CN115798139A - Method and system for monitoring on-duty state of control room personnel - Google Patents
Method and system for monitoring on-duty state of control room personnel Download PDFInfo
- Publication number
- CN115798139A CN115798139A CN202211431898.9A CN202211431898A CN115798139A CN 115798139 A CN115798139 A CN 115798139A CN 202211431898 A CN202211431898 A CN 202211431898A CN 115798139 A CN115798139 A CN 115798139A
- Authority
- CN
- China
- Prior art keywords
- alarm
- time
- early warning
- warning
- information
- 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
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Alarm Systems (AREA)
Abstract
The invention discloses a method and a system for monitoring the on-duty state of personnel in a control room, and belongs to the technical field of intelligent video monitoring. The method judges whether the current image is an alarm or not by using the recognition algorithms under different scenes, and further fuses the historical state of the current analysis item into the judgment of the monitoring result, thereby solving the problems of low accuracy and less effective alarm ratio of the industrial field which only depends on the on-duty state of personnel in an AI analysis control room. The invention can effectively reduce the invalid picture alarm in the intelligent video analysis result, so that managers in the industrial field can pay attention to the real risk points and hidden danger points of the personnel in the control room, and the working efficiency is improved.
Description
Technical Field
The invention discloses a method and a system for monitoring the on-duty state of personnel in a control room, and belongs to the technical field of intelligent video monitoring.
Background
With the application of artificial intelligence identification technology in the industrial field, especially in the hazardous chemical industry, the functions of automatically identifying visual information such as worker wearing non-compliance, console eating, smoke, naked flame, smoking and the like and marking an alarm object identified in an image have been realized at present: such as identifying tooling, food, smoke, fire, safety helmets, etc.
However, with the increase of the amount of alarm information, a new technical problem occurs in the technical field, that is, the artificial intelligence recognition technology can only recognize a single-frame picture at present, and cannot meet the actual requirements for recognition of relevant behaviors at a plurality of time points and judgment of the accuracy and the continuity of alarm, and further cannot reasonably judge the severity of illegal work of employees.
Therefore, the method is particularly important for monitoring the on-duty state of the personnel according to the mass data information. How to provide an effective auxiliary method to improve the alarm accuracy rate and provide effective alarm information for control room personnel in the industrial field becomes an urgent problem to be solved by technical personnel in the field.
Disclosure of Invention
Aiming at the defects of the prior art, the invention discloses a method for monitoring the on-duty state of personnel in a control room.
The invention also discloses a system for realizing the monitoring method.
The monitoring method of the invention judges whether the current image is an alarm or not by using the recognition algorithms under different scenes, and further fuses the historical state of the current analysis item into the judgment of the monitoring result, thereby solving the problems that the accuracy of the on-duty state of the personnel in the control room is low and the effective alarm ratio is low because the personnel only rely on the AI analysis in the industrial field. Behaviors closely related to time such as leaving behind, sleeping behind, routing inspection and the like cannot be effectively solved through a simple artificial intelligence identification mode.
The detailed technical scheme of the invention is as follows:
a monitoring method for the on-duty state of the personnel in a control room is characterized by comprising the following steps:
1) Setting an alarm interval, a repeated alarm interval and a warning-eliminating interval corresponding to the identification item; according to different application scenes, the invention can enable a user to preset parameters for different identification items, wherein the identification items are as follows: whether the personnel leave the post or whether the personnel sleep on the post for the item to be identified, the hardware and the algorithm for identifying the identification item are not the contents to be protected by the invention, because the part of the contents can further select the corresponding hardware and call the AI algorithm according to different application scenes of the user; the alarm interval refers to the time from the first recognition of the AI algorithm to the recognition item alarm to the real triggering of the first alarm; the repeated alarm interval refers to the interval time between the time when the same alarm appears again next time after the alarm is judged last time; the alarm elimination interval refers to the time for identifying the identification item alarm for the first time and determining the elimination time for the alarm for how long; in addition, the invention also comprises alarm elimination time which is a time node from alarm generation to alarm elimination;
2) Utilizing the existing AI algorithm to correspondingly identify the identification items to obtain identification results, preprocessing the identification results, and associating the corresponding equipment information to the identification results;
3) And storing alarm information according to the identification result:
when the identification item is identified as not-compliant by using the conventional AI algorithm, defining the identification item as an early warning alarm;
recording early warning information according to equipment and identification items in a classified manner, wherein the early warning information comprises whether the early warning information is firstly identified as an early warning and time node and/or historical state information, and the historical state information is historical identification as the early warning and time node;
in step 3), further comprises
3-1) the method for further judging the early warning alarm corresponding to the identification result as the first warning comprises the following steps:
when the early warning information does not contain the last early warning in the historical state information, subtracting the time which is firstly identified as the early warning from the current early warning time, and judging that the early warning under the equipment is the first warning if the result is greater than the warning interval; otherwise, judging the early warning alarm of the first recognition under the equipment as a first alarm; when the time for early warning is not identified for the first time, judging the early warning under the equipment as the first warning;
when the early warning information contains the last early warning in the historical state information, subtracting the time of the last early warning from the current early warning time, and if the result is greater than the warning interval, sending the warning under the equipment again; otherwise, not sending out the alarm under the equipment;
3-2) when the recognition result is non-early warning, the method for judging whether the warning is removed or not comprises the following steps:
when the current time minus the alarm-eliminating time is greater than the alarm-eliminating interval, judging the alarm-eliminating time is an alarm-eliminating time; otherwise, judging that the alarm is not cleared; and when the alarm-eliminating time does not exist, judging the alarm-eliminating time to be alarm-eliminating time.
Preferably, according to the present invention, the monitoring method further includes step 4) forming single alarm information:
the single warning information is the time information of key points in the whole life cycle of warning; the single alarm is an identification result set which is identified as an early warning alarm for the first time, judged as an alarm for the first time and eliminated.
According to the invention, the single alarm further comprises a recognition result set of repeated alarms.
According to the present invention, preferably, the single warning message includes: the first judgment is the time of alarming, the time of repeated alarming and the time of alarm elimination.
According to the invention, the corresponding penalty level is preset in advance according to the specific situation of the single alarm. This design is in order to link up with industrial control room management scene phase-match, makes industrial control more intelligent diversified.
A system for realizing the monitoring method is characterized by comprising the following steps:
the system comprises video equipment for acquiring image information of personnel in a control room, an AI module loaded with an AI algorithm and a monitoring module loaded with the monitoring method;
the AI module is arranged on the video equipment or on a remote monitoring platform;
the monitoring module is installed in the AI module or independently installed in the remote monitoring platform.
The invention has the beneficial effects that:
1. because the control room is frequently grabbed, and the AI can not be identified as an alarm every time and is pushed to the client, the invention effectively solves the technical problems that the detection of off-duty and regular inspection alarm information in the scene of the industrial control room is huge and difficult to accurately judge by adding the comparison of the alarm interval and the historical state information, namely effectively solves the problems of excessive alarm quantity and less effective alarm ratio.
2. The invention can make the program think that the personnel state in the control room is unchanged in the time range of the alarm elimination interval, eliminates the error caused by the movement and dynamic shielding of the personnel in a period of time range (the time configured in the alarm elimination interval) by adding the alarm elimination interval and the historical state information, and solves the problem of false alarm elimination caused by the movement, dynamic shielding and other reasons of the personnel to a certain extent.
3. The invention integrates the three parameters and the historical state information to solve the problem of false alarm, reduces the number of invalid alarms and lightens the workload of illegal personnel for processing scenes in an industrial control room. In particular to alarm filtering and screening based on artificial intelligence identification result data, which reduces false alarm and improves alarm accuracy.
4. The invention adds alarm-eliminating interval parameter judgment to the non-alarm result of the existing AI algorithm to solve the problem of false identification of the AI algorithm, and the higher the doubtful degree of the identification result of the AI algorithm is, the larger the value of the parameter should be theoretically set.
Drawings
FIG. 1 is a schematic flow diagram of a monitoring method according to the present invention;
FIG. 2 is a human-computer interaction interface for setting repeat alarm intervals, alarm elimination intervals, and alarm intervals in an embodiment of the present invention;
FIG. 3 is a diagram illustrating a human-computer interaction interface obtained after monitoring in an embodiment of the present invention, in which a part of the alarm result is displayed.
Detailed description of the invention
The present invention is further illustrated by, but not limited to, the following examples.
Examples 1,
As shown in fig. 1, a method for monitoring the on-duty status of a person in a control room, includes:
1) Setting an alarm interval, a repeated alarm interval and a warning-eliminating interval corresponding to the identification item;
2) Correspondingly identifying the identification items by utilizing the conventional AI algorithm to obtain an identification result, preprocessing the identification result, and associating corresponding equipment information into the identification result;
3) And storing alarm information according to the identification result:
when the identification item is identified as not in compliance by using the conventional AI algorithm, an early warning alarm is defined;
recording early warning information according to equipment and identification items in a classified manner, wherein the early warning information comprises whether the early warning information is firstly identified as an early warning and time node and/or historical state information, and the historical state information is historical identification as the early warning and time node;
in step 3), further comprises
3-1) the method for further judging the early warning alarm corresponding to the identification result as the first warning comprises the following steps:
when the early warning information does not contain the last early warning in the historical state information, subtracting the time which is firstly identified as the early warning from the current early warning time, and judging the early warning under the equipment as the first warning if the result is larger than the warning interval; otherwise, judging the early warning alarm of the first recognition under the equipment as a first alarm; when the time for early warning is not identified for the first time, judging the early warning under the equipment as the first warning;
when the early warning information contains the last early warning in the historical state information, subtracting the time of the last early warning from the current early warning time, and if the result is greater than the warning interval, sending the warning under the equipment again; otherwise, not sending out the alarm under the equipment;
3-2) when the recognition result is non-early warning, the method for judging whether the warning is removed or not comprises the following steps:
when the current time minus the alarm-eliminating time is greater than the alarm-eliminating interval, judging the alarm-eliminating time is an alarm-eliminating time; otherwise, judging that the alarm is not canceled; and when the alarm-eliminating time does not exist, judging the alarm-eliminating time to be alarm-eliminating time.
The monitoring method also comprises the following step 4) of forming single alarm information:
the single warning information is the time information of key points in the whole life cycle of warning; the single alarm is an identification result set which is identified as an early warning alarm for the first time, judged as an alarm for the first time and eliminated.
Examples 2,
The monitoring method according to embodiment 1, wherein the single alarm further comprises a recognition result set of repeated alarms.
Examples 3,
The monitoring method according to embodiment 1, wherein the single warning message includes: the first judgment is the time of alarm, the time of repeated alarm and the time of alarm elimination.
Examples 4,
The monitoring method according to embodiment 1 presets a corresponding penalty level in advance according to the specific situation of a single alarm.
Examples 5,
A system for implementing the monitoring method of embodiments 1-4, comprising:
the system comprises video equipment for acquiring image information of personnel in a control room, an AI module loaded with an AI algorithm and a monitoring module loaded with the monitoring method;
the AI module is arranged on the video equipment or on a remote monitoring platform;
the monitoring module is installed in the AI module or independently installed in the remote monitoring platform.
With reference to embodiments 1 to 5, the specific application scenario is described as follows:
a. a control room off duty identification item of a certain factory is respectively provided with a repeated alarm interval, a warning elimination interval and alarm intervals of 300 seconds, 0 second and 300 seconds, namely, the alarm is considered to be generated after 5 minutes of first detection of the alarm, the alarm is pushed and still in an illegal state, the alarm is pushed again after 5 minutes, and the artificial intelligence algorithm detects non-illegal behavior for the first time and immediately cancels the alarm of the behavior, as shown in figure 2;
b.23:05:
c. after the alarm lasts for five minutes, judging the alarm, and updating the state:
d. next, if the alarm continuously occurs, the state is updated every 5 minutes; e. the alarm is eliminated and the final state of the alarm is updated
f. And finally giving an alarm result. As shown in fig. 3.
Claims (6)
1. A monitoring method for the on-duty state of the personnel in a control room is characterized by comprising the following steps:
1) Setting an alarm interval, a repeated alarm interval and a warning-eliminating interval corresponding to the identification item;
2) Correspondingly identifying the identification items by utilizing the conventional AI algorithm to obtain an identification result, preprocessing the identification result, and associating corresponding equipment information into the identification result;
3) And storing alarm information according to the identification result:
when the identification item is identified as not-compliant by using the conventional AI algorithm, defining the identification item as an early warning alarm;
recording early warning information according to equipment and identification items in a classified manner, wherein the early warning information comprises whether the early warning information is firstly identified as an early warning and time node and/or historical state information, and the historical state information is historical identification as the early warning and time node;
in step 3), further comprising
3-1) the method for further judging the early warning alarm corresponding to the identification result as the first warning comprises the following steps:
when the early warning information does not contain the last early warning in the historical state information, subtracting the time which is firstly identified as the early warning from the current early warning time, and judging the early warning under the equipment as the first warning if the result is larger than the warning interval; otherwise, judging the early warning alarm of the first recognition under the equipment as a first alarm; when the time for early warning is not identified for the first time, judging the early warning under the equipment as the first warning;
when the early warning information contains the last early warning in the historical state information, subtracting the time of the last early warning from the current early warning time, and if the result is greater than the warning interval, sending the warning under the equipment again; otherwise, not sending out the alarm under the equipment;
3-2) when the recognition result is non-early warning, the method for judging whether the warning is removed or not comprises the following steps:
when the current time minus the alarm-eliminating time is greater than the alarm-eliminating interval, judging the alarm-eliminating time is an alarm-eliminating time; otherwise, judging that the alarm is not cleared; and when the alarm-eliminating time does not exist, judging the alarm-eliminating time to be alarm-eliminating time.
2. The method for monitoring the on-duty state of the personnel in the control room according to the claim 1, characterized in that the monitoring method further comprises the step 4) of forming a single alarm message:
the single warning information is the time information of key points in the whole life cycle of warning; the single alarm is an identification result set which is identified as an early warning alarm for the first time, judged as an alarm for the first time and eliminated.
3. The method as claimed in claim 2, wherein the single alarm further comprises a recognition result set of repeated alarms.
4. The method for monitoring the on-duty state of the personnel in the control room according to claim 2, wherein the single alarm message comprises: the first judgment is the time of alarm, the time of repeated alarm and the time of alarm elimination.
5. The method for monitoring the on-duty state of the personnel in the control room according to claim 2, wherein the corresponding penalty level is preset in advance according to the specific situation of a single alarm.
6. A system for implementing the monitoring method according to any one of claims 1 to 5, comprising:
the system comprises video equipment for acquiring image information of personnel in a control room, an AI module loaded with an AI algorithm and a monitoring module loaded with the monitoring method;
the AI module is arranged on the video equipment or on a remote monitoring platform;
the monitoring module is installed in the AI module or independently installed in the remote monitoring platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211431898.9A CN115798139A (en) | 2022-11-16 | 2022-11-16 | Method and system for monitoring on-duty state of control room personnel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211431898.9A CN115798139A (en) | 2022-11-16 | 2022-11-16 | Method and system for monitoring on-duty state of control room personnel |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115798139A true CN115798139A (en) | 2023-03-14 |
Family
ID=85438012
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211431898.9A Pending CN115798139A (en) | 2022-11-16 | 2022-11-16 | Method and system for monitoring on-duty state of control room personnel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115798139A (en) |
-
2022
- 2022-11-16 CN CN202211431898.9A patent/CN115798139A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4668978B2 (en) | Flame detection method and apparatus | |
| CN110705834B (en) | Industrial and mining enterprise dangerous state identification method | |
| KR102058452B1 (en) | IoT Convergence Intelligent Video Analysis Platform System | |
| CN102447570A (en) | Monitoring device and method based on health degree analysis | |
| CN108288361A (en) | A kind of passageway for fire apparatus door state detection method | |
| CN112396796A (en) | Automatic fire hazard identification early warning monitoring system based on computer vision | |
| CN113470302A (en) | Intelligent security management method and system based on Internet of things and storage medium | |
| CN113537009A (en) | Household isolation supervision system | |
| CN109615817A (en) | Monitoring system combines alarm method, device, electronic equipment and storage medium | |
| CN114973567B (en) | Automatic alarm method and terminal based on face recognition | |
| US20240312331A1 (en) | Methods and internet of things (iot) systems for gas safety warning based on gas determination scenarios | |
| CN111228706A (en) | GIS-based building interior fire fighting equipment control method and device | |
| CN110348585A (en) | A kind of power system monitor Information base event generates and update method | |
| CN113554834A (en) | Chemical industry park visual identification application system based on artificial intelligence technology | |
| CN116758490B (en) | Construction site safety helmet wearing detection and early warning method and system based on AI algorithm | |
| CN106713492A (en) | Garbage station relay monitoring system and monitoring method based on cloud computing | |
| CN115798139A (en) | Method and system for monitoring on-duty state of control room personnel | |
| CN116667536A (en) | Whole process management method, device and equipment for transformer substation monitoring information and storage medium | |
| KR102081577B1 (en) | Intelligence Fire Detecting System Using CCTV | |
| CN114999095B (en) | Building electrical fire monitoring method and system based on time and space fusion | |
| CN114495007A (en) | Substation fault point monitoring method and device | |
| CN115531787A (en) | Wisdom fire extinguishing system based on internet of things | |
| CN115346170A (en) | Intelligent monitoring method and device for gas facility area | |
| CN118379863B (en) | Safety monitoring and early warning method and device for stadium | |
| CN111882800A (en) | Fire-fighting early warning method and system based on multi-dimensional data linkage |
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 |