CN109472220A - A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN - Google Patents
A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN Download PDFInfo
- Publication number
- CN109472220A CN109472220A CN201811237280.2A CN201811237280A CN109472220A CN 109472220 A CN109472220 A CN 109472220A CN 201811237280 A CN201811237280 A CN 201811237280A CN 109472220 A CN109472220 A CN 109472220A
- Authority
- CN
- China
- Prior art keywords
- faster
- cnn
- network module
- image data
- substation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- 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
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (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 substation's worker safety helmet detection method and its system based on Faster R-CNN are worn and the worker without safe wearing cap, realization security alarm function by depth learning technology identification.The present invention is input to RPN network module by will acquire image data and is used to handle image data and generates feature candidate region rectangle, image data is trained by Faster R-CNN network module, image data after training is trained into RPN network again, and updates the parameter of RPN network module；Faster R-CNN network module data are finely tuned with the output data of RPN network module, determine the common parameter of Faster R-CNN network module, and update the parameter of Faster R-CNN network module；Judge whether worker takes safety cap according to the image data after training.It is compared with the traditional method, detection efficiency of the invention and detection accuracy are all greatly improved.
Description
Technical field
The present invention relates to field of image detection, more particularly, to a kind of substation worker based on Faster R-CNN
Safety cap detection method and its system.
Background technique
For a long time, in the problem of the generally existing awareness of safety weakness in construction area, especially weary foundation protection sets operating personnel
The use consciousness for applying such as safety cap, considerably increases operating risk.In the existing method in safety cap detection, R-CNN is lacked
Point is more obvious, and predominantly 1. training steps are cumbersome (trim network+training SVM+ training bbox)；2. training, test speed are slow；
3 training take up space.And existing method is based primarily upon Face datection, and there are resolution fluctuations greatly, vulnerable to environmental impact issues.
Summary of the invention
The present invention is that resolution described in the above-mentioned prior art is overcome to fluctuate big, defect easily affected by environment, provides one
Substation worker safety helmet detection method and its system of the kind based on Faster R-CNN.
The present invention is directed to solve above-mentioned technical problem at least to a certain extent.
In order to solve the above technical problems, technical scheme is as follows: a kind of substation based on Faster R-CNN
Worker safety helmet detection method, comprising the following steps:
S1: image data is acquired by image capture module；
S2: image data being input in RPN network module and is trained, and obtains the candidate region of image；
S3: the data in image candidate region are input in Faster R-CNN network module and are trained, after being trained
Image data；
S4: the image data after the training in S3 is trained into RPN network again, and updates the parameter of RPN network module；
S5: Faster R-CNN network module data are finely tuned with the output data of RPN network module, determine Faster R-CNN net
The common parameter of network module, and update the parameter of Faster R-CNN network module.
S6: judge whether worker takes safety cap according to the image data after training.
Preferably, image data size is fixed in the S1.
Preferably, image data is input in RPN network module and is trained by the S2 specifically:
S21: image data progress convolution algorithm is obtained into prospect anchor foreground anchors and bounding box returns
The offset of bounding box regression, and suggestion amount proposals is calculated；
S22: the suggestion amount proposals that pond layer utilization is found out extracts suggestion feature proposal from characteristic pattern
feature；
S23: suggestion feature is input to subsequent full connection and softmax network as classification classification.
A kind of system of substation's worker safety helmet detection method based on Faster R-CNN, including Image Acquisition mould
Block, RPN network module, Faster R-CNN network module；
Preferably, described image acquisition module is using image data in high-definition camera acquisition substation.
Preferably, RPN network module is stated for handling image data and generating feature candidate region rectangle, and utilizes output
Image data adjust Faster R-CNN parameter.
Compared with prior art, the beneficial effect of technical solution of the present invention is: the present invention passes through Faster R-CNN model
It is detected, the calculating having are completed in GPU more completely, substantially increase the speed of service, detection accuracy is also centainly mentioned
It rises.
Detailed description of the invention
Fig. 1 is flow chart of the invention.
Fig. 2 is system block diagram of the invention.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent；
The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
Embodiment 1
Fig. 2 is system block diagram of the invention, it includes image capture module, RPN network module, Faster R-CNN network mould
Block；Wherein image capture module is for acquiring image data image data；RPN network module is for handling image data and generating
Feature candidate region rectangle；Faster R-CNN network module carries out parameter update using the output data of RPN network and exports
Testing result.Its work step is as shown in Figure 1:
S1: image data is acquired by image capture module；
S2: image data being input in RPN network module and is trained, and obtains the candidate region of image；
S3: the data in image candidate region are input in Faster R-CNN network module and are trained, after being trained
Image data；
S4: the image data after the training in S3 is trained into RPN network again, and updates the parameter of RPN network module；
S5: Faster R-CNN network module data are finely tuned with the output data of RPN network module, determine Faster R-CNN net
The common parameter of network module, and update the parameter of Faster R-CNN network module；
S6: judge whether worker takes safety cap according to the image data after training.
In the specific implementation process, image data is acquired by high-definition camera, input an image into first in detection model
It is first scaled to the picture of unified size dimension, then into RPN network；Image data convolutional layer is input to first to pass through
Having a size of 3 × 3 convolution algorithm, wherein convolutional layer includes 13 convolutional layers, 4 pond layers and 13 active coatings.Again by image
Data carry out convolution algorithm and obtain prospect anchor foreground anchors and bounding box recurrence bounding box
The offset of regression, and suggestion amount proposals is calculated；The suggestion amount proposals that pond layer utilization is found out,
Suggestion feature proposal feature is extracted from characteristic pattern；Suggestion feature is input to subsequent full connection and softmax
Network obtains the candidate region image of image, the data of candidate region are input to Faster as classification classification
It is trained in R-CNN network module, the image data after being trained；Image data after training in S3 is instructed again
Practice RPN network, and updates the parameter of RPN network module；Faster R-CNN network is finely tuned with the output data of RPN network module
Module data, determines the common parameter of Faster R-CNN network module, and updates the parameter of Faster R-CNN network module.
Judge whether worker takes safety cap according to the image data after training.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (7)
1. a kind of substation's worker safety helmet detection method based on Faster R-CNN, it is characterised in that: the following steps are included:
S1: image data is acquired by image capture module；
S2: image data being input in RPN network module and is trained, and obtains the candidate region of image；
S3: the data in image candidate region are input in Faster R-CNN network module and are trained, after being trained
Image data；
S4: the image data after the training in S3 is trained into RPN network again, and updates the parameter of RPN network module；
S5: Faster R-CNN network module data are finely tuned with the output data of RPN network module, determine Faster R-CNN net
The common parameter of network module, and update the parameter of Faster R-CNN network module；
S6: judge whether worker takes safety cap according to the image data after training.
2. a kind of substation's worker safety helmet detection method based on Faster R-CNN according to claim 1, special
Sign is: image capture module is using high-definition camera.
3. a kind of substation's worker safety helmet detection method based on Faster R-CNN according to claim 1, special
Sign is: the image data size being input in RPN network module is fixed.
4. a kind of substation's worker safety helmet detection method based on Faster R-CNN according to claim 1, special
Sign is: image data is input in RPN network module and is trained by the S2 specifically:
S21: image data progress convolution algorithm is obtained into prospect anchor foreground anchors and bounding box returns
The offset of bounding box regression, and suggestion amount proposals is calculated；
S22: the suggestion amount proposals that pond layer utilization is found out extracts suggestion feature proposal from characteristic pattern
feature；
S23: suggestion feature is input to subsequent full connection and softmax network as classification classification.
5. a kind of system of substation's worker safety helmet detection method based on Faster R-CNN, it is characterised in that: including figure
As acquisition module, RPN network module, Faster R-CNN network module.
6. a kind of substation's worker safety helmet detection method based on Faster R-CNN according to claim 1 is
System, it is characterised in that: described image acquisition module is using image data in high-definition camera acquisition substation.
7. a kind of substation's worker safety helmet detection method based on Faster R-CNN according to claim 1 is
System, it is characterised in that: the RPN network module utilizes defeated for handling image data and generating feature candidate region rectangle
Image data out adjusts Faster R-CNN parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811237280.2A CN109472220A (en) | 2018-10-23 | 2018-10-23 | A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811237280.2A CN109472220A (en) | 2018-10-23 | 2018-10-23 | A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109472220A true CN109472220A (en) | 2019-03-15 |
Family
ID=65664039
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811237280.2A Pending CN109472220A (en) | 2018-10-23 | 2018-10-23 | A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109472220A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110309719A (en) * | 2019-05-27 | 2019-10-08 | 安徽继远软件有限公司 | A kind of electric network operation personnel safety cap wears management control method and system |
| CN110399905A (en) * | 2019-07-03 | 2019-11-01 | 常州大学 | The detection and description method of safety cap wear condition in scene of constructing |
| CN110992349A (en) * | 2019-12-11 | 2020-04-10 | 南京航空航天大学 | Underground pipeline abnormity automatic positioning and identification method based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778852A (en) * | 2016-12-07 | 2017-05-31 | 中国科学院信息工程研究所 | A kind of picture material recognition methods for correcting erroneous judgement |
| CN107274451A (en) * | 2017-05-17 | 2017-10-20 | 北京工业大学 | Isolator detecting method and device based on shared convolutional neural networks |
| CN108010060A (en) * | 2017-12-06 | 2018-05-08 | 北京小米移动软件有限公司 | Object detection method and device |
| US20180268292A1 (en) * | 2017-03-17 | 2018-09-20 | Nec Laboratories America, Inc. | Learning efficient object detection models with knowledge distillation |
-
2018
- 2018-10-23 CN CN201811237280.2A patent/CN109472220A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778852A (en) * | 2016-12-07 | 2017-05-31 | 中国科学院信息工程研究所 | A kind of picture material recognition methods for correcting erroneous judgement |
| US20180268292A1 (en) * | 2017-03-17 | 2018-09-20 | Nec Laboratories America, Inc. | Learning efficient object detection models with knowledge distillation |
| CN107274451A (en) * | 2017-05-17 | 2017-10-20 | 北京工业大学 | Isolator detecting method and device based on shared convolutional neural networks |
| CN108010060A (en) * | 2017-12-06 | 2018-05-08 | 北京小米移动软件有限公司 | Object detection method and device |
Non-Patent Citations (4)
| Title |
|---|
| OUTTHINKER: "Faster rcnn代码理解（2）", 《HTTPS://WWW.CNBLOGS.COM/ZF-BLOG/P/7286405.HTML》 * |
| WAKOJOSIN: "faster rcnn中核心部分RPN网络的整理与理解", 《HTTPS://BLOG.CSDN.NET/WAKOJOSIN/ARTICLE/DETAILS/79363224》 * |
| 欧西特: "Faster-Rcnn中RCN（Region Proposal Network）的理解", 《HTTPS://BLOG.CSDN.NET/MLLEARNERTJ/ARTICLE/DETAILS/53709766》 * |
| 王忠玉: "智能视频监控下的安全帽佩戴检测系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110309719A (en) * | 2019-05-27 | 2019-10-08 | 安徽继远软件有限公司 | A kind of electric network operation personnel safety cap wears management control method and system |
| CN110399905A (en) * | 2019-07-03 | 2019-11-01 | 常州大学 | The detection and description method of safety cap wear condition in scene of constructing |
| CN110992349A (en) * | 2019-12-11 | 2020-04-10 | 南京航空航天大学 | Underground pipeline abnormity automatic positioning and identification method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101577812B (en) | Method and system for post monitoring | |
| CN111047818A (en) | Forest fire early warning system based on video image | |
| CN103093192B (en) | The recognition methods that high voltage transmission line is waved | |
| CN109472220A (en) | A kind of substation's worker safety helmet detection method and its system based on Faster R-CNN | |
| CN106295567A (en) | The localization method of a kind of key point and terminal | |
| CN109002801A (en) | A kind of face occlusion detection method and system based on video monitoring | |
| CN103473564B (en) | A kind of obverse face detection method based on sensitizing range | |
| CN113903081A (en) | Visual identification artificial intelligence alarm method and device for images of hydraulic power plant | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN104966372A (en) | Multi-data fusion forest fire intelligent recognition system and method | |
| CN103412296B (en) | Automatically method of power tower is extracted in random laser point cloud data | |
| CN109522838A (en) | A kind of safety cap image recognition algorithm based on width study | |
| CN103390151B (en) | Method for detecting human face and device | |
| CN108229524A (en) | A kind of chimney and condensing tower detection method based on remote sensing images | |
| CN104850841A (en) | Elder abnormal behavior monitoring method with combination of RFID and video identification | |
| CN103425959A (en) | Flame video detection method for identifying fire hazard | |
| CN103049748B (en) | Behavior monitoring method and device | |
| CN116935473A (en) | Real-time detection method and system for wearing safety helmet based on improved YOLO v7 under complex background | |
| CN108921826A (en) | The transmission line of electricity that super-pixel segmentation is combined with deep learning invades object detecting method | |
| CN113128693A (en) | Information processing method, device, equipment and storage medium | |
| CN104573658A (en) | Blind drive detection method based on support vector machine | |
| CN108108696A (en) | A kind of safety protecting method, apparatus and system | |
| CN106096499A (en) | A kind of video image culminant star moon pattern detection method and system | |
| CN110516538A (en) | The double violation assessment method of leaving the post in prison based on deep learning target detection | |
| CN103123689B (en) | A kind of run detection method and device based on the detection of people's leg |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190315 |
|
| RJ01 | Rejection of invention patent application after publication |