An abnormal work behavior monitoring method, device, medium and equipment

Abstract

The application relates to the technical field of construction safety monitoring, and particularly discloses an abnormal operation behavior monitoring method, device, medium and equipment. The method comprises the following steps: acquiring multi-source operation data of a construction worker in an operation process; based on image data in the multi-source operation data, a target behavior recognition model obtained through pre-training is used for abnormal operation behavior recognition to obtain a first recognition result; based on non-image data in the multi-source data, abnormal operation behavior recognition is performed to obtain a second recognition result; and based on the first recognition result and the second recognition result, a target monitoring result of abnormal operation behavior is obtained. The application can realize comprehensive monitoring of abnormal operation behavior by collecting multi-source operation data of a construction worker in an operation process, so that the final monitoring result is more comprehensive and accurate, and the monitoring efficiency of abnormal operation behavior is improved since manual inspection is not required.

Description

Technical Field

[0001] This application relates to the field of construction safety monitoring technology, specifically to a method, device, medium, and equipment for monitoring abnormal operational behavior. Background Technology

[0002] With the rapid development of the construction industry, construction scenarios are becoming increasingly complex, the number of workers is increasing, and safety accidents caused by violations of regulations are frequent, seriously threatening the lives of construction workers and the smooth progress of projects.

[0003] Currently, construction safety monitoring mainly relies on fixed monitoring equipment and manual inspections. However, fixed monitoring has problems such as limited coverage and inability to adapt to temporary work areas, while manual inspections have drawbacks such as low efficiency, easy to miss detections, and poor real-time performance.

[0004] Therefore, there is an urgent need for a method to monitor abnormal work behavior in order to solve the problem that existing technologies cannot quickly and accurately monitor abnormal construction behavior. Summary of the Invention

[0005] In view of this, this application provides a method, device, medium and equipment for monitoring abnormal work behavior, the main purpose of which is to solve the problem that it is currently impossible to monitor abnormal construction behavior quickly and accurately.

[0006] To address the above problems, this application provides a method for monitoring abnormal operational behavior, comprising: Acquire multi-source operational data of construction workers during the work process; Based on the image data in the multi-source operation data, a target behavior recognition model obtained through pre-training is used to identify abnormal operation behavior and obtain a first recognition result. Based on the non-image data in the multi-source data, abnormal operation behavior is identified to obtain a second identification result; Based on the first identification result and the second identification result, the target monitoring result of abnormal operation behavior is obtained.

[0007] Optionally, before using the pre-trained target behavior recognition model to identify abnormal work behavior and obtain the first recognition result, the method further includes: pre-training the target behavior recognition model, specifically including: Acquire several sample image data containing abnormal operational behaviors; Abnormal operational behaviors in each of the sample image data are labeled to obtain labeling information; Based on the sample image data and the corresponding annotation information, the initial convolutional neural network model is trained to obtain the target behavior recognition model.

[0008] Optionally, the non-image data in the multi-source data includes any one or more of the following: wearable data from wearable devices, sound data from construction equipment, gas data from the construction work space, and vibration data from construction equipment. The acquisition of multi-source work data of construction workers during the work process specifically includes: Based on visible light imaging devices and / or infrared thermal imaging devices, the construction workers' work process is captured in real time to obtain image data; Based on the target wearable device corresponding to the construction worker, acquire the wear data of the target wearable device; The sound data of the construction equipment is acquired based on the sound sensor installed on the construction equipment. Based on gas sensors installed in the construction work space, gas data of the construction work space is acquired; Vibration data of the construction equipment is acquired based on vibration sensors installed on the construction equipment. The image data, the wearable data of the target wearable device, the sound data of the construction equipment, the gas data of the construction work space, and the vibration data of the construction equipment are used as the multi-source operation data.

[0009] Optionally, the step of identifying abnormal work behavior based on non-image data from the multi-source data to obtain a second identification result specifically includes: Based on the wear data of the target wearable device, it is determined whether the target wearable device is being worn, in order to obtain the second identification result; And / or, calculate sound feature indicators based on the sound data of the construction equipment, and identify abnormal operation behavior based on the sound feature indicators and a predetermined sound indicator threshold to obtain the second identification result; And / or, based on the gas data of the construction work space, obtain the gas type and the gas concentration of the predetermined target gas type, and based on the gas type and the gas concentration of the target gas type, perform abnormal operation behavior identification to obtain the second identification result; And / or, calculate vibration characteristic indicators based on the vibration data of the construction equipment, and identify abnormal operation behavior based on the vibration characteristic indicators and a predetermined vibration indicator threshold to obtain the second identification result.

[0010] Optionally, obtaining the target monitoring result of abnormal operation behavior based on the first identification result and the second identification result specifically includes: When the first identification result indicates abnormal work behavior and / or the second identification result indicates abnormal work behavior, it is determined that the construction worker has abnormal work behavior, and the target monitoring result is obtained; When both the first identification result and the second identification result indicate that there is no abnormal work behavior, it is determined that the construction worker has not engaged in any abnormal work behavior, and the target monitoring result is obtained.

[0011] Optionally, when the target monitoring result indicates the presence of abnormal operational behavior, the method further includes: Obtain the movement trajectory of the construction workers and the duration of the violation; The hazard level of the abnormal work behavior is determined based on the movement trajectory and / or the duration. The target early warning method is determined based on the aforementioned hazard level; A warning signal is output based on the aforementioned target warning method.

[0012] Optionally, when the target monitoring result indicates the presence of abnormal operational behavior, the method further includes: Obtain the identity information of the construction workers; The identity information and the target monitoring results are associated and stored in a predetermined storage area, and the identity information and the target monitoring results are sent to a predetermined target device.

[0013] To address the aforementioned problems, this application provides an abnormal operation behavior monitoring device, comprising: The acquisition module is used to acquire multi-source work data of construction workers during the work process; The first identification module is used to identify abnormal work behavior based on the image data in the multi-source work data and to obtain a first identification result by using a pre-trained target behavior identification model. The second identification module is used to identify abnormal work behavior based on non-image data in the multi-source data and obtain a second identification result; The acquisition module is used to obtain target monitoring results of abnormal operation behavior based on the first identification result and the second identification result.

[0014] To address the aforementioned problems, this application provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for monitoring abnormal operating behavior.

[0015] To address the aforementioned problems, this application provides an electronic device, comprising at least a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of any of the above-described methods for monitoring abnormal operating behavior.

[0016] The method, apparatus, medium, and equipment for monitoring abnormal work behavior disclosed in this application can achieve comprehensive monitoring of abnormal work behavior by collecting multi-source work data from construction workers during the construction process, making the final monitoring results more comprehensive and accurate. Since manual inspection is no longer required, the monitoring efficiency of abnormal work behavior is improved.

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating an abnormal operation behavior monitoring method according to an embodiment of this application; Figure 2 This is a structural block diagram of an abnormal operation behavior monitoring device according to another embodiment of this application; Figure 3 This is a structural block diagram of an electronic device according to another embodiment of this application. Detailed Implementation

[0019] Various embodiments and features of this application are described herein with reference to the accompanying drawings.

[0020] It should be understood that various modifications can be made to the embodiments described herein. Therefore, the above description should not be considered as limiting, but merely as an example of embodiments. Other modifications within the scope and spirit of this application will be apparent to those skilled in the art.

[0021] The accompanying drawings, which are included in and form part of this specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0022] These and other features of this application will become apparent from the following description of preferred forms of embodiments given as non-limiting examples, with reference to the accompanying drawings.

[0023] It should also be understood that although this application has been described with reference to some specific examples, those skilled in the art can certainly implement many other equivalent forms of this application.

[0024] The above and other aspects, features and advantages of this application will become more apparent when taken in conjunction with the accompanying drawings and in view of the following detailed description.

[0025] Specific embodiments of this application are described thereafter with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of this application, which can be implemented in various ways. Well-known and / or repeated functions and structures are not described in detail to avoid unnecessary or redundant details that could obscure the application. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as the basis and representative basis for the claims to teach those skilled in the art to use this application in a variety of substantially any suitable detailed structures.

[0026] This specification may use the phrases “in one embodiment,” “in another embodiment,” “in yet another embodiment,” or “in other embodiments,” all of which may refer to one or more of the same or different embodiments according to this application.

[0027] This application provides a method for monitoring abnormal operating behavior, which can be specifically applied to electronic devices such as terminals and servers. Figure 1 As shown, the method in this embodiment includes the following steps: Step S101: Obtain multi-source operation data of construction workers during the operation process; In this step, multi-source operational data can include image data during the operation, wearable device data, sound data from construction equipment, gas data from the construction work space, and vibration data from construction equipment, etc.

[0028] Step S102: Based on the image data in the multi-source operation data, use a pre-trained target behavior recognition model to identify abnormal operation behavior and obtain a first recognition result; In the specific implementation process of this step, the initial convolutional neural network model can be trained in advance based on several sample image data containing abnormal work behavior to obtain a target behavior recognition model. Subsequently, the image data from the multi-source work data can be input into the target recognition model, and the target recognition model can be used to identify the abnormal work behavior in the image data to obtain the first recognition result.

[0029] Step S103: Based on the non-image data in the multi-source data, perform abnormal operation behavior identification to obtain a second identification result; In this step, non-image data includes one or more of the following: wearable device data, sound data from construction equipment, gas data from the work area, and vibration data from the construction equipment. Specifically, this step can determine whether construction workers are not wearing safety helmets, protective clothing, etc., based on wearable device data; determine if the construction equipment is malfunctioning based on sound data, thereby identifying any violations or abnormal work practices; determine if gas concentrations are within acceptable limits or if toxic gases are present based on gas data, further identifying any violations or abnormal work practices; and determine if the construction equipment is malfunctioning based on vibration data, thus identifying any violations or abnormal work practices.

[0030] Step S104: Based on the first identification result and the second identification result, obtain the target monitoring result of abnormal operation behavior.

[0031] In this step, after obtaining the first identification result and the second identification result, the two identification results can be combined to determine whether there is any abnormal behavior. That is, if the first identification result indicates abnormal work behavior and / or the second identification result indicates abnormal work behavior, it is determined that the construction worker has engaged in abnormal work behavior, and the target monitoring result is obtained; if the first identification result indicates no abnormal work behavior and the second identification result indicates no abnormal work behavior, it is determined that the construction worker has not engaged in abnormal work behavior, and the target monitoring result is obtained.

[0032] The abnormal work behavior monitoring device in this embodiment can achieve comprehensive monitoring of abnormal work behavior by collecting multi-source work data from construction workers during the construction process, making the final monitoring results more comprehensive and accurate. Since manual inspection is no longer required, the monitoring efficiency of abnormal work behavior is improved.

[0033] Another embodiment of this application provides a method for monitoring abnormal operational behavior, which specifically includes the following steps: Step S201: Obtain several sample image data containing abnormal operation behavior; Step S202: Annotate the abnormal work behaviors in each of the sample image data to obtain annotation information; Step S203: Based on each of the sample image data and the annotation information corresponding to each of the sample image data, train the initial convolutional neural network model to obtain the target behavior recognition model; Step S204: Based on a visible light camera and / or an infrared thermal imaging camera, the construction workers' work process is captured in real time to obtain image data; based on the target wearable device corresponding to the construction workers, wear data of the target wearable device is obtained; based on a sound sensor installed on the construction work equipment, sound data of the construction work equipment is obtained; based on a gas sensor installed in the construction work space, gas data of the construction work space is obtained; based on a vibration sensor installed on the construction work equipment, vibration data of the construction work equipment is obtained; the image data, wear data of the target wearable device, sound data of the construction work equipment, gas data of the construction work space, and vibration data of the construction work equipment are used as the multi-source work data; In the specific implementation of this step, surveillance cameras can be set up at the construction site to capture real-time images of the construction workers' work process and obtain image data.

[0034] Specifically, it can integrate multimodal camera and multi-sensor technologies, combined with dual-mode precise positioning, and through adaptive transmission technology, collect multi-dimensional data from all directions to ensure stable data transmission and comprehensive data collection in complex scenarios.

[0035] In this step, the location information of construction workers can also be obtained, and then the image data, sensor data, and positioning data are fused together, thereby eliminating the limitations of single data recognition.

[0036] Step S205: Based on the image data in the multi-source operation data, use a pre-trained target behavior recognition model to identify abnormal operation behavior and obtain a first recognition result; Step S206: Based on the wear data of the target wearable device, determine whether the target wearable device is being worn to obtain the second identification result; and / or, calculate sound feature indicators based on the sound data of the construction equipment, and perform abnormal operation behavior identification based on the sound feature indicators and a predetermined sound indicator threshold to obtain the second identification result; and / or, based on the gas data of the construction work space, obtain the gas type and the gas concentration of a predetermined target gas type, and perform abnormal operation behavior identification based on the gas type and the gas concentration of the target gas type to obtain the second identification result; and / or, based on the vibration data of the construction equipment, calculate vibration feature indicators, and perform abnormal operation behavior identification based on the vibration feature indicators and a predetermined vibration indicator threshold to obtain the second identification result. In this step, sound characteristic indicators can include any one or more of the following: noise peak value, short-time energy, zero-crossing rate, root mean square (RMS), spectral bandwidth, and spectral peak value. Vibration characteristic indicators can include any one or more of the following: vibration peak value, RMS value, mean, kurtosis, dominant frequency, frequency variance, and frequency band energy. Gas types can include carbon monoxide, oxygen, hydrogen sulfide, formaldehyde, etc. That is, if the gas types include toxic gases such as hydrogen sulfide and formaldehyde, then abnormal operational behavior can be determined.

[0037] Step S207: When the first identification result indicates abnormal work behavior and / or the second identification result indicates abnormal work behavior, determine that the construction worker has abnormal work behavior and obtain the target monitoring result; when the first identification result indicates no abnormal work behavior and the second identification result indicates no abnormal work behavior, determine that the construction worker has no abnormal work behavior and obtain the target monitoring result.

[0038] Step S208: Obtain the movement trajectory of the construction workers and the duration of the violation; determine the danger level of the abnormal operation based on the movement trajectory and / or the duration; determine the target early warning method based on the danger level; output an early warning signal based on the target early warning method; In this step, based on location data, the movement trajectory of violators can be tracked to analyze the duration and scope of the violation, and to determine whether there is a group violation. Simultaneously, the danger level / degree can be determined based on the length of the movement trajectory and the duration. That is, based on the degree of danger of the violation, it can be divided into three levels: emergency violation, serious violation, and general violation, and the time, location, and personnel information of the violation are automatically marked.

[0039] In this embodiment, a high-decibel alarm and a high-brightness warning light can be installed inside the surveillance sphere to trigger different warning modes / methods based on the level of violation.

[0040] Step S209: Obtain the identity information of the construction workers; associate and store the identity information and the target monitoring results in a predetermined storage area, and send the identity information and the target monitoring results to a predetermined target device.

[0041] In the specific implementation process, this step can integrate facial recognition and the electronic certificate database of construction personnel to compare the captured facial features with the personnel identity information registered in the background in real time.

[0042] In this step, warning information can be sent to the target equipment held by the management personnel via SMS, app push notifications, and platform pop-ups. Specifically, it can also be linked with equipment at the construction site to automatically trigger the corresponding equipment in the event of an emergency violation, preventing the accident from escalating.

[0043] In this step, when storing identity information and target monitoring results, a dual-storage architecture of local + cloud can be adopted. Local storage meets offline viewing needs, while the cloud stores all violation data, equipment operation data, and system logs using encryption algorithms. A data visualization engine can be set up to automatically count the frequency of violations in different areas, time periods, and types, and generate violation analysis reports and trend charts. It can respond to user retrieval operations for historical violation data, retrieving corresponding image and video evidence from predetermined storage locations to meet the needs of accident investigation, liability determination, and performance evaluation. It can respond to role-based permission configuration operations, configuring different data viewing and operation permissions for different roles. It can monitor the system / equipment operating status in real time, automatically diagnose equipment faults, and dynamically optimize algorithm model parameters based on violation recognition accuracy.

[0044] In this embodiment, the deployment location of the control ball and the real-time coordinates of the operators can be accurately located during implementation; the data encoding format can be dynamically adjusted according to the network bandwidth of the construction site, and 5G+WiFi dual-mode transmission can be adopted; the control system of the construction machinery can be connected through the Internet of Things protocol to collect equipment operating parameters; and when the equipment parameters exceed the safety threshold, they can be synchronously marked as equipment-related violation data and collected in conjunction with personnel violation data.

[0045] In the specific implementation process of this embodiment, a violation behavior trend prediction model can also be constructed to predict upcoming violations by analyzing the real-time movement trajectory of personnel and the frequency of violations in the area, and to trigger a mild warning in advance.

[0046] The method in this embodiment is adaptable to various construction environments. By integrating multimodal camera and multi-sensor technologies, combined with dual-mode precise positioning, it can collect multi-dimensional data from all angles, ensuring stable data transmission and comprehensive data collection in complex scenarios. A recognition model is constructed based on deep learning algorithms and multi-source data is integrated to accurately identify various construction violations. Simultaneously, the violation trajectory is tracked, the impact range is analyzed, and violation levels are intelligently classified, achieving efficient identification and scientific grading of violations. Differentiated warnings are initiated based on the violation level, linking local audio-visual alerts with remote multi-channel reminders, and simultaneously connecting to on-site emergency equipment for rapid response to risks and hazards. Violation evidence is automatically retained, forming a closed-loop execution chain of warning-handling-evidence collection. A dual-storage architecture is used to manage massive amounts of data, and decision-making basis is generated through statistical analysis. Multi-dimensional historical data traceability is supported, along with multi-role access control, system status monitoring, and self-optimization, contributing to refined safety management.

[0047] Another embodiment of this application provides an abnormal operation behavior monitoring device, such as... Figure 2 As shown, it includes: Module 11 is used to acquire multi-source work data of construction workers during the work process; The first identification module 12 is used to identify abnormal work behavior based on the image data in the multi-source work data and to obtain a first identification result by using a pre-trained target behavior identification model. The second identification module 13 is used to identify abnormal work behavior based on non-image data in the multi-source data and obtain a second identification result; The module 14 is used to obtain the target monitoring results of abnormal operation behavior based on the first identification result and the second identification result.

[0048] In this embodiment, the abnormal operation behavior monitoring device further includes a training module for pre-training the target behavior recognition model. The training module is specifically used for: acquiring several sample image data containing abnormal operation behaviors; labeling the abnormal operation behaviors in each sample image data to obtain labeling information; and training an initial convolutional neural network model based on each sample image data and the labeling information corresponding to each sample image data to obtain the target behavior recognition model.

[0049] In this embodiment, the non-image data in the multi-source data includes any one or more of the following: wearable device data, sound data of construction equipment, gas data of the construction work space, and vibration data of construction equipment. The acquisition module is specifically used for: capturing real-time images of the construction workers' work process using a visible light camera and / or an infrared thermal imaging camera to obtain image data; acquiring wear data of the target wearable device corresponding to the construction workers; acquiring sound data of the construction work equipment based on a sound sensor installed on the construction work equipment; acquiring gas data of the construction work space based on a gas sensor installed in the construction work space; acquiring vibration data of the construction work equipment based on a vibration sensor installed on the construction work equipment; and using the image data, wear data of the target wearable device, sound data of the construction work equipment, gas data of the construction work space, and vibration data of the construction work equipment as the multi-source work data.

[0050] In this embodiment, the second identification module is specifically used for: determining whether the target wearable device is being worn based on the wear data of the target wearable device, thereby obtaining the second identification result; and / or calculating sound feature indicators based on the sound data of the construction equipment, and identifying abnormal work behavior based on the sound feature indicators and a predetermined sound indicator threshold, thereby obtaining the second identification result; and / or obtaining gas types and the gas concentration of a predetermined target gas type based on the gas data of the construction work space, and identifying abnormal work behavior based on the gas types and the gas concentration of the target gas type, thereby obtaining the second identification result; and / or calculating vibration feature indicators based on the vibration data of the construction equipment, and identifying abnormal work behavior based on the vibration feature indicators and a predetermined vibration indicator threshold, thereby obtaining the second identification result.

[0051] In this embodiment, the obtaining module is specifically used to: determine that the construction worker has abnormal work behavior when the first identification result is that there is abnormal work behavior and / or the second identification result is that there is abnormal work behavior, and obtain the target monitoring result; and determine that the construction worker has no abnormal work behavior when the first identification result is that there is no abnormal work behavior and the second identification result is that there is no abnormal work behavior, and obtain the target monitoring result.

[0052] In this embodiment, the abnormal work behavior monitoring device further includes an early warning module. The early warning module is used to: acquire the movement trajectory of the construction worker and the duration of the violation when the target monitoring result indicates the existence of abnormal work behavior; determine the danger level of the abnormal work behavior based on the movement trajectory and / or the duration; determine the target early warning method based on the danger level; and output an early warning signal based on the target early warning method.

[0053] In this embodiment, the abnormal work behavior monitoring device further includes a storage module, which is used to: obtain the identity information of the construction worker when the target monitoring result indicates the existence of abnormal work behavior; associate and store the identity information and the target monitoring result in a predetermined storage area; and send the identity information and the target monitoring result to a predetermined target device.

[0054] The abnormal work behavior monitoring device in this application can achieve comprehensive monitoring of abnormal work behavior by collecting multi-source work data of construction workers during the construction process, making the final monitoring results more comprehensive and accurate. Since manual inspection is no longer required, the monitoring efficiency of abnormal work behavior is improved.

[0055] Another embodiment of this application provides a storage medium storing a computer program, which, when executed by a processor, implements the following method steps: Step 1: Obtain multi-source operation data of construction workers during the operation process; Step 2: Based on the image data in the multi-source operation data, use the pre-trained target behavior recognition model to identify abnormal operation behavior and obtain the first recognition result; Step 3: Based on the non-image data in the multi-source data, perform abnormal operation behavior identification to obtain a second identification result; Step 4: Based on the first identification result and the second identification result, obtain the target monitoring result of abnormal operation behavior.

[0056] The specific implementation process of the above method steps can be found in the embodiments of the above-mentioned abnormal operation behavior monitoring methods, which will not be repeated here.

[0057] The storage medium in this application can achieve comprehensive monitoring of abnormal work behavior by collecting multi-source work data from construction workers during the construction process, making the final monitoring results more comprehensive and accurate. Since manual inspection is no longer required, the monitoring efficiency of abnormal work behavior is improved.

[0058] Another embodiment of this application provides an electronic device, such as... Figure 3 As shown, it includes at least a memory 1 and a processor 2. The memory 1 stores a computer program, and the processor 2 performs the following method steps when executing the computer program in the memory 1: Step 1: Obtain multi-source operation data of construction workers during the operation process; Step 2: Based on the image data in the multi-source operation data, use the pre-trained target behavior recognition model to identify abnormal operation behavior and obtain the first recognition result; Step 3: Based on the non-image data in the multi-source data, perform abnormal operation behavior identification to obtain a second identification result; Step 4: Based on the first identification result and the second identification result, obtain the target monitoring result of abnormal operation behavior.

[0059] The specific implementation process of the above method steps can be found in the embodiments of the above-mentioned abnormal operation behavior monitoring methods, which will not be repeated here.

[0060] The electronic device in this application can achieve comprehensive monitoring of abnormal work behavior by collecting multi-source work data from construction workers during the construction process, making the final monitoring results more comprehensive and accurate. Since manual inspection is no longer required, the monitoring efficiency of abnormal work behavior is improved.

[0061] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.

Claims

1. A method for monitoring abnormal work behavior, characterized in that, include: Acquire multi-source operational data of construction workers during the work process; Based on the image data in the multi-source operation data, a target behavior recognition model obtained through pre-training is used to identify abnormal operation behavior and obtain a first recognition result. Based on the non-image data in the multi-source data, abnormal operation behavior is identified to obtain a second identification result; Based on the first identification result and the second identification result, the target monitoring result of abnormal operation behavior is obtained.

2. The method for monitoring abnormal work behavior as described in claim 1, characterized in that, Before using a pre-trained target behavior recognition model to identify abnormal work behavior and obtain a first recognition result, the method further includes: pre-training the target behavior recognition model, specifically including: Acquire several sample image data containing abnormal operational behaviors; Abnormal operational behaviors in each of the sample image data are labeled to obtain labeling information; Based on the sample image data and the corresponding annotation information, the initial convolutional neural network model is trained to obtain the target behavior recognition model.

3. The method for monitoring abnormal work behavior as described in claim 1, characterized in that, The non-image data in the multi-source data includes any one or more of the following: wearable data from wearable devices, sound data from construction equipment, gas data from the construction work space, and vibration data from construction equipment. The acquisition of multi-source work data of construction workers during the work process specifically includes: Based on visible light imaging devices and / or infrared thermal imaging devices, the construction workers' work process is captured in real time to obtain image data; Based on the target wearable device corresponding to the construction worker, acquire the wear data of the target wearable device; The sound data of the construction equipment is acquired based on the sound sensor installed on the construction equipment. Based on gas sensors installed in the construction work space, gas data of the construction work space is acquired; Vibration data of the construction equipment is acquired based on vibration sensors installed on the construction equipment. The image data, the wearable data of the target wearable device, the sound data of the construction equipment, the gas data of the construction work space, and the vibration data of the construction equipment are used as the multi-source operation data.

4. The method for monitoring abnormal work behavior as described in claim 3, characterized in that, The step of identifying abnormal work behavior based on non-image data from the multi-source data to obtain a second identification result specifically includes: Based on the wear data of the target wearable device, it is determined whether the target wearable device is being worn, in order to obtain the second identification result; And / or, calculate sound feature indicators based on the sound data of the construction equipment, and identify abnormal operation behavior based on the sound feature indicators and a predetermined sound indicator threshold to obtain the second identification result; And / or, based on the gas data of the construction work space, obtain the gas type and the gas concentration of the predetermined target gas type, and based on the gas type and the gas concentration of the target gas type, perform abnormal operation behavior identification to obtain the second identification result; And / or, calculate vibration characteristic indicators based on the vibration data of the construction equipment, and identify abnormal operation behavior based on the vibration characteristic indicators and a predetermined vibration indicator threshold to obtain the second identification result.

5. The method for monitoring abnormal work behavior as described in claim 1, characterized in that, The step of obtaining target monitoring results for abnormal operational behavior based on the first identification result and the second identification result specifically includes: When the first identification result indicates abnormal work behavior and / or the second identification result indicates abnormal work behavior, it is determined that the construction worker has abnormal work behavior, and the target monitoring result is obtained; When both the first identification result and the second identification result indicate that there is no abnormal work behavior, it is determined that the construction worker has not engaged in any abnormal work behavior, and the target monitoring result is obtained.

6. The method for monitoring abnormal work behavior as described in claim 1, characterized in that, When the target monitoring result indicates the presence of abnormal operational behavior, the method further includes: Obtain the movement trajectory of the construction workers and the duration of the violation; The hazard level of the abnormal work behavior is determined based on the movement trajectory and / or the duration. The target early warning method is determined based on the aforementioned hazard level; A warning signal is output based on the aforementioned target warning method.

7. A method for monitoring abnormal work behavior as described in any one of claims 1-6, characterized in that, When the target monitoring result indicates the presence of abnormal operational behavior, the method further includes: Obtain the identity information of the construction workers; The identity information and the target monitoring results are associated and stored in a predetermined storage area, and the identity information and the target monitoring results are sent to a predetermined target device.

8. A device for monitoring abnormal work behavior, characterized in that, include: The acquisition module is used to acquire multi-source work data of construction workers during the work process; The first identification module is used to identify abnormal work behavior based on the image data in the multi-source work data and to obtain a first identification result by using a pre-trained target behavior identification model. The second identification module is used to identify abnormal work behavior based on non-image data in the multi-source data and obtain a second identification result; The acquisition module is used to obtain target monitoring results of abnormal operation behavior based on the first identification result and the second identification result.

9. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the abnormal operation behavior monitoring method according to any one of claims 1-7.

10. An electronic device, characterized in that, It includes at least a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of the abnormal operation behavior monitoring method according to any one of claims 1-7.

Images