CN111598040B - Construction worker identity recognition and safety helmet wearing detection method and system - Google Patents
Construction worker identity recognition and safety helmet wearing detection method and system Download PDFInfo
- Publication number
- CN111598040B CN111598040B CN202010447288.2A CN202010447288A CN111598040B CN 111598040 B CN111598040 B CN 111598040B CN 202010447288 A CN202010447288 A CN 202010447288A CN 111598040 B CN111598040 B CN 111598040B
- Authority
- CN
- China
- Prior art keywords
- construction
- yolo
- shwd
- worker
- safety helmet
- 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.)
- Active
Links
- 238000010276 construction Methods 0.000 title claims abstract description 83
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 25
- 238000000034 method Methods 0.000 claims abstract description 25
- 238000012545 processing Methods 0.000 claims abstract description 17
- 238000013528 artificial neural network Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 238000003062 neural network model Methods 0.000 claims abstract description 11
- 238000010606 normalization Methods 0.000 claims abstract description 10
- 238000012795 verification Methods 0.000 claims abstract description 10
- 238000013135 deep learning Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 10
- 238000013527 convolutional neural network Methods 0.000 claims description 9
- 230000001815 facial effect Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 3
- 230000005484 gravity Effects 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 2
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 238000005286 illumination Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 208000019462 Occupational injury Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/30—Individual registration on entry or exit not involving the use of a pass
- G07C9/32—Individual registration on entry or exit not involving the use of a pass in combination with an identity check
- G07C9/37—Individual registration on entry or exit not involving the use of a pass in combination with an identity check using biometric data, e.g. fingerprints, iris scans or voice recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Human Computer Interaction (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method and a system for identifying the identity of a construction worker and detecting the wearing of a safety helmet, comprising the following steps: collecting the identity information of a construction worker and a worker image monitored by a real-name entrance; adding labels to the images in the public dataset SHWD, and carrying out normalization processing; building a system architecture based on a YOLO V3 and Light cnn-29 neural network algorithm; training YOLO v3 on the common dataset SHWD, inputting the marked image, detecting the helmet and face parts; training Light CNN-29; testing the recognition precision and recall rate under various construction environments, and correcting; and (3) carrying out identity verification on personnel entering the construction site by using the corrected neural network model, and carrying out safety helmet detection on the personnel entering the construction site. According to the invention, by means of the deep learning algorithm, the identity of the person entering the construction site is verified, the safety helmet detection is carried out on the person entering the construction site, the person not in the project and the person not wearing the safety helmet are prevented from entering the construction site, and the safety management capability of the construction site is effectively improved.
Description
Technical Field
The invention relates to the technical field of computer AI, in particular to a method for identifying the identity of a construction worker and detecting the wearing of a safety helmet.
Background
Safety has been a concern for the construction industry for decades. Occupational injuries in the construction field are mainly derived from unsafe behaviour, unsafe environments and unsafe mechanical tools, wherein unsafe environments and behaviour are considered to be the main causes of building accidents. At present, unsafe behaviors and environments in the building field are mainly checked by manpower detection on a construction site, and due to heavy construction engineering tasks, the manpower resources capable of being allocated to safety detection are limited. With the development of computer and information technology, more and more intelligent monitoring and detection technologies are applied to safety detection in the building field.
At construction sites, inspection of workers 'safety activities and safety protection settings often depends on the strength of the security officer's inspection. The identity of the personnel entering the construction site is confirmed and the safety helmet is checked, the personnel entering the construction site is mainly checked by manpower on site, the mobility of the construction workers is high, the personnel can not be confirmed by the manpower check, and the safety helmet of the construction workers is often carelessly checked.
Disclosure of Invention
According to the method for identifying the identity of the construction worker and detecting the wearing of the safety helmet, which is provided by the invention, the computer vision is cited into the construction safety detection, and the artificial intelligence technology is used for replacing the manual detection in the traditional construction safety.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a construction worker identity recognition and safety helmet wearing detection method comprises the following steps:
The method comprises the following steps:
downloading a public dataset SHWD, and collecting building worker identity information and worker images monitored by a real-name system entrance;
adding labels to the images in the public dataset SHWD, naming and storing the collected images according to the identity information of workers, and carrying out normalization processing;
Building a system architecture based on a YOLO V3 and Light cnn-29 neural network algorithm;
Training YOLO v3 on common dataset SHWD and inputting the labeled SHWD dataset image to YOLO v3;
Taking the image detected as a face area by the Yolo v3 as a face recognition training set to train the Light CNN-29;
detecting the performance of the YOLO v3 and Light cnn-29 neural network systems respectively, testing the recognition accuracy and recall rate under various construction environments, and correcting;
Further, downloading a public dataset SHWD, and collecting construction worker identity information and worker images monitored by a real-name system entrance; comprising the following steps:
First, collecting project department management personnel and construction personnel of each team, self-photographing the front face under the condition of wearing the safety helmet, and using photo information to: the name, the team and the age are put into a database;
Secondly, the face image information of the construction workers is intercepted through a camera arranged at the entrance of a real-name system channel of the construction site, so that a data set is enlarged, and the data set is classified and stored according to the actual site condition;
finally, the common data set SHWD, which is the common data set for the helmet wear detection, is downloaded.
Further, adding a label to the image in the public dataset SHWD, naming and storing the collected image according to the worker identity information, and performing normalization processing, including:
Standardizing the collected images to 416 x 416 size, wherein the maximized image content comprises faces and helmets;
Further comprises:
Labeling the classified samples and the data samples labeled by the position boundaries; for the collected dataset, the facial features and helmet features of the worker are tagged using LabelImg, and these annotations are saved as XML files in PASCALVOC format for Python reading.
Further, the introducing HAAR CASCADE cascades the classifier, including:
The HAAR CASCADE cascade classifier locates the position of the face through an expandable window, and takes the highest score from the prediction window to locate the face.
On the other hand, the invention also discloses a system for identifying the identity of a construction worker and detecting the wearing of a safety helmet, which comprises the following units:
the data collection unit is used for respectively collecting the identity information of the construction workers and the corresponding wearing image information of the safety helmet;
the data processing unit is used for adding labels to the collected image information, storing the image information and carrying out normalization processing;
The model training unit is used for inputting the image added with the label into the convolutional neural network to train the neural network model;
The model correction unit is used for detecting the performance of the trained convolutional neural network of the trained training neural network model, testing the recognition precision and recall rate under various environments and correcting;
The data detection unit is used for verifying the identity of the personnel entering the construction site by utilizing the corrected neural network model and detecting the safety helmet of the personnel entering the construction site.
According to the technical scheme, the method and the system for identifying the identity of the construction worker and detecting the wearing of the safety helmet have the following beneficial effects:
In summary, according to the deep learning-based construction worker identity recognition and helmet wearing detection method, through collecting face images of construction workers, adopting a supervised learning mode, utilizing a Python programming language and a YOLO v3 network frame to extract face and helmet characteristics, learning image information and testing, and therefore identity verification of workers and helmet wearing detection are achieved. Through a large amount of data acquisition and training, the method has the advantages that the accuracy rate of the identity verification of workers and the detection and recognition rate of safety helmets are both over 90 percent.
Compared with the traditional face recognition method, the method can recognize the face of a building worker in various weather and illumination environments, and can reach more than 85% of recognition paths under the condition of shielding, and the algorithm has strong robustness; compared with a common safety helmet detection algorithm, the safety helmet detection method and the safety helmet detection system can expand the function of worker identity verification on the basis of safety helmet identification, and can realize real-time monitoring.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a schematic diagram of the method of the present invention;
Fig. 3 is a view showing a YOLO v3 network configuration of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention.
As shown in fig. 1, the method for identifying the identity of a construction worker and detecting the wearing of a helmet according to the embodiment includes:
The method comprises the following steps:
downloading a public dataset SHWD, and collecting building worker identity information and worker images monitored by a real-name system entrance;
adding labels to the images in the public dataset SHWD, naming and storing the collected images according to the identity information of workers, and carrying out normalization processing;
Building a system architecture based on a YOLO V3 and Light cnn-29 neural network algorithm;
Training YOLO v3 on common dataset SHWD and inputting the labeled SHWD dataset image to YOLO v3;
Taking the image detected as a face area by the Yolo v3 as a face recognition training set to train the Light CNN-29;
detecting the performance of the YOLO v3 and Light cnn-29 neural network systems respectively, testing the recognition accuracy and recall rate under various construction environments, and correcting;
and (3) carrying out identity verification on personnel entering the construction site by using the corrected neural network model, and carrying out safety helmet detection on the personnel entering the construction site.
The following is a specific description:
Safety management for construction sites, see fig. 2, comprising the steps of:
Step 1: downloading a public dataset SHWD, and collecting building worker identity information and worker images monitored by a real-name system entrance;
The existing public data set does not support the training requirement of the invention, and a worker face information and a safety helmet library are built by the worker. First, we collect project department manager and team constructors, self-timer the face with the helmet on, and use photo information to: name + team + age (e.g., li Gang constructor 31) into the database. Secondly, in order to ensure the generalization capability of the convolutional neural network, the face image information of a construction worker is intercepted through a camera installed at the entrance of a real-name channel of the construction site, the standard of the intercepted image is that the face characteristics can be clearly obtained and the safety helmet is worn, the data set is enlarged, and the data set is classified and stored according to the actual condition of the construction site.
The method specifically comprises the following steps:
First, collecting project department management personnel and construction personnel of each team, self-photographing the front face under the condition of wearing the safety helmet, and using photo information to: the name, the team and the age are put into a database;
Secondly, the face image information of the construction workers is intercepted through a camera arranged at the entrance of a real-name system channel of the construction site, so that a data set is enlarged, and the data set is classified and stored according to the actual site condition;
Finally, the common dataset SHWD is downloaded. SHWD is a common dataset for headgear wear detection, comprising 7581 images, 9044 of which were face-worn headgear subjects (positive), 111514 of which were normal head subjects (not worn or negative).
Step 2: adding labels to the images in the public dataset SHWD, naming and storing the collected images according to the identity information of workers, and carrying out normalization processing;
the input image size of YOLO v3 neural network is 416 x 416, we normalize the collected image to 416 x 416 size, and the image content is maximized to include face and helmet. The classification samples and the data samples marked by the position boundaries need to be marked. For the collected dataset, the facial features and helmet features of the worker are tagged using LabelImg, and these annotations are saved as XML files in PASCAL VOC format for Python reading.
Step 3: building a system architecture based on a YOLO V3 and Light cnn-29 neural network algorithm;
the neural network of the invention adopts the neural network developed by the OpenCV and CAFFE frameworks of the python 3.7 programming language based on the Windows system. Firstly, a system Python environment is built, an image description function is defined, then a training-completed YOLO v3 and Light cnn-29 algorithm is imported, a HAAR CASCADE cascade classifier is imported, finally a Utils module is defined, and the Utils module can read results of the YOLO v3 and Light cnn-29 recognition and output in a combined mode.
The HAAR CASCADE cascade classifier locates the position of the face through an expandable window, and takes the highest score from the prediction window to locate the face.
Step 4: training YOLO v3 on common dataset SHWD and inputting the labeled SHWD dataset image to YOLO v3;
The neural network is trained on devices i7-8700k,NVIDIA GTX1060 6GB.
Creating a VOC format training set, which mainly comprises:
first creating Annotations folder, mainly for storing image marks in xml format, then creating IMAGESETS folder, storing test. Txt, train. Txt, val. Txt in folder for determining SHWD dataset training image and verification image, and finally creating JPEGImages folder for storing SHWD dataset all images.
Step 5: taking the image detected as a face area by the Yolo v3 as a face recognition training set to train the Light CNN-29;
Mainly comprises the following steps:
And an open source deep learning framework Caffe is adopted to realize a Light cnn model. Dropout was used for the fully attached layers, with a ratio set to 0.7. For the convolutional layer and the fully-connected layer other than the fc2 layer, the momentum is set to 0.9 and the weight decay is set to 5×10 4. As the identity in the training dataset increases, fc2 contains a large number of parameters that are not used for feature extraction. To avoid overfitting, we increase the weights of the fc2 layer to attenuate it to 5×10 3. The initial learning rate is set to 1×10 3 and then gradually decreases to 5×10 5.
Step6: detecting the performance of the YOLO v3 and Light cnn-29 neural network systems respectively, testing the recognition accuracy and recall rate under various construction environments, and correcting;
When evaluating a target algorithm recognition algorithm, two indexes of accuracy and recall rate are generally used for measuring the accuracy of the algorithm, namely a real example TP, a false positive example FP and a false negative example FN are defined firstly; TP represents the number of construction workers identified as not wearing a helmet after the algorithm is run, and FP represents the number of construction worker objects identified as not wearing a helmet, but the result is not accurate. FN is the number indicating that a construction worker without a helmet is misjudged. The target recognition accuracy represents the specific gravity of the real sample TP and the total sample (tp+fp) for measuring the reliability of the recognition performance. The recall rate indicates the proportion of the real sample TP to the total positive sample (TP+FN), which are commonly used evaluation indexes for target recognition.
The specific calculation formula is as follows:
the time of target recognition is also an important performance of the measurement algorithm, and here we also use the recognition time as a standard of the detection algorithm;
wherein the detection of the accuracy of test identification under various environments comprises:
the illumination can be changed in different weather environments and in different times of the same day, the definition of the image can be influenced, the change of the number of people and the shielding of the face can influence the recognition accuracy of the algorithm, the two factors are taken into consideration, and the specific classification is shown in the table 1:
table 1 data classification table
The face recognition performance is mainly achieved by comparing a face recognition data set with other face recognition algorithms, and the face recognition performance is specifically shown in the following table:
The detection results are shown in the following table:
table 2 YOLO network structure performance test table
YOLO v3 in this embodiment has different input and output sizes for each layer of the network, and the sizes of each layer of the network are shown in fig. 3. The theory section of the convolutional neural network introduces that the network model receives the original pixels of the input video stream image and finally outputs the original pixels in the form of probability vectors of categories and coordinates. And inputting the original image into a convolution layer, performing convolution operation, performing feature matrix based on an activation function sigmoid, and finally extracting image features.
The working principle of YOLO is to divide the input image into S x S grids, each grid consisting of (x, y, w, h) and confidence C (Object). Coordinates (x, y) represent the center position of the detection bounding box relative to the grid. (W, h) is the width and height of the bounding box. If the center of an object falls within a grid cell, the grid cell is responsible for detecting the object. Each cell of the grid predicts the bounding box and the confidence of that box. The calculation formula is as follows:
C(Object)＝Pr(Object)*IOU(Pred，Truth)
where Pr (Object) indicates whether an Object is contained in the mesh. Pr (Object) =1 if the mesh contains an Object, and Pr (Object) =0 if the mesh does not contain an Object. The IOU (intersection/union) represents the accuracy of the bounding box containing the object, i.e., the overlap ratio of the detected candidate boundaries with the ground truth, i.e., their intersection/union ratio.
The final confidence level is calculated as follows:
After obtaining the confidence level of each prediction frame, one low-scoring prediction frame is removed by setting a threshold value, and then the rest of the boundary frames are suppressed to the non-maximum extent.
The YOLO algorithm uses target detection as a regression problem, using a mean square loss function, with different weights at different places.
In summary, according to the method for identifying the identity of the construction worker and detecting the wearing of the safety helmet, the construction worker face images are collected, a supervised learning mode is adopted, the Python programming language and the YOLO v3 algorithm are utilized for face and safety helmet feature extraction, image information is learned and tested, the wearing detection of the safety helmet is achieved, the Light cnn-29 algorithm is used for identifying the identity of the worker, and the identity verification of the construction worker is achieved. Through a large amount of data acquisition and training, the method has the advantages that the accuracy rate of the identity verification of workers reaches 85 percent, and the average recognition rate of the detection of the safety helmet reaches more than 90 percent.
Compared with the traditional face recognition method, the method can recognize the face of a building worker in various weather and illumination environments, and can achieve recognition rate of more than 85% under the condition that people are shielded from each other, and the algorithm has strong robustness; compared with a common safety helmet detection algorithm, the safety helmet detection method and the safety helmet detection system can expand the function of worker identity verification on the basis of safety helmet identification, and can realize real-time monitoring.
Correspondingly, the invention realizes intelligent safety supervision of the construction site. The face information and the safety helmet information base of the construction workers are established, and a new database is provided for deep learning research in the construction field. The method comprises the steps of establishing a database, and collecting front face pictures of safety helmets worn by project department management personnel and team construction personnel; and collecting real-name channel entrance monitoring videos, and intercepting worker images under different illumination conditions and various weather environments. For the captured images, we normalized to 416 x 416 size, using LabelImg to mark the facial features and helmet features of the worker, these annotations were saved as XML files in the PASCAL VOC format. In deep learning research, image label making is time-consuming and labor-consuming, and the labeled data set can save much time for later scholars to study.
On the other hand, the embodiment of the invention also discloses a system for identifying the identity of a construction worker and detecting the wearing of a safety helmet, which comprises the following units:
the data collection unit is used for respectively collecting the identity information of the construction workers and the corresponding wearing image information of the safety helmet;
the data processing unit is used for adding labels to the collected image information, storing the image information and carrying out normalization processing;
The model training unit is used for inputting the image added with the label into the convolutional neural network to train the neural network model;
The model correction unit is used for detecting the performance of the trained convolutional neural network of the trained training neural network model, testing the recognition precision and recall rate under various environments and correcting;
The data detection unit is used for verifying the identity of the personnel entering the construction site by utilizing the corrected neural network model and detecting the safety helmet of the personnel entering the construction site.
It may be understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and explanation, examples and beneficial effects of the related content may refer to corresponding parts in the above method.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (5)
1. A construction worker identity recognition and safety helmet wearing detection method is characterized in that:
The method comprises the following steps:
downloading a public dataset SHWD, and collecting building worker identity information and worker images monitored by a real-name system entrance;
adding labels to the images in the public dataset SHWD, naming and storing the collected images according to the identity information of workers, and carrying out normalization processing;
building a system architecture based on a YOLO v3 and Light CNN-29 neural network algorithm;
Training YOLO v3 on common dataset SHWD and inputting the labeled SHWD dataset image to YOLO v3;
Taking the image detected as a face area by the Yolo v3 as a face recognition training set to train the Light CNN-29;
detecting the performance of the YOLO v3 and Light CNN-29 neural network systems respectively, testing the recognition accuracy and recall rate under various construction environments, and correcting;
Using the corrected neural network model to carry out identity verification on personnel entering a construction site, and carrying out safety helmet detection on the personnel entering the construction site;
based on a YOLO v3 and Light CNN-29 neural network algorithm, a system architecture is built, comprising:
Firstly, building a system Python environment, defining an image description function, then importing a training-completed YOLO v3 and Light CNN-29 algorithm, importing HAAR CASCADE cascade classifiers, finally defining a Utils module, and reading and combining and outputting results of the YOLO v3 and Light CNN-29 identification by the Utils module;
The introducing HAAR CASCADE of cascade classifiers includes:
HAAR CASCADE the cascade classifier locates the position of the face through an expandable window, and takes the highest score from a prediction window to locate the face;
training YOLO v3 on common dataset SHWD and inputting a labeled SHWD dataset image into YOLO v3, comprising:
creating a VOC-formatted training set comprising:
Creating Annotations folder for storing image mark in xml format, creating IMAGESETS folder, storing test. Txt, train. Txt, val. Txt, determining SHWD data set training image and verifying image, creating JPEGImages folder, storing SHWD data set all images;
Taking the image detected as the face area by Yolo v3 as a face recognition training set, training Light CNN-29, comprising:
The open source deep learning framework Caffe is adopted to realize a Light CNN model, the Dropout method is used for completely connected layers, and the proportion is set to be 0.7; for the convolution layer and the full connection layer other than the fc2 layer, the momentum is set to 0.9, and the weight decay is set to 5×10 4 times;
In order to avoid overfitting, the weights of the fc2 layer are increased to attenuate to 5×10 3 times;
The initial learning rate was set to 1×10 3 times and then gradually decayed to 5×10 5 times.
2. The construction worker identification and helmet wear detection method according to claim 1, wherein: downloading the public dataset SHWD and collecting construction worker identity information and worker images monitored by a real-name portal, comprising:
First, collecting project department management personnel and construction personnel of each team, self-photographing the front face under the condition of wearing the safety helmet, and using photo information to: the name, the team and the age are put into a database;
Secondly, the face image information of the construction workers is intercepted through a camera arranged at the entrance of a real-name system channel of the construction site, so that a data set is enlarged, and the data set is classified and stored according to the actual site condition;
finally, the common data set SHWD, which is the common data set for the helmet wear detection, is downloaded.
3. The construction worker identification and helmet wear detection method according to claim 1, wherein: tagging the images in the common dataset SHWD, naming and storing the collected images with the worker identity information, and performing normalization processing, including:
Standardizing the collected images to 416 x 416 size, wherein the maximized image content comprises faces and helmets;
Further comprises:
Labeling the classified samples and the data samples labeled by the position boundaries; for the collected dataset, the facial features and helmet features of the worker are tagged using LabelImg, and these annotations are saved as XML files in PASCAL VOC format for Python reading.
4. The construction worker identification and helmet wear detection method according to claim 1, wherein: detecting the performance of the YOLO v3 and Light CNN-29 neural network systems respectively, testing the recognition accuracy and recall rate under various construction environments, and correcting;
wherein the detection of recall comprises:
Firstly, defining a real example TP, a false positive example FP and a false negative example FN; wherein TP represents the number of construction workers which are identified as not wearing the safety helmet after the algorithm is operated, FP represents the number of construction workers which are identified as not wearing the safety helmet, but the result is inaccurate; FN is the number indicating that a construction worker without a helmet is misjudged; the target recognition accuracy represents the specific gravity of a real sample TP and a total sample TP+FP and is used for measuring the reliability of recognition performance;
the recall indicates that the real sample TP accounts for the total positive sample (tp+fn) specific gravity;
The specific calculation formula is as follows:
5. a construction worker identity recognition and safety helmet wearing detection system capable of realizing the construction worker identity recognition and safety helmet wearing detection method according to any one of claims 1-4, characterized in that:
comprising the following units:
the data collection unit is used for respectively collecting the identity information of the construction workers and the corresponding wearing image information of the safety helmet;
the data processing unit is used for adding labels to the collected image information, storing the image information and carrying out normalization processing;
The model training unit is used for inputting the image added with the label into the convolutional neural network to train the neural network model;
The model correction unit is used for detecting the performance of the trained convolutional neural network, testing the recognition precision and recall rate under various environments and correcting;
The data detection unit is used for verifying the identity of the personnel entering the construction site by utilizing the corrected neural network model and detecting the safety helmet of the personnel entering the construction site.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010447288.2A CN111598040B (en) | 2020-05-25 | 2020-05-25 | Construction worker identity recognition and safety helmet wearing detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010447288.2A CN111598040B (en) | 2020-05-25 | 2020-05-25 | Construction worker identity recognition and safety helmet wearing detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111598040A CN111598040A (en) | 2020-08-28 |
| CN111598040B true CN111598040B (en) | 2024-05-14 |
Family
ID=72191399
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010447288.2A Active CN111598040B (en) | 2020-05-25 | 2020-05-25 | Construction worker identity recognition and safety helmet wearing detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111598040B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112036360B (en) * | 2020-09-10 | 2023-11-28 | 杭州云栖智慧视通科技有限公司 | Riding helmet attribute identification method |
| CN112183284B (en) * | 2020-09-22 | 2022-09-23 | 上海钧正网络科技有限公司 | Safety information verification and designated driving order receiving control method and device |
| CN112347943A (en) * | 2020-11-09 | 2021-02-09 | 哈尔滨理工大学 | Anchor optimization safety helmet detection method based on YOLOV4 |
| CN112434828B (en) * | 2020-11-23 | 2023-05-16 | 南京富岛软件有限公司 | Intelligent safety protection identification method in 5T operation and maintenance |
| CN112396658B (en) * | 2020-11-30 | 2024-03-19 | 同济人工智能研究院(苏州)有限公司 | Indoor personnel positioning method and system based on video |
| CN112672052A (en) * | 2020-12-24 | 2021-04-16 | 华中光电技术研究所（中国船舶重工集团公司第七一七研究所） | Image data enhancement method and system, electronic equipment and storage medium |
| CN112668465A (en) * | 2020-12-25 | 2021-04-16 | 秒影工场(北京)科技有限公司 | Film face extraction method based on multistage CNN |
| CN112818175B (en) * | 2021-02-07 | 2023-09-01 | 中国矿业大学 | Factory staff searching method and training method of staff identification model |
| CN113256934A (en) * | 2021-05-18 | 2021-08-13 | 哈尔滨理工大学 | Safety detection system for aerial work personnel |
| CN113516076B (en) * | 2021-07-12 | 2023-09-01 | 大连民族大学 | Attention mechanism improvement-based lightweight YOLO v4 safety protection detection method |
| CN114283485B (en) * | 2022-03-04 | 2022-10-14 | 杭州格物智安科技有限公司 | Safety helmet wearing detection method and device, storage medium and safety helmet |
| CN114627499B (en) * | 2022-03-07 | 2024-04-09 | 上海应用技术大学 | On-line safety helmet face recognition method based on convolutional neural network |
| CN114511427A (en) * | 2022-04-21 | 2022-05-17 | 四川省大数据中心 | Safety education supervision method, device and system for project site |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119686A (en) * | 2019-04-17 | 2019-08-13 | 电子科技大学 | A kind of safety cap real-time detection method based on convolutional neural networks |
| 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 |
| CN110533811A (en) * | 2019-08-28 | 2019-12-03 | 深圳市万睿智能科技有限公司 | The method and device and system and storage medium of safety cap inspection are realized based on SSD |
| CN110728223A (en) * | 2019-10-08 | 2020-01-24 | 济南东朔微电子有限公司 | Helmet wearing identification method based on deep learning |
| CN110738127A (en) * | 2019-09-19 | 2020-01-31 | 福建师范大学福清分校 | Helmet identification method based on unsupervised deep learning neural network algorithm |
| CN111160440A (en) * | 2019-12-24 | 2020-05-15 | 广东省智能制造研究所 | Helmet wearing detection method and device based on deep learning |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609481B (en) * | 2017-08-14 | 2020-11-20 | 百度在线网络技术（北京）有限公司 | Method, apparatus and computer storage medium for generating training data for face recognition |
-
2020
- 2020-05-25 CN CN202010447288.2A patent/CN111598040B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119686A (en) * | 2019-04-17 | 2019-08-13 | 电子科技大学 | A kind of safety cap real-time detection method based on convolutional neural networks |
| 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 |
| CN110533811A (en) * | 2019-08-28 | 2019-12-03 | 深圳市万睿智能科技有限公司 | The method and device and system and storage medium of safety cap inspection are realized based on SSD |
| CN110738127A (en) * | 2019-09-19 | 2020-01-31 | 福建师范大学福清分校 | Helmet identification method based on unsupervised deep learning neural network algorithm |
| CN110728223A (en) * | 2019-10-08 | 2020-01-24 | 济南东朔微电子有限公司 | Helmet wearing identification method based on deep learning |
| CN111160440A (en) * | 2019-12-24 | 2020-05-15 | 广东省智能制造研究所 | Helmet wearing detection method and device based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111598040A (en) | 2020-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111598040B (en) | Construction worker identity recognition and safety helmet wearing detection method and system | |
| CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
| CN113516076B (en) | Attention mechanism improvement-based lightweight YOLO v4 safety protection detection method | |
| CN112085010B (en) | Mask detection and deployment system and method based on image recognition | |
| CN111881730A (en) | Wearing detection method for on-site safety helmet of thermal power plant | |
| CN110852183B (en) | Method, system, device and storage medium for identifying person without wearing safety helmet | |
| CN110728252B (en) | Face detection method applied to regional personnel motion trail monitoring | |
| CN112347943A (en) | Anchor optimization safety helmet detection method based on YOLOV4 | |
| CN112613454A (en) | Electric power infrastructure construction site violation identification method and system | |
| CN111652046A (en) | Safe wearing detection method, equipment and system based on deep learning | |
| CN110458794B (en) | Quality detection method and device for accessories of rail train | |
| US20230186634A1 (en) | Vision-based monitoring of site safety compliance based on worker re-identification and personal protective equipment classification | |
| CN117035419B (en) | Intelligent management system and method for enterprise project implementation | |
| CN113807240A (en) | Intelligent transformer substation personnel dressing monitoring method based on uncooperative face recognition | |
| CN111860187A (en) | High-precision worn mask identification method and system | |
| CN111931573A (en) | Helmet detection and early warning method based on YOLO evolution deep learning model | |
| CN113052125B (en) | Construction site violation image recognition and alarm method | |
| CN116311081B (en) | Medical laboratory monitoring image analysis method and system based on image recognition | |
| CN112686180A (en) | Method for calculating number of personnel in closed space | |
| CN112183532A (en) | Safety helmet identification method based on weak supervision collaborative learning algorithm and storage medium | |
| Rivera et al. | Work safety assessment through contextual analysis with computer vision | |
| CN116977904A (en) | Yolov 5-based rapid large-scene-identification multi-man-made garment detection method | |
| CN114882442A (en) | Personnel and equipment situation identification method based on electric power operation site | |
| CN113033289A (en) | Safety helmet wearing inspection method, device and system based on DSSD algorithm | |
| CN116579609B (en) | Illegal operation analysis method based on inspection process |
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 |