classifying structure name of traditional house drawing and method thereof |
| CN116109899B (en) * | 2022-12-14 | 2024-07-09 | 内蒙古建筑职业技术学院(内蒙古自治区建筑职工培训中心) | Ancient architecture repairing method, system, computer equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017097853A (en) * | 2015-11-18 | 2017-06-01 | 同方威視技術股▲フン▼有限公司 | Inspection method for cargo and its system |
| CN107423760A (en) * | 2017-07-21 | 2017-12-01 | 西安电子科技大学 | Based on pre-segmentation and the deep learning object detection method returned |
| CN107633267A (en) * | 2017-09-22 | 2018-01-26 | 西南交通大学 | A kind of high iron catenary support meanss wrist-arm connecting piece fastener recognition detection method |
| CN107720552A (en) * | 2017-10-16 | 2018-02-23 | 西华大学 | A kind of assembled architecture intelligence hanging method based on computer machine vision |
-
2018
- 2018-06-27 CN CN201810677252.6A patent/CN109002841B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017097853A (en) * | 2015-11-18 | 2017-06-01 | 同方威視技術股▲フン▼有限公司 | Inspection method for cargo and its system |
| CN107423760A (en) * | 2017-07-21 | 2017-12-01 | 西安电子科技大学 | Based on pre-segmentation and the deep learning object detection method returned |
| CN107633267A (en) * | 2017-09-22 | 2018-01-26 | 西南交通大学 | A kind of high iron catenary support meanss wrist-arm connecting piece fastener recognition detection method |
| CN107720552A (en) * | 2017-10-16 | 2018-02-23 | 西华大学 | A kind of assembled architecture intelligence hanging method based on computer machine vision |
Non-Patent Citations (1)
| Title |
|---|
| 雷兆明等: "基于改进RBF神经网络的钢构件质量预测研究", 《自动化与仪表》 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119718A (en) * | 2019-05-15 | 2019-08-13 | 燕山大学 | A kind of overboard detection and Survivable Control System based on deep learning |
| CN110414551A (en) * | 2019-06-14 | 2019-11-05 | 田洪涛 | A kind of method and system classified automatically based on RCNN network to medical instrument |
| CN110598634B (en) * | 2019-09-12 | 2020-08-07 | 山东文多网络科技有限公司 | Machine room sketch identification method and device based on graph example library |
| CN110598634A (en) * | 2019-09-12 | 2019-12-20 | 山东文多网络科技有限公司 | Machine room sketch identification method and device based on graph example library |
| CN110781888A (en) * | 2019-10-25 | 2020-02-11 | 北京字节跳动网络技术有限公司 | Method and device for regressing screen in video picture, readable medium and electronic equipment |
| 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 |
| CN111476801A (en) * | 2020-03-31 | 2020-07-31 | 万翼科技有限公司 | Image segmentation method, electronic equipment and related product |
| CN111476801B (en) * | 2020-03-31 | 2023-04-18 | 万翼科技有限公司 | Image segmentation method, electronic equipment and related product |
| CN111914612A (en) * | 2020-05-21 | 2020-11-10 | 淮阴工学院 | Construction graph primitive self-adaptive identification method based on improved convolutional neural network |
| CN111914612B (en) * | 2020-05-21 | 2024-03-01 | 淮阴工学院 | Construction graphic primitive self-adaptive identification method based on improved convolutional neural network |
| CN111652251A (en) * | 2020-06-09 | 2020-09-11 | 星际空间（天津）科技发展有限公司 | Method and device for building remote sensing image building feature extraction model and storage medium |
| CN111652250A (en) * | 2020-06-09 | 2020-09-11 | 星际空间（天津）科技发展有限公司 | Remote sensing image building extraction method and device based on polygon and storage medium |
| CN111652250B (en) * | 2020-06-09 | 2023-05-26 | 星际空间（天津）科技发展有限公司 | Remote sensing image building extraction method and device based on polygons and storage medium |
| CN111652251B (en) * | 2020-06-09 | 2023-06-27 | 星际空间（天津）科技发展有限公司 | Remote sensing image building feature extraction model construction method, device and storage medium |
| CN111832437A (en) * | 2020-06-24 | 2020-10-27 | 万翼科技有限公司 | Building drawing identification method, electronic equipment and related product |
| CN111832437B (en) * | 2020-06-24 | 2024-03-01 | 万翼科技有限公司 | Building drawing identification method, electronic equipment and related products |
| CN111832447B (en) * | 2020-06-30 | 2023-01-24 | 万翼科技有限公司 | Building drawing component identification method, electronic equipment and related product |
| CN111832447A (en) * | 2020-06-30 | 2020-10-27 | 万翼科技有限公司 | Building drawing component identification method, electronic equipment and related product |
| CN112036268B (en) * | 2020-08-14 | 2022-11-18 | 万翼科技有限公司 | Component identification method and related device |
| CN112036268A (en) * | 2020-08-14 | 2020-12-04 | 万翼科技有限公司 | Component identification method and related device |
| CN112241565A (en) * | 2020-10-27 | 2021-01-19 | 万翼科技有限公司 | Modeling method and related device |
| CN112733735A (en) * | 2021-01-13 | 2021-04-30 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by machine learning |
| CN112733735B (en) * | 2021-01-13 | 2024-04-09 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by adopting machine learning |
| CN113392761A (en) * | 2021-06-15 | 2021-09-14 | 万翼科技有限公司 | Component identification method, device, equipment and storage medium |
| JP2023076396A (en) * | 2021-11-22 | 2023-06-01 | コリア プラットフォーム サービステクノロジー株式会社 | Deep learning frame work application database server classifying structure name of traditional house drawing and method thereof |
| JP7414212B2 (en) | 2021-11-22 | 2024-01-16 | コリア プラットフォーム サービステクノロジー株式会社 | Deep learning framework application database server and method for classifying structural names of traditional house drawings |
| CN116109899A (en) * | 2022-12-14 | 2023-05-12 | 内蒙古建筑职业技术学院(内蒙古自治区建筑职工培训中心) | Ancient architecture repairing method, system, computer equipment and storage medium |
| CN116109899B (en) * | 2022-12-14 | 2024-07-09 | 内蒙古建筑职业技术学院(内蒙古自治区建筑职工培训中心) | Ancient architecture repairing method, system, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109002841B (en) | 2021-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109002841A (en) | A kind of building element extracting method based on Faster-RCNN model | |
| Zhu et al. | Vision-based defects detection for bridges using transfer learning and convolutional neural networks | |
| US20210390458A1 (en) | Data analytics methods for spatial data, and related systems and devices | |
| CN108734694A (en) | Thyroid tumors ultrasonoscopy automatic identifying method based on faster r-cnn | |
| CN110390275A (en) | A kind of gesture classification method based on transfer learning | |
| CN111460927B (en) | Method for extracting structured information of house property evidence image | |
| CN112308826B (en) | Bridge structure surface defect detection method based on convolutional neural network | |
| Liu et al. | Study of shrimp recognition methods using smart networks | |
| CN111914612B (en) | Construction graphic primitive self-adaptive identification method based on improved convolutional neural network | |
| CN105760877A (en) | Wool and cashmere identification algorithm based on gray level co-occurrence matrix model | |
| CN111914613B (en) | Multi-target tracking and facial feature information recognition method | |
| Bhagat et al. | WheatNet-lite: A novel light weight network for wheat head detection | |
| CN110889437B (en) | Image processing method and device, electronic equipment and storage medium | |
| CN109685065A (en) | Printed page analysis method, the system of paper automatic content classification | |
| Hoang et al. | A novel approach for detection of pavement crack and sealed crack using image processing and salp swarm algorithm optimized machine learning | |
| CN102722578B (en) | Unsupervised cluster characteristic selection method based on Laplace regularization | |
| CN110334656A (en) | Multi-source Remote Sensing Images Clean water withdraw method and device based on information source probability weight | |
| CN109389050B (en) | Method for identifying connection relation of flow chart | |
| CN104504381A (en) | Non-rigid target detection method and system thereof | |
| Ouadiay et al. | Simultaneous object detection and localization using convolutional neural networks | |
| ElAlami | Unsupervised image retrieval framework based on rule base system | |
| CN104463091A (en) | Face image recognition method based on LGBP feature subvectors of image | |
| CN114299394A (en) | Intelligent interpretation method for remote sensing image | |
| Lin et al. | Radical-based extract and recognition networks for Oracle character recognition | |
| CN105631451A (en) | Plant leave identification method based on android system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20181214 Assignee: Suzhou Hongtu Intelligent Technology Co.,Ltd. Assignor: HUAIYIN INSTITUTE OF TECHNOLOGY Contract record no.: X2021980014034 Denomination of invention: A building component extraction method based on fast RCNN model Granted publication date: 20211112 License type: Common License Record date: 20211208 |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231228 Address after: Room 306, Building B3, Huai'an Smart Valley, No. 19 Meigao Road, Economic and Technological Development Zone, Huai'an City, Jiangsu Province, 223005 Patentee after: Huai'an Yijian Zhidao Technology Co.,Ltd. Address before: 223005 Jiangsu Huaian economic and Technological Development Zone, 1 East Road. Patentee before: HUAIYIN INSTITUTE OF TECHNOLOGY |
|
| TR01 | Transfer of patent right |