CN115690687A - Safe wearing standard detection system based on deep learning technology - Google Patents
Safe wearing standard detection system based on deep learning technology Download PDFInfo
- Publication number
- CN115690687A CN115690687A CN202211411823.4A CN202211411823A CN115690687A CN 115690687 A CN115690687 A CN 115690687A CN 202211411823 A CN202211411823 A CN 202211411823A CN 115690687 A CN115690687 A CN 115690687A
- Authority
- CN
- China
- Prior art keywords
- detection
- model
- safety
- identification
- deep learning
- 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
Images
Abstract
The invention discloses a safe wearing standard detection system based on a deep learning technology, which comprises: the system comprises a monitoring camera image acquisition module, an algorithm detection and identification module, a one-stage personnel detection and identification module, a two-stage safety wearing detection and identification module and an alarm module. The invention realizes the purpose that whether the unmanned monitoring staff accords with the safety dressing standard or not by combining deep learning and hardware, and realizes intelligent management. Secondly, through real-time supervision, very big prevention the emergence of potential safety hazard to the personal safety of staff has further been ensured.
Description
Technical Field
The invention relates to a safety wearing standard detection system, in particular to a safety wearing standard detection system based on a deep learning technology.
Background
Nowadays, production safety becomes a most important link for each enterprise, and along with the development of modernization and the requirement of industrial intelligent manufacturing, safe production also becomes more and more the most important link in each enterprise. The protection measures of the safety dressing standard detection system of the enterprise are still in a relatively laggard stage at present, and many original manual supervision modes are still available for carrying out safety dressing standard detection. In reality, many workers at the production line lack safety consciousness and enter high-risk areas or do not wear safety helmets or protective clothing, so that safety accidents occur. Therefore, the safety awareness of workers is improved, and meanwhile, the key effect is played on the safety supervision and early warning of the site. However, the dependence of the current safety wearing standard detection system on manual work is still very serious, and due to the large working area, the large number of workers, the long working time and the like, the manual supervision difficulty is large, and the supervision efficiency is low. In recent years, with the continuous development of deep learning, the combination of deep learning and computer vision technology is becoming more and more compact. In the field of vision, deep learning plays a very important role, and is widely applied to different directions such as security detection, target detection and the like. Meanwhile, with the continuous development of hardware and electronic technology, the combination of monitoring equipment and calculator vision is promoted, and the intellectualization of the monitoring system becomes practical by utilizing a target detection algorithm of deep learning.
At present, the research of being applied to safe dress standard detecting system with deep learning is still comparatively few, mainly combines together manual supervision and traditional machine vision, has protected staff's safety to a certain extent, but still has following several problems:
the intellectualization is insufficient. The existing safety wearing standard detection system mainly depends on manual supervision and judgment, and fatigue and judgment errors exist in manual work, so that great potential safety hazards exist.
The function is single. The traditional machine vision carries out safety dressing standard detection and depends on a fixed environment, the application scene is single, and the use is not flexible enough.
The detection precision is lower. Traditional machine vision is to wearing the stability that the standard detection relies on operational environment safely, and the stability of light detects the precision and all can reduce under weather variation and different job scenes, and different personnel gestures can all influence the judgement, will produce like this and miss the inspection with the false retrieval scheduling problem.
Therefore, a set of efficient, flexible and reliable intelligent safety wearing standard detection system is urgently needed to be developed, the requirements of the function and the performance of the system are determined by analyzing the problems, a deep learning target detection algorithm is applied to the safety wearing standard detection system, and the safety wearing standard detection system of the deep learning technology is provided. This system is more intelligent and high-efficient to safety protection's supervision, is applicable to different scenes and detects, detects the precision height and is greater than artifical and traditional machine vision detection rate far away, carries out safety precaution through this system simultaneously, the relevant inspection of the manpower that has significantly reduced and administrative work to reduce the supervision cost of enterprise. For this reason, a corresponding technical scheme needs to be designed for solution.
Disclosure of Invention
The invention aims to solve the problems and provides a safe and reliable safe wearing standard detection system based on a deep learning technology, which is convenient to operate.
In order to achieve the purpose, the invention adopts the following technical scheme: a safe wearing standard detection system based on deep learning technology comprises:
(1) Monitoring camera image acquisition module
The monitoring camera needs to acquire a certain number of pictures to supply to model training, and the model training enables the model to have intelligent detection and identification capability;
the model finishes generating a pt model, the pt model is deployed at a server end, and when the whole detection system is on line, a monitoring camera needs to acquire images in real time and transmit the images to the model for identification and detection;
(2) Algorithm detection and identification module
The algorithm model mainly selects a deep learning target detection model YOLO model, and mainly adopts a two-step walking principle, firstly, the model is used for detecting and identifying personnel, and the safety wearing detection and identification are carried out;
(2.1) one-stage person detection and identification
After the trained pt model is deployed to a server, the deployed pt model is connected with a monitoring camera through software, the monitoring camera acquires images and transmits the images to the model for forward reasoning, and finally the pixel position coordinates of all the personnel in the image of the current frame are obtained;
(2.2) two-stage safety wearing detection identification
And (4) screenshot is carried out according to the pixel position coordinates of all the personnel obtained in the one-stage, and the screenshot pictures are sequentially transmitted into a safety wearing detection model to carry out standard detection on the picture safety wearing. Sending the recognition result to an alarm module;
(3) Alarm module
And after the alarm module receives the identification result, if the worker is found not to conform to the safety dressing standard behavior, voice alarm is carried out, and if all the workers are normal, real-time monitoring is continued.
Compared with the prior art, the invention has the following beneficial effects:
(1) The invention realizes the purpose that whether the unmanned monitoring staff accords with the safety dressing standard or not by combining deep learning and hardware, and realizes intelligent management.
(2) According to the invention, through real-time monitoring, the occurrence of potential safety hazards is greatly prevented, so that the life safety of workers is further ensured.
(3) The effect of industrial production is improved, manual supervision is liberated, and the labor cost of enterprises is reduced.
Drawings
Fig. 1 is an overall flowchart of a safety wear specification detection system.
Detailed Description
In order to make the objects, technical solutions, advantages and significant advances of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the drawings provided in the embodiments of the present invention, it is obvious that all the described embodiments are only some embodiments of the present invention, rather than all embodiments, and all other embodiments obtained by a person of ordinary skill in the art without making creative efforts based on the embodiments of the present invention belong to the protection scope of the present invention.
It should be noted that:
the terms "first," "second," and "third" (if any), etc. in the description and claims of the present invention and the accompanying drawings of embodiments of the present invention are used for distinguishing between different objects and not for describing a particular order;
moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements but may alternatively include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It is to be understood that:
in the description of the embodiments of the present invention, the terms "upper", "lower", "top", "bottom", and other indicative orientations or positions are used only for the convenience of describing the embodiments of the present invention and for the simplicity of explanation, and are not intended to indicate or imply that the described devices or elements must have a particular orientation, a configuration, and an operation, and therefore, should not be construed as limiting the present invention.
In the present invention, unless otherwise specifically stated or limited, the terms "mounted," "connected," "fixed," and the like are to be understood broadly, and for example, may be fixedly connected, detachably connected, movably connected, or integrated; the term "a" or "an" refers to a compound that can be directly connected or indirectly connected through an intermediate, and can be used in combination with or without other elements, unless otherwise specifically limited, and the specific meaning of the term in the present invention can be understood by those skilled in the art according to specific situations.
It should also be noted that:
the following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.
The technical means of the present invention will be described in detail with reference to specific examples.
A safe wearing standard detection system based on deep learning technology comprises:
(1) Monitoring camera image acquisition module
The monitoring camera needs to acquire a certain number of pictures to supply to model training, and the model training enables the model to have the capacity of intelligent detection and identification;
the model finishes generating a pt model, the pt model is deployed at a server end, and when the whole detection system is on line, a monitoring camera needs to acquire images in real time and transmit the images to the model for identification and detection;
(2) Algorithm detection and identification module
The algorithm model mainly selects a deep learning target detection model YOLOV6 model, and mainly adopts a two-step walking principle, firstly, the model is used for detecting and identifying personnel, and then safety wearing detection and identification are carried out;
(2.1) one-stage person detection and identification
After the trained pt model is deployed to a server, the deployed pt model is connected with a monitoring camera through software, the monitoring camera acquires images and transmits the images to the model for forward reasoning, and finally the pixel position coordinates of all the personnel in the image of the current frame are obtained;
(2.2) two-stage safety wearing detection identification
And (4) screenshot is carried out according to the pixel position coordinates of all the personnel obtained in the one-stage, and the screenshot pictures are sequentially transmitted into a safety wearing detection model to carry out standard detection on the picture safety wearing. Sending the recognition result to an alarm module;
(3) Alarm module
And after the alarm module receives the identification result, if the worker is found not to conform to the safety dressing standard behavior, voice alarm is carried out, and if all the workers are normal, real-time monitoring is continued.
Example 1
Referring to fig. 1, the present embodiment relates to a deep learning technology-based safety wearing specification detection system, which includes a monitoring camera image acquisition module, an algorithm detection model identification module, and an alarm module, and mainly detects whether a standard behavior of a safety helmet is safely worn.
The method comprises the steps of utilizing a monitoring camera to conduct image acquisition when a worker wears a safety helmet in the early safety production process, acquiring 1 million pictures, respectively marking the safety helmet and the worker in the pictures, independently placing files in a json format to be marked into two folders, conducting data processing on the json files, then respectively training hat marking data and pendant marking data through a yolov6 target detection model, and obtaining a worker recognition detection model and a safety helmet recognition detection model.
The model is deployed by using a tensorrt deployment framework, a version of tensorrt7.2.2.3 is selected, a cuda10.2 version is installed on a server, a helmet identification detection model is trained by using yolov6 and renamed to hat.pt, a worker identifies the detection model to be hat.pt, the two models are respectively converted into hat.onnx and hat.onnx models after training is completed, the models are deployed on the server by using c + + in combination with tensorrt7.2.2.3, and forward reasoning is carried out.
The method comprises the steps of butting tensiorrt deployed software with a front end, collecting pictures in real time through a monitoring camera, sending the collected pictures to a people. And (3) scaling and filling the picture output by crop into a (640 ) picture, sequentially transmitting the pictures into a hat.
And finally, a voice alarm part, wherein the server is in contact with the HMI client through socket communication, the server transmits an instruction to control the HMI client to send an alarm instruction, when the fact that the pictures of the workers do not detect the safety helmet is found in the inference result of the hat.
During the description of the above description:
the description of the terms "present embodiment," like "\8230; \8230"; "shown," "further improved technical sub-scheme," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention, and exemplary expressions of the terms in this specification are not necessarily directed to the same embodiment or example, and that the particular feature, structure, material, or characteristic described may be combined or coupled in any suitable manner in any one or more embodiments or examples;
furthermore, those of ordinary skill in the art may combine or combine features of different embodiments or examples and features of different embodiments or examples described in this specification without undue conflict.
Finally, it should be noted that:
although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made on the technical solutions described in the foregoing embodiments, or some or all of the technical features of the embodiments can be equivalently replaced, and the corresponding technical solutions do not depart from the technical solutions of the embodiments of the present invention.
Claims (3)
1. The utility model provides a safe standard detecting system of wearing based on deep learning technique which characterized in that: the method comprises the following steps:
(1) Monitoring camera image acquisition module
The monitoring camera needs to acquire a certain number of pictures to supply to model training, and the model training enables the model to have intelligent detection and identification capability; the model generates a pt model, the pt model is deployed at a server end, and when the whole detection system is on line, a monitoring camera needs to acquire images in real time and transmit the images to the model for identification and detection;
(2) Algorithm detection and identification module
The method mainly selects a deep learning target detection model YOLO model, and adopts a two-step walking principle, firstly, the model is used for detecting and identifying personnel, and then the safety wearing detection identification is carried out;
(3) Alarm module
And after the alarm module receives the identification result, if the worker is found not to conform to the safety dressing standard behavior, voice alarm is carried out, and if all the workers are normal, real-time monitoring is continued.
2. The deep learning technology-based safety wear specification detection system according to claim 1, characterized in that: the algorithm detection and identification module further comprises: one-stage personnel detection and identification
one-stage personnel detection identification
After the trained pt model is deployed to a server, the deployed pt model is connected with a monitoring camera through software, the monitoring camera acquires images and transmits the images to the model for forward reasoning, and finally the pixel position coordinates of all the personnel in the image of the current frame are obtained.
3. The deep learning technology-based safety wear specification detection system according to claim 1, characterized in that: the algorithm detection and identification module further comprises: two-stage safety wearing detection identification
two-stage safety wearing detection identification
And (4) screenshot is carried out according to the pixel position coordinates of all the personnel obtained in the one-stage, the screenshot pictures are sequentially transmitted into a safety wearing detection model to carry out picture safety wearing standard detection, and the identification result is sent to an alarm module. A safety wearing specification detection system based on a deep learning technology.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211411823.4A CN115690687A (en) | 2022-11-11 | 2022-11-11 | Safe wearing standard detection system based on deep learning technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211411823.4A CN115690687A (en) | 2022-11-11 | 2022-11-11 | Safe wearing standard detection system based on deep learning technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115690687A true CN115690687A (en) | 2023-02-03 |
Family
ID=85052367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211411823.4A Pending CN115690687A (en) | 2022-11-11 | 2022-11-11 | Safe wearing standard detection system based on deep learning technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115690687A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116502810A (en) * | 2023-06-28 | 2023-07-28 | 威胜信息技术股份有限公司 | Standardized production monitoring method based on image recognition |
-
2022
- 2022-11-11 CN CN202211411823.4A patent/CN115690687A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116502810A (en) * | 2023-06-28 | 2023-07-28 | 威胜信息技术股份有限公司 | Standardized production monitoring method based on image recognition |
| CN116502810B (en) * | 2023-06-28 | 2023-11-03 | 威胜信息技术股份有限公司 | Standardized production monitoring method based on image recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113705372B (en) | AI identification system for join in marriage net job site violating regulations | |
| CN112183317B (en) | Live working field violation behavior detection method based on space-time diagram convolutional neural network | |
| CN211509189U (en) | Electric power operation field supervision equipment and system | |
| CN110874866B (en) | Video-based three-dimensional monitoring method and system for transformer substation | |
| CN112149522A (en) | Intelligent visual external-damage-prevention monitoring system and method for cable channel | |
| CN111920129A (en) | Intelligent safety helmet system | |
| CN104110271A (en) | Under-mine equipment monitoring alarm system based on face recognition technology | |
| CN115690687A (en) | Safe wearing standard detection system based on deep learning technology | |
| CN113206978A (en) | Security intelligent monitoring early warning system and method for oil and gas pipeline station | |
| CN114863489B (en) | Virtual reality-based movable intelligent auxiliary inspection method and system for construction site | |
| CN112188164A (en) | AI vision-based violation real-time monitoring system and method | |
| CN112434828A (en) | Intelligent identification method for safety protection in 5T operation and maintenance | |
| CN115880722A (en) | Intelligent identification method, system and medium worn by power distribution operating personnel | |
| CN115731670A (en) | Safety alarm method, device and monitoring system for construction site | |
| CN110310450A (en) | A kind of electrical intelligent fire disaster cruising inspection system and method based on wearable device | |
| CN213128247U (en) | Intelligent safety helmet system | |
| CN113807240A (en) | Intelligent transformer substation personnel dressing monitoring method based on uncooperative face recognition | |
| CN113506416A (en) | Engineering abnormity early warning method and system based on intelligent visual analysis | |
| CN116624225A (en) | Tunnel defect picture and text inspection information real-time transmission early warning system | |
| CN115641608A (en) | Intelligent identification equipment for personal protection articles for distribution network operation based on edge intelligence | |
| CN114997582A (en) | Early warning method and system for distribution network operation risk and computer equipment | |
| CN114422750A (en) | Movable edge calculation camera system and device for construction safety management | |
| CN113188662A (en) | Infrared thermal imaging fault automatic identification system and method | |
| CN215871468U (en) | Operation site digital supervision system based on intelligent robot dog | |
| CN215871711U (en) | Operation safety monitoring device based on high-speed wireless network and Bluetooth network |
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 |