Water conservancy building construction site panoramic monitoring and illegal behavior recognition system and method

Abstract

The invention discloses a panoramic monitoring and illegal behavior recognition system and method for a water conservancy building construction site. The method comprises the steps that a data processing box calls a preset typical illegal construction behavior database and a typical illegal equipment image database; the method comprises the following steps: acquiring construction personnel behavior images, identifying and identifying the acquired construction personnel behavior images and a typical illegal construction behavior database, comparing construction equipment behavior images with a typical illegal equipment image database, and sending alarm information to a server side when the acquired construction personnel behavior images and the construction equipment behavior images are identified and judged to have illegal behaviors; the intercepted violation picture and the basic attribute information of the construction personnel and the construction equipment are obtained and sent to a server side; according to the method, the recognition result of the illegal behavior displayed on the interface, the heat map of the risk area and the like can be monitored in real time, construction site managers are assisted in quick response, and a powerful guarantee is provided for safe operation of site building construction.

Description

Technical Field

[0001] The present invention relates to the field of water conservancy construction safety, and particularly to a panoramic monitoring and illegal behavior recognition system and method for a water conservancy construction site. Background Art

[0002] At present, the safety supervision of construction sites has always been a key concern in the industry. With the continuous acceleration of the urbanization process and the concentrated development of large-scale construction projects, construction sites are often densely populated, with a large number of equipment, and the construction environment is complex and changeable, increasing the potential for various safety accidents. The traditional on-site supervision method mainly relies on video collection by surveillance cameras and manual inspections by security personnel. However, this method not only has deficiencies such as monitoring blind spots, fatigue in manual inspections, and delays in video playback, but also it is difficult to comprehensively and real-time capture and judge various behaviors in large-scale construction sites. Especially when construction workers are working in high-risk areas such as high-altitude operations and deep foundation pit construction, whether their behaviors comply with safety regulations is often difficult to be discovered and noticed in the first time, thus leaving huge safety hazards.

[0003] In recent years, with the rapid development of computer vision, artificial intelligence, and deep learning technologies, safety supervision systems based on image recognition and object detection have gradually become a research and application hotspot. Scholars and engineering and technical personnel at home and abroad have tried to use deep learning methods such as convolutional neural networks (CNNs) to implement the behavior recognition of construction workers in surveillance images by constructing object detection models (such as algorithms like Faster R-CNN, YOLO, SSD, etc.). Some studies have shown that using deep learning technology can, to a certain extent, overcome the deficiencies of traditional image processing methods in harsh environments such as light changes, image noise, and partial occlusion, and improve the accuracy and real-time response ability of illegal behavior detection. However, there are still some obvious problems and deficiencies in the existing technologies, such as insufficient sample data, inaccurate data annotation, poor model robustness, and a relatively high false positive rate in complex construction scenarios, etc., and there is an urgent need to further improve the performance of the automatic recognition system. Summary of the Invention

[0004] The purpose of the present invention is to provide a panoramic monitoring and illegal behavior recognition system and method for a water conservancy construction site, which solves the above-mentioned technical problems pointed out in the existing technologies.

[0005] The present invention proposes a panoramic monitoring and illegal behavior recognition system for a water conservancy construction site, including: high-definition surveillance cameras, fence warning surveillance cameras, and first and second sensor devices, a server side, and a data processing box:

[0006] The high-definition surveillance camera is used to collect images of the behaviors of construction workers and construction equipment at the construction site, obtaining construction worker behavior images and construction equipment behavior images; the high-definition surveillance camera is also connected to the data processing box through a real-time video stream, and the data processing box calls a preset database of typical illegal construction behavior and a database of typical illegal equipment images; and it identifies and appraises the collected construction worker behavior images with the database of typical illegal construction behavior, compares the construction equipment behavior images with the database of typical illegal equipment images, and when it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, it sends an alarm message to the server side and sends the basic attribute information of the current construction workers and construction equipment to the server side; among them, the typical illegal behaviors in the database of typical illegal construction behavior include that construction workers do not wear safety helmets, do not wear reflective vests, do not fasten seat belts, smoke, and doze off; among them, the typical illegal equipment behaviors in the database of typical illegal equipment images include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs.

[0007] Based on the second sensor device, a prohibitive electronic fence is constructed. The second sensor device is an infrared sensor, which is used to send an alarm message to the server side when it detects that there are construction workers and equipment entering the current electronic fence.

[0008] Based on the second sensor device and the fence warning surveillance camera, a warning electronic fence is constructed. The fence warning surveillance camera is used to collect images of the current warning electronic fence, and compare the collected images with a preset image database to identify and judge whether the construction workers and construction equipment entering the construction area currently are illegal personnel. If the judgment is yes, an alarm message is sent to the server side.

[0009] It also includes an illegal behavior sampling module; the illegal behavior sampling module is used to intercept the collected construction worker behavior images and construction equipment behavior images with illegal behaviors, obtain the intercepted illegal pictures and the basic attribute information of the construction workers and construction equipment, and send them to the server side.

[0010] It should be noted that the high-definition surveillance camera is connected to the data processing box through a real-time video stream, and the data processing box automatically analyzes the behaviors of construction workers and the states of equipment in the video images based on a pre-trained computer vision model (i.e., including a database).

[0011] The above warning electronic fence identifies construction workers and construction equipment within the boundary range of the construction site through a high-definition surveillance camera; based on the preset electronic fence rules, it judges whether there are unauthorized personnel or construction equipment entering the construction site; it issues an alarm to the illegally entering personnel or equipment and notifies relevant management personnel in real time through the management platform.

[0012] The above attendance management module combines face recognition processing and device information scanning processing to achieve intelligent attendance management for the entry and exit of construction workers and construction equipment;

[0013] The above image data storage module is used to store the image data collected by the video surveillance equipment as the claim basis during the construction process and the image data for later review; reasonably arrange high-definition surveillance cameras and the first sensor devices to ensure a full-coverage monitoring network at the construction site, and based on video image recognition technology, conduct real-time analysis on the behavior of personnel and the status of equipment at the construction site; once a violation is detected, the platform automatically records the violation, intercepts the violation pictures, and sends the violation information to the management terminal for the convenience of management personnel to handle in a timely manner.

[0014] Preferably, as an implementable solution; it further includes an attendance management module: the attendance management module is used to implement monitoring and attendance processing of the construction workers entering and leaving the site by combining face recognition processing; the attendance management module is also used to implement monitoring processing of the construction equipment entering and leaving the site by combining the device information scanning processing method.

[0015] Preferably, as an implementable solution; the attendance management module includes a first attendance management sub-module and a second attendance management sub-module:

[0016] The first attendance management sub-module is used to collect the facial features of construction workers through high-definition surveillance cameras and compare them with the preset facial feature database of construction workers to confirm the identity information of construction workers; the identity information includes name information and job number information;

[0017] The second attendance management sub-module is used to collect the unique identification information of construction equipment through device information scanning technology, compare it with the device management database through the unique identification information to confirm the entry and exit status of the equipment; it is also used to obtain the entry and exit records of construction workers and construction equipment, and then generate automated attendance data according to the entry and exit records of construction workers and construction equipment and store it, or send it to the server side.

[0018] Preferably, as an implementable solution; the first sensor device is used to collect the environmental parameters of the construction site, and the environmental parameters include environmental temperature, environmental humidity, dust concentration, noise level, and rainfall; it further includes a safety assessment module, and the safety assessment module is used to comprehensively evaluate the safety evaluation score of the construction site according to the rainfall and the foggy pictures of the construction environment collected by the high-definition surveillance camera; and send the safety evaluation score to the server side.

[0019] Preferably, as an implementable solution; the safety assessment module includes a calculation processing module and an analysis processing module;

[0020] The calculation processing module is specifically used to obtain the normalized rainfall R and fog index F; construct a safety evaluation score based on the rainfall R and fog index F; the safety evaluation score formula is expressed as S=100×[1-(w R ·R+w F ·F)]; 100 represents full score; R is normalized rainfall; F is the fog index obtained by image processing algorithm, 0 represents no fog at all, 1 represents the most severe fog); w R is the impact weight of rainfall on safety assessment; w F is the influence weight of fog index on safety assessment, and it is required that w R +w F ≤1;

[0021] The analysis and processing module is used to determine that there is no rainfall and no fog when the rainfall R and fog index F are both 0, and output the safety evaluation score S=100; and is also used to determine that there is no rainfall and no fog when the rainfall R and fog index F are both 0. R ·R+w F The larger F is, the worse the construction site environment is and the lower the safety evaluation score S is.

[0022] Preferably, as an implementable solution; the violation sampling module is specifically used to capture video screenshots when violations occur; record the time, location, and basic attribute information of construction personnel and construction equipment when the violations occur; generate a violation report based on the time, location, and basic attribute information of construction personnel and construction equipment, and push it to the construction site management terminal or server.

[0023] Preferably, as an implementable solution, the high-definition surveillance camera and the fence warning surveillance camera also include an image data storage module; the image data storage module supports segmented storage of long-term video data, supports playback function, and provides a rapid positioning function for violations.

[0024] Preferably, as an implementable embodiment, it also includes multiple management terminals that establish cloud service communication connections with the server; the management terminals perform two-way signal call control with the server, and the management terminals are used to view in real time the images captured by the high-definition surveillance cameras at the construction site, the fence warning surveillance cameras, and the data, historical records and violation reports collected by the first sensor device and the second sensor device.

[0025] Preferably, as an implementable solution, it further includes an environmental parameter evaluation module; the environmental parameter evaluation module is used to determine the environmental index E of the construction site according to the environmental parameters collected by the first sensor device: E = w1T + w2H + w3D + w4N + w5R, where: T is the environmental temperature, H is the environmental humidity, D is the dust concentration, N is the noise level, R is the rainfall; w1, w2, w3, w4, w5 are the weight coefficients of each parameter; calculate the environmental index E through the formula, and associate the environmental index E with the violation record; if the environmental index E is lower than the environmental health score threshold, trigger a global safety alarm and send an alarm message to the server side.

[0026] Correspondingly, the present invention provides a panoramic monitoring and violation behavior recognition method for a water conservancy construction site, including the following steps:

[0027] Arrange fence warning monitoring cameras, high-definition monitoring cameras, first sensor devices and second sensor devices at the water conservancy construction site, and merge and load the positions of the fence warning monitoring cameras, high-definition monitoring cameras, first sensor devices and second sensor devices with the construction site topographic map and the construction site building map into a preset 3D BIM model for binding to realize real-time monitoring of construction operations and environmental conditions;

[0028] Install high-definition monitoring cameras for behavior recognition in the construction area; the high-definition monitoring cameras are used to collect images of the behaviors of construction workers and construction equipment at the construction site to obtain construction worker behavior images and construction equipment behavior images; the high-definition monitoring cameras are connected to the data processing box through a real-time video stream, and the data processing box calls a preset database of typical illegal construction behavior and a database of typical illegal equipment images; and identify and authenticate the collected construction worker behavior images with the database of typical illegal construction behavior, compare the construction equipment behavior images with the database of typical illegal equipment images, and when it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, send an alarm message to the server side and send the basic attribute information of the current construction workers and construction equipment to the server side; among them, the typical illegal behaviors in the database of typical illegal construction behavior include that construction workers do not wear safety helmets, do not wear reflective vests, do not fasten seat belts, smoke, and doze off; among them, the typical illegal equipment behaviors in the database of typical illegal equipment images include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs;

[0029] Construct a prohibitive electronic fence based on the second sensor device. The second sensor device is an infrared sensor, and sends an alarm message to the server side when it detects that there are construction workers and equipment entering the current electronic fence;

[0030] Construct a vigilant electronic enclosure based on the second sensor device and the fence warning monitoring camera. The fence warning monitoring camera collects images of the current vigilant electronic enclosure, compares the collected images with a preset image database, and identifies and determines whether the construction personnel and construction equipment entering the construction area are illegal personnel. If the determination is yes, an alarm message is sent to the server side.

[0031] The violation behavior sampling module intercepts the behavior images of construction personnel and construction equipment with violation behaviors collected, obtains the intercepted violation pictures and the basic attribute information of the construction personnel and construction equipment, and sends them to the server side.

[0032] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0033] Analyzing the above technical solutions, it can be seen that the present invention provides a method for panoramic monitoring and violation behavior identification in a water conservancy construction site, including the following steps: Arrange a fence warning monitoring camera, a high-definition monitoring camera, a first sensor device and a second sensor device in the water conservancy construction site, and merge and load the positions of the fence warning monitoring camera, the high-definition monitoring camera, the first sensor device and the second sensor device with the topographic map of the construction site and the building map of the construction site into a preset 3D BIM model for binding to realize real-time monitoring of construction operations and environmental conditions.

[0034] Install a high-definition monitoring camera for behavior recognition at the construction area. The high-definition monitoring camera is used to collect images of the behaviors of construction personnel and construction equipment at the construction site to obtain construction personnel behavior images and construction equipment behavior images. The high-definition monitoring camera is connected to a data processing box through a real-time video stream. The data processing box calls a preset typical illegal construction behavior database and a typical illegal equipment image database, identifies and determines the collected construction personnel behavior images with the typical illegal construction behavior database, compares the construction equipment behavior images with the typical illegal equipment image database, and when it is identified and determined that there are illegal behaviors in the collected construction personnel behavior images and construction equipment behavior images, an alarm message is sent to the server side, and the basic attribute information of the current construction personnel and construction equipment is sent to the server side. Among them, the typical illegal behaviors in the typical illegal construction behavior database include that construction personnel do not wear safety helmets, do not wear reflective vests, do not fasten seat belts, smoke, and doze off. Among them, the typical illegal equipment behaviors in the typical illegal equipment image database include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing a prohibited construction area, and construction equipment not installing warning signs.

[0035] Build a prohibitive electronic enclosure based on the second sensor device, where the second sensor device is an infrared sensor, and send an alarm message to the server when it detects that construction workers and equipment have entered the current electronic fence; build a warning electronic enclosure based on the second sensor device and the fence warning monitoring camera. The fence warning monitoring camera collects images of the current warning electronic enclosure, and compares the collected images with a preset image database to identify and determine whether the construction workers and construction equipment entering the construction area are illegal personnel. If the determination is yes, an alarm message is sent to the server; the violation behavior sampling module intercepts the behavior images of construction workers and construction equipment with violation behaviors, obtains the intercepted violation pictures and the basic attribute information of construction workers and construction equipment, and sends them to the server.

[0036] Analyzing the above technical solution, it can be seen that: The present invention provides a panoramic monitoring and violation behavior identification system for a water conservancy construction site, which is a platform information system for intelligent construction site construction, and widely uses video monitoring, image recognition and sensor technologies; various intelligent cameras and monitoring devices are arranged in the on-site construction area, so as to ensure that a non-blind spot monitoring network is woven at the construction site. In the video monitoring management module of the management platform, the three-dimensional model is bound to the positions of the cameras and environmental monitoring devices one by one, and the cameras and monitoring devices are numbered. The above system closely focuses on the safety management, violation behavior identification, environmental monitoring and intelligent attendance management of the construction site, and forms an overall multi-dimensional intelligent construction site monitoring and management system, significantly improving the safety and management efficiency of the construction site. Brief Description of the Drawings

[0037] In order to more clearly illustrate the specific implementation manners of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required to be used in the description of the specific implementation manners or the prior art. Obviously, the following drawings are some implementation manners of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0038] Figure 1 It is a schematic diagram of the architecture principle of a panoramic monitoring and violation behavior identification system for a water conservancy construction site;

[0039] Figure 2 It is a schematic diagram of the information monitoring interface in a panoramic monitoring and violation behavior identification system for a water conservancy construction site;

[0040] Figure 3 It is a schematic diagram of the main process of a panoramic monitoring and violation behavior identification method for a water conservancy construction site.

[0041] Labels: High-definition surveillance camera 10; Fence warning surveillance camera 20; First sensor device 30; Second sensor device 40; Server side 50; Data processing box 60; Illegal behavior sampling module 70. Detailed implementation

[0042] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] The present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0044] As Figure 1 shown, the present invention provides a panoramic surveillance and illegal behavior identification system for a water conservancy construction site, including: a high-definition surveillance camera 10, a fence warning surveillance camera 20, a first sensor device 30, a second sensor device 40, a server side 50, and a data processing box 60:

[0045] The high-definition surveillance camera 10 is used to collect images of the behaviors of construction workers and construction equipment at the construction site to obtain construction worker behavior images and construction equipment behavior images; the high-definition surveillance camera is also connected to the data processing box 60 through a real-time video stream. The data processing box 60 calls a preset typical illegal construction behavior database and a typical illegal equipment image database; and identifies and appraises the collected construction worker behavior images with the typical illegal construction behavior database, compares the construction equipment behavior images with the typical illegal equipment image database. When it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, an alarm message is sent to the server side 50, and the basic attribute information of the current construction workers and construction equipment is sent to the server side 50; among them, the typical illegal behaviors in the typical illegal construction behavior database include that construction workers do not wear safety helmets, do not wear reflective vests, do not fasten seat belts, smoke, and doze; among them, the typical illegal equipment behaviors in the typical illegal equipment image database include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs;

[0046] Based on the second sensor device 40, a prohibitive electronic fence is constructed. The second sensor device is an infrared sensor, which is used to send an alarm message to the server side 50 when it detects that there are construction workers and equipment entering the current electronic fence;

[0047] Construct a warning electronic enclosure based on the second sensor device 40 and the fence warning monitoring camera 20. The fence warning monitoring camera is used to collect images of the current warning electronic enclosure, compare the collected images with a preset image database, identify and determine whether the construction personnel and construction equipment entering the construction area are illegal personnel, and if the determination is yes, send an alarm message to the server side 50;

[0048] It further includes a violation behavior sampling module 70; the violation behavior sampling module 70 is used to intercept the behavior images of construction personnel and construction equipment with violation behaviors, obtain the intercepted violation pictures and the basic attribute information of the construction personnel and construction equipment, and send them to the server side.

[0049] It should be noted that the above high-definition monitoring camera is connected to the data processing box through a real-time video stream. The data processing box automatically analyzes the behaviors of construction personnel and the states of equipment in the video image based on a pre-trained computer vision model (i.e., including a database).

[0050] The above warning electronic enclosure identifies construction personnel and construction equipment within the boundary of the construction site through a high-definition monitoring camera; based on preset electronic enclosure rules, determines whether there are unauthorized personnel or construction equipment entering the construction site; issues an alarm for illegally entering personnel or equipment, and notifies relevant management personnel in real time through the management platform.

[0051] The above attendance management module combines face recognition processing and device information scanning processing to realize intelligent attendance management for the entry and exit of construction personnel and construction equipment;

[0052] The above image data storage module is used to store the image data collected by the video monitoring device as the claim basis during the construction process and the image data for later review; reasonably arrange high-definition monitoring cameras and the first sensor device to ensure a full-coverage monitoring network at the construction site, and based on video image recognition technology, conduct real-time analysis on the behaviors of personnel and the states of equipment at the construction site; once a violation behavior is detected, automatically record the violation behavior through the platform, intercept the violation pictures, and send the violation information to the management terminal for the management personnel to handle in a timely manner.

[0053] The present invention provides a panoramic monitoring and violation behavior recognition system for a water conservancy construction site, which is a platform information system for intelligent construction site construction, and widely uses video monitoring, image recognition, and sensor technologies;

[0054] When the panoramic monitoring and illegal behavior identification system for water conservancy construction sites provided by the present invention is built, various intelligent cameras and monitoring devices need to be arranged in the on-site construction area, so as to ensure that a surveillance network without dead angles is woven at the construction site. In the video surveillance management module of the management platform, the three-dimensional model is corresponding bound with the positions of the cameras and environmental monitoring devices one by one, and the cameras and monitoring devices are numbered. The management personnel only need to select the corresponding module on the model to view the construction progress, as well as the environmental and weather changes during the construction process, providing a true and effective basis for construction project claims during the construction process and the retention of construction image data in the later stage.

[0055] At the same time, an electronic fence is constructed for the construction area. When unauthorized personnel and equipment enter the area, a warning procedure will be triggered, and a warning message will be broadcast through the built-in broadcaster to expel and warn the illegal intruders. Relevant construction personnel and equipment enter the construction area by swiping their faces or scanning device information at the pedestrian and vehicle access gates. By connecting the access gates with the terminal digital management platform, the management personnel can conduct attendance management for the entry and exit of laborers on the platform, so as to improve labor efficiency and reduce labor risks.

[0056] The above-mentioned electronic fence includes (prohibitive electronic fence and warning electronic fence); the prohibitive electronic fence detects personnel or equipment illegally entering the construction area through infrared sensors, triggers the alarm procedure, and effectively prevents non-related personnel from entering the construction site. The warning electronic fence collects images through cameras and compares them with a preset database, can identify illegally entering personnel or equipment, and send alarm information to the management platform in a timely manner. A warning message is sent through the broadcaster to further strengthen the warning effect.

[0057] Preferably, as an implementable solution; it further includes an attendance management module: the attendance management module is used to implement monitoring and attendance processing of construction workers entering and leaving the site by combining face recognition processing; the attendance management module is also used to implement monitoring processing of construction equipment entering and leaving the site by combining device information scanning processing methods.

[0058] During the project construction process, after relevant personnel enter the construction area, the cameras will feedback the intercepted pictures to the AI box for illegal behavior identification. If illegal behaviors (such as not wearing safety helmets, reflective vests, safety belts, smoking, dozing off, etc.) are determined to appear in the pictures, the platform terminal will automatically intercept the illegal pictures and send them to the on-site management personnel for rectification and education through the platform. At the same time, a warning message is broadcast through the broadcaster in the site area to remind the illegal personnel to stop their illegal behaviors.

[0059] Preferably, as an implementable solution; the attendance management module includes a first attendance management sub-module and a second attendance management sub-module:

[0060] The first attendance management sub-module is used to collect the facial features of construction workers through a high-definition surveillance camera, compare them with the pre-set facial feature database of construction workers, and confirm the identity information of construction workers; the identity information includes name information and job number information;

[0061] The second attendance management sub-module is used to collect the unique identification information of construction equipment through equipment information scanning technology, compare it with the equipment management database through the unique identification information, and confirm the entry and exit status of the equipment; it is also used to obtain the entry and exit records of construction workers and construction equipment, and then generate and store automated attendance data according to the entry and exit records of construction workers and construction equipment, or send it to the server side.

[0062] The attendance management module combines face recognition technology and equipment information scanning technology to realize intelligent attendance management for the entry and exit of construction workers and equipment, reduce the workload of manual management, and improve the efficiency and accuracy of attendance management. At the same time, it also has the technical effects of reducing omissions in the attendance management of personnel and equipment and reducing employment risks.

[0063] Preferably, as an implementable solution; the first sensor device is used to collect the environmental parameters of the construction site, and the environmental parameters include environmental temperature, environmental humidity, dust concentration, noise level, and rainfall; it also includes a safety assessment module, and the safety assessment module is used to comprehensively evaluate the safety evaluation score of the construction site according to the rainfall and the foggy pictures of the construction environment collected by the high-definition surveillance camera; and send the safety evaluation score to the server side.

[0064] Preferably, as an implementable solution; the safety assessment module includes a calculation processing module and an analysis processing module;

[0065] The calculation processing module is specifically used to obtain the normalized rainfall R and fog index F; construct a safety evaluation score based on the rainfall R and fog index F; the safety evaluation score formula is expressed as S = 100×[1-(w R ·R + w F ·F)]; where 100 represents the full score; R is the normalized rainfall; F is the fog index obtained by the image processing algorithm, 0 means completely fog-free, 1 means the foggiest); w R is the influence weight of rainfall on safety evaluation; w F is the influence weight of the fog index on safety evaluation, and it is required that w R + w F ≤1;

[0066] The analysis processing module is used to determine that there is no rainfall and no fog when both the rainfall R and the fog index F are 0, and output the safety evaluation score S = 100; it is also used to detect when w R·R + w F ·The larger the F, the more severe the construction site environment is determined to be and the lower the safety evaluation score S is.

[0067] The following gives an example formula that comprehensively considers the rainfall and the zero-air index obtained by processing the zero-air image of the construction environment collected by high-definition surveillance cameras to obtain the safety evaluation score.

[0068] Assume: R is the normalized rainfall (for example, dividing the actual rainfall value by a maximum value to make it fall within the range of [0, 1]); F is the fog index obtained by the image processing algorithm (also normalized to the range of [0, 1], where 0 means completely fog-free and 1 means the most severe fog); w R is the influence weight of rainfall on safety evaluation; w F is the influence weight of the zero-air index on safety evaluation, and it is required that w R + w F ≤ 1 to ensure that the score does not appear negative. The above safety evaluation score formula can be expressed as: S = 100 × [1 - (w R ·R + w F ·F)] where 100 represents the full score.

[0069] When both R and F are 0, that is, there is no rainfall and no fog, the safety evaluation score S = 100; when w R ·R + w F ·F is larger, it means the construction site environment is more severe and the safety evaluation score S is lower.

[0070] The above safety assessment module dynamically calculates the safety evaluation score based on the rainfall and fog index of the construction environment, providing a quantitative indicator of the safety status of the construction environment. It can promptly alert managers in harsh environments (such as heavy rain, thick fog, etc.) to improve the safety of the construction site. Furthermore, through the comprehensive evaluation results, the construction arrangement can be optimized to reduce safety hazards caused by environmental factors.

[0071] Preferably, as an implementable solution; the illegal behavior sampling module is specifically used to intercept video screenshots when illegal behaviors occur; record the time, location of the occurrence of illegal behaviors, and the basic attribute information of construction workers and construction equipment; generate an illegal behavior report based on the occurrence time, location, and the basic attribute information of construction workers and construction equipment, and push it to the construction site management terminal or the server side.

[0072] Preferably, as an implementable solution; the high-definition surveillance cameras and the fence warning surveillance cameras further include an image data storage module; the image data storage module supports segmented storage of long-time video data, supports the playback function, and provides a quick positioning function for illegal behaviors.

[0073] The image data storage module supports the storage, segmented storage, and playback functions of long-term video data. It can provide a function for quickly locating violations, facilitating post-event review and the preservation of evidence for claims, thereby forming a complete chain of construction image data and enhancing the transparency and traceability of the construction process.

[0074] Preferably, as an implementable solution, it further includes multiple management terminals that establish a cloud service communication connection with the server. The management terminals perform two-way signal call control with the server and are used to view in real time the images collected by high-definition monitoring cameras and fence warning monitoring cameras at the construction site, as well as the data, historical records, and violation reports collected by the first sensor device and the second sensor device.

[0075] Preferably, as an implementable solution, it further includes an environmental parameter evaluation module. The environmental parameter evaluation module is used to determine the environmental index E of the construction site according to the environmental parameters collected by the first sensor device: E = w1T + w2H + w3D + w4N + w5R, where: T is the environmental temperature, H is the environmental humidity, D is the dust concentration, N is the noise level, R is the rainfall; w1, w2, w3, w4, w5 are the weight coefficients of each parameter. Calculate the environmental index E through the formula and associate the environmental index E with the violation records. If the environmental index E is lower than the environmental health score threshold, trigger a global security alarm and send an alarm message to the server side.

[0076] The environmental parameter evaluation module can calculate the environmental index E based on the environmental parameters, associate the environmental indicators with the violation records, and form a data-based analysis of the environment and violation behaviors. When the environmental index E is lower than the threshold, it automatically triggers a global security alarm to promptly respond to potential safety hazards. In this way, through dynamic environmental parameter evaluation, the monitoring and management capabilities of the environmental health at the construction site are improved.

[0077] Embodiment 2

[0078] Based on the same concept as the above method embodiment, the embodiment of the present invention also provides a panoramic monitoring and violation behavior recognition method for a water conservancy construction site, which is used to implement the above method of the present invention. Since the principle and method of solving problems in this system embodiment are similar, it has at least all the beneficial effects brought by the technical solutions of the above embodiments, and will not be elaborated here one by one.

[0079] See Figure 3 , the present invention provides a panoramic monitoring and violation behavior recognition method for a water conservancy construction site, including the following steps:

[0080] S10. Arrange a fence warning monitoring camera, a high-definition monitoring camera, a first sensor device, and a second sensor device at the water conservancy construction site, and merge and load the positions of the fence warning monitoring camera, the high-definition monitoring camera, the first sensor device, and the second sensor device with the construction site topographic map and the construction site building map into a preset 3D BIM model for binding to achieve real-time monitoring of construction operations and environmental conditions.

[0081] S20. Install a high-definition monitoring camera for behavior recognition in the construction area. The high-definition monitoring camera is used to collect images of the behaviors of construction workers and construction equipment at the construction site to obtain construction worker behavior images and construction equipment behavior images. The high-definition monitoring camera is connected to a data processing box through a real-time video stream. The data processing box calls a preset typical illegal construction behavior database and a typical illegal equipment image database, and identifies and appraises the collected construction worker behavior images with the typical illegal construction behavior database, and compares the construction equipment behavior images with the typical illegal equipment image database. When it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, an alarm message is sent to the server side, and the basic attribute information of the current construction workers and construction equipment is sent to the server side. Among them, the typical illegal behaviors in the typical illegal construction behavior database include that construction workers do not wear safety helmets, do not wear reflective vests, do not fasten seat belts, smoke, and doze off. Among them, the typical illegal equipment behaviors in the typical illegal equipment image database include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs.

[0082] S30. Build a prohibitive electronic fence based on the second sensor device. The second sensor device is an infrared sensor, and when it detects that there are construction workers and equipment entering the current electronic fence, it sends an alarm message to the server side.

[0083] In the embodiment of the present invention, the above-mentioned electronic fence is a virtual space fence composed of multiple optoelectronic sensor groups, which can help detect when people or objects enter or leave the corresponding designated area, and also has an alarm function.

[0084] S40. Build a warning electronic fence based on the second sensor device and the fence warning monitoring camera. The fence warning monitoring camera collects images of the current warning electronic fence, and compares the collected images with a preset image database to identify and judge whether the construction workers and construction equipment entering the construction area are illegal personnel. If the judgment is yes, an alarm message is sent to the server side.

[0085] The S50, the illegal behavior sampling module intercepts the behavior images of construction workers and construction equipment with illegal behaviors collected, obtains the intercepted illegal pictures and the basic attribute information of the construction workers and construction equipment, and sends them to the server side.

[0086] It should be noted that the above high-definition monitoring cameras are used to collect images of the behaviors of construction workers and construction equipment at the construction site, obtain the behavior images of construction workers and construction equipment, and then perform image database analysis. The specific steps are as follows.

[0087] Use on-site monitoring cameras or other collection devices to obtain real-time images or video data of the construction site. Adjust the collected images, such as denoising, enhancement, cropping, size standardization, etc., to ensure that the image frequency quality meets the requirements during subsequent recognition and analysis.

[0088] Analyzing the above technical solution, it can be seen that the present invention provides a panoramic monitoring and illegal behavior recognition method for a water conservancy construction site, which collects a variety of monitoring data and monitoring data at multiple installation positions, obtains the spatial position information of the operators; calls the electronic fence and the monitoring data for further image analysis of abnormal behavior operations.

[0089] During preprocessing, a typical illegal behavior database is constructed to realize sample collection and database establishment, collect and sort out various illegal construction behavior images. Typical illegal behaviors may include: working at heights without wearing safety protection equipment (such as safety helmets, seat belts, etc.), lack of effective protection measures, improper working postures or dangerous actions, illegal entry into dangerous areas, etc.; then data annotation is carried out: detailed annotation is carried out on various collected illegal behavior images, including illegal behavior types, key areas (such as human body, tool area) and other auxiliary information. The accuracy of data annotation directly affects the learning effect of the model.

[0090] Meanwhile, a deep learning model for object detection of constructing images of construction workers' behaviors and construction equipment behaviors is built, such as Faster R-CNN and YOLO. These models perform excellently in real-time object detection tasks and can locate and classify violations simultaneously. The above-mentioned Faster R-CNN is used to extract features in the image and learn to distinguish between violation behaviors and normal behaviors. Finally, for real-time recognition and judgment, the image to be detected is input: the preprocessed image is input into the trained deep learning model. The model will automatically extract the features in the image and match them with the features in the typical violation behavior database, and output a similarity score. When the similarity score output by the model exceeds the standard score threshold, it can be determined that there is a violation behavior in the image. When the system determines a violation behavior, an alarm mechanism is triggered, and at the same time, detailed violation behavior information (including time, location, and specific description of the violation behavior) is recorded for subsequent statistical analysis and safety management. The above-mentioned server side can display the recognition results of violation behaviors, heat maps of risk areas, etc. on the real-time monitoring interface to assist the construction site management personnel to respond quickly.

[0091] In summary, the present invention provides a panoramic monitoring and violation behavior recognition system for a water conservancy construction site, which realizes the interaction between a high-definition monitoring camera and a data processing box; performs real-time image acquisition on the behaviors of construction workers and equipment at the construction site. By calling a preset violation behavior database, it automatically identifies construction behaviors that do not meet the specifications (such as not wearing a safety helmet, not wearing a reflective vest, etc.). It realizes the rapid detection, alarm, and information sending of violation behaviors, and improves the safety management efficiency of the construction site.

[0092] The present invention provides a panoramic monitoring and violation behavior recognition system for a water conservancy construction site, which realizes real-time remote monitoring of the construction site, and management personnel can view the images and sensor data collected by the camera at any time. At the same time, it supports the back-check of historical records and the viewing of violation behavior reports, improving the management transparency of the construction site. Through two-way signal calls, management personnel can quickly respond to on-site problems and improve management efficiency.

[0093] In summary, the present invention provides a panoramic monitoring and violation behavior recognition system for a water conservancy construction site, which closely focuses on the safety management, violation behavior recognition, environmental monitoring, and intelligent attendance management of the construction site, and forms an overall multi-dimensional intelligent construction site monitoring and management system, significantly improving the safety and management efficiency of the construction site.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; those of ordinary skill in the art can modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A panoramic monitoring and illegal behavior recognition system for a water conservancy construction site, characterized in that, Including: A high-definition surveillance camera, a fence warning surveillance camera, a first sensor device, a second sensor device, a server end, and a data processing box: The high-definition surveillance camera is used to collect images of the behaviors of construction workers and construction equipment at the construction site to obtain construction worker behavior images and construction equipment behavior images; the high-definition surveillance camera is also connected to the data processing box through a real-time video stream, and the data processing box calls a preset typical illegal construction behavior database and a typical illegal equipment image database; and it identifies and appraises the collected construction worker behavior images with the typical illegal construction behavior database, compares the construction equipment behavior images with the typical illegal equipment image database, and when it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, it sends an alarm message to the server end and sends the basic attribute information of the current construction workers and construction equipment to the server end; among them, the typical illegal behaviors in the typical illegal construction behavior database include construction workers not wearing safety helmets, not wearing reflective vests, not wearing seat belts, smoking, and dozing off; among them, the typical illegal equipment behaviors in the typical illegal equipment image database include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs; Based on the second sensor device, a prohibitive electronic fence is constructed. The second sensor device is an infrared sensor, which is used to send an alarm message to the server end when it detects that there are construction workers and equipment entering the current electronic fence; Based on the second sensor device and the fence warning surveillance camera, a warning electronic fence is constructed. The fence warning surveillance camera is used to collect images of the current warning electronic fence, and compare the collected images with a preset image database to identify and judge whether the construction workers and construction equipment entering the construction area are illegal personnel. If the judgment is yes, it sends an alarm message to the server end; It also includes an illegal behavior sampling module; the illegal behavior sampling module is used to intercept the collected construction worker behavior images and construction equipment behavior images with illegal behaviors, obtain the intercepted illegal pictures and the basic attribute information of the construction workers and construction equipment, and send them to the server end.

2. The panoramic monitoring and illegal behavior identification system for a water conservancy construction site according to claim 1, characterized in that, It also includes an attendance management module: the attendance management module is used to monitor and manage the attendance of construction workers entering and leaving the site in real time by combining face recognition processing; the attendance management module is also used to monitor and manage the construction equipment entering and leaving the site in real time by combining the equipment information scanning processing method.

3. The panoramic monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 2, wherein The attendance management module includes a first attendance management sub-module and a second attendance management sub-module: The first attendance management sub-module is used to collect the facial features of construction workers through a high-definition surveillance camera and compare them with a preset facial feature database of construction workers to confirm the identity information of the construction workers; the identity information includes name information and job number information; The second attendance management sub-module is used to collect the unique identification information of construction equipment through device information scanning technology, and confirm the entry and exit status of the equipment by comparing the unique identification information with the equipment management database. It is also used to obtain the entry and exit records of construction personnel and construction equipment, and then generate and store automated attendance data based on the entry and exit records of construction personnel and construction equipment, or send it to the server side.

4. A panoramic monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 3, characterized in that, The first sensor device is used to collect environmental parameters at the construction site, and the environmental parameters include environmental temperature, environmental humidity, dust concentration, noise level, and rainfall. It also includes a safety assessment module, which is used to comprehensively evaluate the safety evaluation score of the construction site according to the rainfall and the foggy pictures of the construction environment collected by the high-definition monitoring camera, and send the safety evaluation score to the server side.

5. The panoramic monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 4, characterized in that, The safety assessment module includes a calculation and processing module and an analysis and processing module. The calculation and processing module is specifically configured to obtain the normalized rainfall amount R and the fog index F; construct a safety evaluation score based on the rainfall amount R and the fog index F; the formula of the safety evaluation score is expressed as S = 100×[1-(w R ·R + w F ·F)]; Among them, 100 represents the full score; R is the normalized rainfall; F is the fog index obtained by the image processing algorithm, where 0 means completely fog-free and 1 means the thickest fog); w R is the influence weight of rainfall on safety evaluation; w F is the influence weight of the fog index on safety evaluation, and it is required that w R + w F ≤ 1; The analysis and processing module is used to determine that there is no rainfall and no fog when the rainfall R and fog index F are both 0, and output the safety evaluation score S=100; and is also used to determine that there is no rainfall and no fog when the rainfall R and fog index F are both 0. R ·R+w F The larger F is, the worse the construction site environment is and the lower the safety evaluation score S is.

6. The panoramic monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 5, characterized in that, The illegal behavior sampling module is specifically used to intercept video screenshots when illegal behaviors occur, record the time, location of the occurrence of illegal behaviors, and the basic attribute information of construction personnel and construction equipment, generate an illegal behavior report based on the time, location, and basic attribute information of construction personnel and construction equipment, and push it to the construction site management terminal or the server side.

7. A panoramic monitoring and illegal behavior identification system for a water conservancy construction site according to claim 5, characterized in that, The high-definition monitoring camera and the fence warning monitoring camera also include an image data storage module. The image data storage module supports segmented storage of long-term video data, supports the playback function, and provides a quick positioning function for illegal behaviors.

8. An overall monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 6, characterized in that, It also includes multiple management terminals that establish a cloud service communication connection with the server side. The management terminals perform two-way signal call control with the server side, and the management terminals are used to view in real time the images collected by the high-definition monitoring camera and the fence warning monitoring camera at the construction site, as well as the data, historical records, and illegal behavior reports collected by the first sensor device and the second sensor device.

9. The panoramic monitoring and illegal behavior recognition system for a water conservancy construction site according to claim 1, characterized in that, It also includes an environmental parameter evaluation module. The environmental parameter evaluation module is used to determine the environmental index E of the construction site according to the environmental parameters collected by the first sensor device: E = w1T + w2H + w3D + w4N + w5R, where: T is the environmental temperature, H is the environmental humidity, D is the dust concentration, N is the noise level, R is the rainfall; w1, w2, w3, w4, w5 are the weight coefficients of each parameter; calculate the environmental index E through the formula, and associate the environmental index E with the illegal records. If the environmental index E is lower than the environmental health score threshold, trigger a global safety alarm and send an alarm message to the server side.

10. A panoramic monitoring and illegal behavior recognition method for a water conservancy construction site, characterized in that, It includes the following steps: Arrange fence warning monitoring cameras, high-definition monitoring cameras, the first sensor device and the second sensor device at the water conservancy construction site, and merge and load the positions of the fence warning monitoring cameras, high-definition monitoring cameras, the first sensor device and the second sensor device with the construction site topographic map and the construction site building map into a preset three-dimensional BIM model for binding to realize real-time monitoring of construction operations and environmental conditions. Install high-definition surveillance cameras for behavior recognition in the construction area; the high-definition surveillance cameras are used to collect images of the behaviors of construction workers and construction equipment at the construction site to obtain construction worker behavior images and construction equipment behavior images; the high-definition surveillance cameras are connected to a data processing box through a real-time video stream, and the data processing box calls a preset database of typical illegal construction behavior and a database of typical illegal equipment images; and identify and authenticate the collected construction worker behavior images with the database of typical illegal construction behavior, compare the construction equipment behavior images with the database of typical illegal equipment images, and when it is identified and judged that there are illegal behaviors in the collected construction worker behavior images and construction equipment behavior images, send an alarm message to the server side and send the basic attribute information of the current construction workers and construction equipment to the server side; wherein, the typical illegal behaviors in the database of typical illegal construction behavior include construction workers not wearing safety helmets, not wearing reflective vests, not wearing seat belts, smoking, and dozing off; wherein, the typical illegal equipment behaviors in the database of typical illegal equipment images include: illegal driving of construction equipment, illegal operation of construction equipment, construction equipment crossing prohibited construction areas, and construction equipment not installing warning signs; Construct a prohibitive electronic fence based on the second sensor device, where the second sensor device is an infrared sensor, and send an alarm message to the server side when it detects that there are construction workers and equipment entering the current electronic fence; Construct a warning electronic fence based on the second sensor device and a fence warning surveillance camera. The fence warning surveillance camera collects images of the current warning electronic fence, and compares the collected images with a preset image database to identify and judge whether the construction workers and construction equipment entering the construction area are illegal personnel. If the judgment is yes, send an alarm message to the server side; The illegal behavior sampling module intercepts the collected construction worker behavior images and construction equipment behavior images with illegal behaviors, obtains the intercepted illegal pictures and the basic attribute information of the construction workers and construction equipment, and sends them to the server side.

Images