CN110096943A - A kind of architecture against regulations detection system based on deep learning - Google Patents
A kind of architecture against regulations detection system based on deep learning Download PDFInfo
- Publication number
- CN110096943A CN110096943A CN201910081655.9A CN201910081655A CN110096943A CN 110096943 A CN110096943 A CN 110096943A CN 201910081655 A CN201910081655 A CN 201910081655A CN 110096943 A CN110096943 A CN 110096943A
- Authority
- CN
- China
- Prior art keywords
- layer
- module
- video
- image
- responsible
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 36
- 238000013135 deep learning Methods 0.000 title claims abstract description 12
- 238000010191 image analysis Methods 0.000 claims abstract description 21
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 238000002372 labelling Methods 0.000 claims abstract description 8
- 238000012360 testing method Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 9
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000000889 atomisation Methods 0.000 claims description 6
- 238000000034 method Methods 0.000 claims description 4
- 238000012512 characterization method Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 8
- 230000009286 beneficial effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000012545 processing Methods 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The architecture against regulations detection system based on deep learning that the invention discloses a kind of, including region labeling module, VAM Video Access Module, image analysis module and evidence obtaining uploading module；The region labeling module: it is responsible for the prewarning area and uncalibrated image of record user's calibration；The VAM Video Access Module: it is responsible for reading land resources monitor video, and video flowing is switched into single frame video image and is sent among image analysis module；Described image analysis module: it is responsible for whether there is the architecture against regulations according to video image analysis；The evidence obtaining uploading module: proof data is synthesized according to the video data of the analysis result of image analysis module and VAM Video Access Module, and is uploaded in top service device in time.Beneficial effects of the present invention are as follows: deep to use depth learning technology, efficient height；It can further be trained, be improved efficient according to user data.
Description
Technical field
The present invention relates to depth learning technology fields, more particularly to one kind to be based on deep learning architecture against regulations detection system.
Background technique
Assert that organ should judge the attribute of building using law, administrative regulation as foundation.If contrary to law, administrative regulation
And regulations violate construction project though not obtaining planning permit of construction engineering or achieving planning permit of construction engineering
The regulation of planning permission is built, and the building and structures of urban planning are seriously affected, and should regard as violating the regulations build
It builds.And at present other than the evidence obtaining of artificial scene photograph, not long-acting automatic evidence-collecting technological means.
Summary of the invention
In order to solve the deficiencies in the prior art, this application provides one kind to be based on deep learning architecture against regulations detection system.
Technical scheme is as follows:
A kind of architecture against regulations detection system based on deep learning, which is characterized in that access mould including region labeling module, video
Block, image analysis module and evidence obtaining uploading module；
The region labeling module: it is responsible for the prewarning area and uncalibrated image of record user's calibration；
The VAM Video Access Module: it is responsible for reading land resources monitor video, and video flowing is switched into single frame video image and is sent to
Among image analysis module；
Described image analysis module: it is responsible for whether there is the architecture against regulations according to video image analysis；
The evidence obtaining uploading module: it is synthesized and is demonstrate,proved according to the video data of the analysis result of image analysis module and VAM Video Access Module
According to data, and uploaded in top service device in time.
A kind of detection method of architecture against regulations detection system based on deep learning, which is characterized in that the figure
As steps are as follows for the concrete analysis of analysis module:
1): the prewarning area demarcated according to single frame video image and user divides the region for needing early warning and record area and sits
Mark will divide and need the region of early warning as detection sample；
2): building includes the training set of several building pictures, utilizes convolutional neural networks of the training set training based on Vgg-16
Contrast model；
3): atomization process is carried out to detection sample；
4): the detection sample of atomization will be gone to sequentially input trained convolutional neural networks contrast model, obtain one group of consecutive hours
Between testing result sequenceL={t i |i=1,2,3,…,n, whereint i Indicate theiThe testing result of detection sample is opened,t i ∈[0,
1],nIndicate detection sample size；
5): testing result sequence being judged, if the sum of element is greater than given threshold value in testing result sequenceT, then determining should
There are the architectures against regulations for detection zone.
A kind of detection method of architecture against regulations detection system based on deep learning, which is characterized in that the step
It is rapid 2) in the convolutional neural networks contrast model of Vgg-16 include two parallel Vgg-16 network characterization extract layers, multiple dimensioned spy
Sign combination equivalent beds and last network output layer；1 one region original image of input of input, mentions by the feature of Vgg-16
After taking layer, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector equivalent beds；2 one region to be measured of input of input
Image, after the feature extraction layer for also passing through Vgg-16, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector
Equivalent beds；Analysis On Multi-scale Features combination equivalent beds are made of one layer of full articulamentum, responsible to compare to input feature vector, then by result
Input last output layer；Output layer normalizes to result in the space of [0,1], and provides two image architecture against regulations confidences
Degree.
Technical scheme is as follows:
1) depth learning technology: using depth learning technology, efficient high；It can further be trained, be mentioned according to user data
It is high efficient.
2) solve the problems, such as candid photograph: architecture against regulations detection system can use existing land resources monitor video, without again
Camera is set up, system utilizes artificial intelligence technology, and the monitoring area set to video acceptance of the bid is analyzed, and compares and examines by front and back
It surveys and whether belongs to that there are the architectures against regulations, then the information of the architecture against regulations is recorded as needed and uploads to top service
Device.
3) powerful system function: analysis result is compared using artificial intelligence technology, is verified, detects violating the regulations build
The situation of change in region is built, then as needed records the information such as architecture against regulations place, time, while being output to illegal
Information platform.
4) excellent Products Compatibility: the product uses national standardization communication protocol and video decoding algorithm, improves and produces
Product compatibility, all major video capturing systems in the compatible country and video.
Detailed description of the invention
Schematic network structure Fig. 1 of the invention；
Analysis system structural schematic diagram Fig. 2 of the invention；
Field deployment schematic diagram Fig. 3 of the invention.
Specific embodiment
With reference to the accompanying drawings and examples, the present invention is further detailed.It should be appreciated that tool described herein
Body embodiment is used only for explaining the present invention, not of the invention in limiting.
Owned building in video is carried out using depth learning technology based on deep learning architecture against regulations detection system
It analyzes and determines, the change of building is then compared out according to network analysis result；Solve the prior art can not automatic evidence-collecting ask
Topic realizes automation, the analytical judgment of the intelligentized architecture against regulations, can carry out the automatic architecture against regulations detection of extensive area.
As shown in Figure 1, be deep neural network structural schematic diagram of the invention, it is special including two parallel Vgg-16 networks
Extract layer is levied, Analysis On Multi-scale Features combine equivalent beds and last network output layer；1 one region original image of input of input, warp
It crosses after the feature extraction layer of Vgg-16, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector equivalent beds；Input 2
An area image to be measured is inputted, after the feature extraction layer for also passing through Vgg-16, by the 4th layer, the 8th layer, the 12nd layer and the
16 layers of feature input feature vector equivalent beds；Analysis On Multi-scale Features equivalent beds are made of one layer of full articulamentum, are responsible for carrying out input feature vector
Comparison, then result is inputted to last output layer；Output layer normalizes to result in the space of [0,1], provides two images
Architecture against regulations confidence level.
It is analysis system structural schematic diagram of the invention as shown in Fig. 2, mainly includes region labeling module, video access
Module, image analysis module and evidence obtaining uploading module.
Region labeling module is mainly responsible for the prewarning area in the image and image of record user's calibration；
VAM Video Access Module by each video camera producer disclose offer SDK access component package form, sample and by amplifier into
Row solution transcoding, input GPU image analysis module are analyzed, and are carried out failure function judgement processing by sampling software, are mainly responsible for
Land resources monitor video is read, and video flowing is switched into single frame video image and is sent among image analysis module；
Image analysis module is responsible for according to video image analysis with the presence or absence of the architecture against regulations, including following steps:
Step 1: the prewarning area demarcated according to single frame video image and user divides the region for needing early warning and recording areas
Domain coordinate will divide and need the region of early warning as detection sample；
Step 2: building includes the training set of several building pictures, utilizes convolutional Neural net of the training set training based on Vgg-16
Network contrast model；
The convolutional neural networks contrast model of Vgg-16 includes two parallel Vgg-16 network characterization extract layers, Analysis On Multi-scale Features
Combine equivalent beds and last network output layer；1 one region original image of input of input, by the feature extraction of Vgg-16
After layer, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector equivalent beds；2 one administrative division map to be measured of input of input
Picture, after the feature extraction layer for also passing through Vgg-16, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector pair
Than layer；Analysis On Multi-scale Features combination equivalent beds are made of one layer of full articulamentum, are responsible for comparing input feature vector, then result is defeated
Enter last output layer；Output layer normalizes to result in the space of [0,1], and provides two image architecture against regulations confidences
Degree.
Step 3: atomization process is carried out to detection sample；
Step 4: the detection sample of atomization will be gone to sequentially input trained network contrast model, obtain the inspection of one group of continuous time
Survey result sequenceL={t i |i=1,2,3,…,n, whereint i Indicate theiThe testing result of detection sample is opened,t i ∈ [0,1],nTable
Show detection sample size；
Step 5: testing result sequence being judged, if the sum of element is greater than given threshold value in testing result sequenceT, then determine
There are the architectures against regulations for the detection zone；TFor peccancy detection threshold value given in advance, in this example,T=n/2;
Uploading module of collecting evidence synthesizes evidence number according to the analysis result of image analysis module and the video data of VAM Video Access Module
According to, and uploaded in top service device in time；
As shown in figure 3, field deployment schematic diagram of the invention, video monitoring are deployed on skyscraper, monitoring range one
Block does not have the region of the architecture against regulations, and the boxed area in region is the prewarning area of user annotation, and video data is passed by cable
It is defeated to arrive video server, then take video flowing to carry out video analysis from video server by Analysis server.
Content described in this specification embodiment is only enumerating to the way of realization of inventive concept, protection of the invention
Range should not be construed as being limited to the specific forms stated in the embodiments, and protection scope of the present invention is also and in this field skill
Art personnel conceive according to the present invention it is conceivable that equivalent technologies mean.
Claims (3)
1. a kind of architecture against regulations detection system based on deep learning, which is characterized in that accessed including region labeling module, video
Module, image analysis module and evidence obtaining uploading module；
The region labeling module: the prewarning area being responsible in the image and image of record user's calibration；
The VAM Video Access Module: it is responsible for reading land resources monitor video, and video flowing is switched into single frame video image and is sent to
Among image analysis module；
Described image analysis module: it is responsible for whether there is the architecture against regulations according to video image analysis；
The evidence obtaining uploading module: it is responsible for being closed according to the analysis result of image analysis module and the video data of VAM Video Access Module
At proof data, and uploaded in top service device in time.
2. a kind of architecture against regulations detection system based on deep learning according to claim 1, which is characterized in that the figure
As steps are as follows for the concrete analysis of analysis module:
1): the region for needing early warning is marked off according to single frame video image and the calibrated prewarning area of user, record area is sat
Mark, and will mark off and need the region of early warning as detection sample；
2): building includes the training set of several building pictures, utilizes convolutional neural networks of the training set training based on Vgg-16
Contrast model；
3): atomization process is carried out to detection sample；
4): the detection sample of atomization will be gone to sequentially input trained convolutional neural networks contrast model, obtain one group of consecutive hours
Between testing result sequenceL={t i |i=1,2,3,…,n, whereint i Indicate theiThe testing result of detection sample is opened,t i ∈[0,
1],nIndicate detection sample size；
5): testing result sequence being judged, if the sum of element is greater than given threshold value in testing result sequenceT, then determining should
There are the architectures against regulations for detection zone.
3. a kind of detection method of architecture against regulations detection system based on deep learning according to claim 2, feature
It is, the convolutional neural networks contrast model of Vgg-16 includes that two parallel Vgg-16 network characterizations extract in the step 2
Layer, Analysis On Multi-scale Features combine equivalent beds and last network output layer；1 one region original image of input of input, by Vgg-
After 16 feature extraction layer, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer feature input feature vector equivalent beds；2 input one of input
Area image to be measured, after the feature extraction layer for also passing through Vgg-16, by the 4th layer, the 8th layer, the 12nd layer and the 16th layer spy
Levy input feature vector equivalent beds；Analysis On Multi-scale Features combination equivalent beds are made of one layer of full articulamentum, responsible to carry out pair to input feature vector
Than, then result is inputted to last output layer；Output layer normalizes to result in the space of [0,1], and provides two images
Architecture against regulations confidence level.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910081655.9A CN110096943A (en) | 2019-01-28 | 2019-01-28 | A kind of architecture against regulations detection system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910081655.9A CN110096943A (en) | 2019-01-28 | 2019-01-28 | A kind of architecture against regulations detection system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110096943A true CN110096943A (en) | 2019-08-06 |
Family
ID=67443797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910081655.9A Pending CN110096943A (en) | 2019-01-28 | 2019-01-28 | A kind of architecture against regulations detection system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110096943A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563448A (en) * | 2020-04-30 | 2020-08-21 | 北京百度网讯科技有限公司 | Method and device for detecting illegal building, electronic equipment and storage medium |
| CN112102572A (en) * | 2020-09-21 | 2020-12-18 | 浙江浩腾电子科技股份有限公司 | Solar energy thing networking isolation barrier control system based on artificial intelligence |
| CN113505643A (en) * | 2021-06-07 | 2021-10-15 | 浙江大华技术股份有限公司 | Violation target detection method and related device |
| CN113537557A (en) * | 2021-06-03 | 2021-10-22 | 浙江浩腾电子科技股份有限公司 | Urban road traffic barrier damage early warning system and early warning method thereof |
| CN115497172A (en) * | 2022-11-18 | 2022-12-20 | 合肥中科类脑智能技术有限公司 | Fishing behavior detection method and device, edge processing equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107194396A (en) * | 2017-05-08 | 2017-09-22 | 武汉大学 | Method for early warning is recognized based on the specific architecture against regulations in land resources video monitoring system |
| CN108052917A (en) * | 2017-12-25 | 2018-05-18 | 东南大学 | A kind of method of the architecture against regulations automatic identification found based on new and old Temporal variation |
| CN108573276A (en) * | 2018-03-12 | 2018-09-25 | 浙江大学 | A kind of change detecting method based on high-resolution remote sensing image |
-
2019
- 2019-01-28 CN CN201910081655.9A patent/CN110096943A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107194396A (en) * | 2017-05-08 | 2017-09-22 | 武汉大学 | Method for early warning is recognized based on the specific architecture against regulations in land resources video monitoring system |
| CN108052917A (en) * | 2017-12-25 | 2018-05-18 | 东南大学 | A kind of method of the architecture against regulations automatic identification found based on new and old Temporal variation |
| CN108573276A (en) * | 2018-03-12 | 2018-09-25 | 浙江大学 | A kind of change detecting method based on high-resolution remote sensing image |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563448A (en) * | 2020-04-30 | 2020-08-21 | 北京百度网讯科技有限公司 | Method and device for detecting illegal building, electronic equipment and storage medium |
| CN111563448B (en) * | 2020-04-30 | 2023-10-31 | 北京百度网讯科技有限公司 | Method and device for detecting illegal building, electronic equipment and storage medium |
| CN112102572A (en) * | 2020-09-21 | 2020-12-18 | 浙江浩腾电子科技股份有限公司 | Solar energy thing networking isolation barrier control system based on artificial intelligence |
| CN113537557A (en) * | 2021-06-03 | 2021-10-22 | 浙江浩腾电子科技股份有限公司 | Urban road traffic barrier damage early warning system and early warning method thereof |
| CN113505643A (en) * | 2021-06-07 | 2021-10-15 | 浙江大华技术股份有限公司 | Violation target detection method and related device |
| CN115497172A (en) * | 2022-11-18 | 2022-12-20 | 合肥中科类脑智能技术有限公司 | Fishing behavior detection method and device, edge processing equipment and storage medium |
| CN115497172B (en) * | 2022-11-18 | 2023-02-17 | 合肥中科类脑智能技术有限公司 | Fishing behavior detection method and device, edge processing equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110096943A (en) | A kind of architecture against regulations detection system based on deep learning | |
| CN104166841B (en) | The quick detection recognition methods of pedestrian or vehicle is specified in a kind of video surveillance network | |
| CN101552910B (en) | Remnant detection device based on comprehensive computer vision | |
| US7822275B2 (en) | Method for detecting water regions in video | |
| CN108564052A (en) | Multi-cam dynamic human face recognition system based on MTCNN and method | |
| CN104504377B (en) | A kind of passenger on public transport degree of crowding identifying system and method | |
| CN111967393A (en) | Helmet wearing detection method based on improved YOLOv4 | |
| CN108549854A (en) | A kind of human face in-vivo detection method | |
| CN107085696A (en) | A kind of vehicle location and type identifier method based on bayonet socket image | |
| CN105959723B (en) | A kind of lip-sync detection method being combined based on machine vision and Speech processing | |
| CN110189355A (en) | Safe escape channel occupies detection method, device, electronic equipment and storage medium | |
| CN114333070A (en) | Examinee abnormal behavior detection method based on deep learning | |
| CN112232333A (en) | Real-time passenger flow thermodynamic diagram generation method in subway station | |
| Kongurgsa et al. | Real-time intrusion—detecting and alert system by image processing techniques | |
| CN113536972A (en) | Self-supervision cross-domain crowd counting method based on target domain pseudo label | |
| CN110674887A (en) | End-to-end road congestion detection algorithm based on video classification | |
| CN110490150A (en) | A kind of automatic auditing system of picture violating the regulations and method based on vehicle retrieval | |
| CN110659546A (en) | Illegal booth detection method and device | |
| CN109712284A (en) | A kind of intelligent campus management system and its method | |
| CN112614102A (en) | Vehicle detection method, terminal and computer readable storage medium thereof | |
| CN109684986A (en) | A kind of vehicle analysis method and system based on automobile detecting following | |
| CN111914649A (en) | Face recognition method and device, electronic equipment and storage medium | |
| CN111476727A (en) | Video motion enhancement method for face changing video detection | |
| CN108229421A (en) | A kind of falling from bed behavior real-time detection method based on deep video information | |
| CN113689382B (en) | Tumor postoperative survival prediction method and system based on medical images and pathological images |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190806 |