An engineering construction safety risk early warning method and system

Abstract

The present invention discloses a method and system for early warning of engineering construction safety risks. The method includes: dividing a construction site into several sub-regions, and obtaining first monitoring images of each sub-region at a first time interval; determining whether the several first monitoring images are abnormal images; if a certain first monitoring image is an abnormal image, defining the sub-region corresponding to the certain first monitoring image as an abnormal sub-region, and obtaining at least one second monitoring image of the abnormal sub-region at a second time interval; obtaining a second monitoring image whose similarity to a certain first monitoring image is less than a preset threshold, and determining whether the time length of the second monitoring image whose similarity to a certain first monitoring image is less than the preset threshold exists is greater than a preset time threshold. It can avoid the phenomenon that the risk warning of the construction site cannot be carried out in time due to the existence of an obstacle for too long.

Description

Technical Field

[0001] The present invention belongs to the technical field of safety risk prediction, and particularly relates to a method and system for early warning of engineering construction safety risks. Background Art

[0002] In construction work, the construction site environment is complex and there are many hazard sources. For example, hazards of dust and aerosol, hazards of flammable and explosive substances, high temperature hazards, noise hazards, and falling object hazards, etc. The number of construction site personnel and various large and small construction equipment is not only large, but it is also difficult to effectively master the construction rules, making management difficult. If there are illegal operations resulting in safety risks, due to the chaos of on-site personnel and equipment, it is difficult to detect safety risks in a timely manner. Even when safety problems have occurred, it is also difficult for the on-duty personnel to immediately know the location and type of the safety problems that have occurred, and they cannot arrange risk elimination matters in a timely and effective manner.

[0003] In view of the current insufficient safety of construction sites and the lack of effective risk early warning means, and how to ensure construction safety has increasingly attracted the attention of construction enterprises, it is urgent to study new technologies that can identify construction site safety risks in a timely and accurate manner. Summary of the Invention

[0004] The present invention provides a method and system for early warning of engineering construction safety risks, which are used to solve the technical problem that when there are obstacles at the construction site, it is impossible to accurately identify the risks of the construction site.

[0005] In a first aspect, the present invention provides a method for early warning of engineering construction safety risks, including: dividing the construction site into several sub-areas, and obtaining first monitoring images of each sub-area at a first time interval; determining whether the several first monitoring images are abnormal images; if a certain first monitoring image is an abnormal image, defining the sub-area corresponding to the certain first monitoring image as an abnormal sub-area, and obtaining at least one second monitoring image of the abnormal sub-area at a second time interval, where the second time interval is less than the first time interval; obtaining second monitoring images with a similarity less than a preset threshold to the certain first monitoring image, and determining whether the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image exists is greater than a preset time threshold; if the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image exists is greater than the preset time threshold, inputting the monitoring images of the normal sub-areas adjacent to the abnormal sub-area into a preset risk identification model, and sending out an early warning signal according to the output result of the risk identification model.

[0006] Further, after determining whether the several first monitoring images are abnormal images, the method further includes:

[0007] If none of the several first monitoring images is an abnormal image, the first monitoring images of each sub-region are input into a preset risk identification model to obtain the output of the risk identification model;

[0008] If the output obtained by inputting the first monitoring image of a certain sub-region into the risk identification model indicates a risk, a warning signal of the first priority is issued.

[0009] Further, after determining whether the duration during which a second monitoring image with a similarity less than a preset threshold to a certain first monitoring image exists is greater than a preset time threshold, the method further includes:

[0010] If the duration during which a second monitoring image with a similarity less than a preset threshold to a certain first monitoring image exists is not greater than the preset time threshold, the second monitoring image is input into the preset risk identification model to obtain the output of the risk identification model;

[0011] If the output obtained by inputting the second monitoring image into the risk identification model indicates a risk, a warning signal of the first priority is issued.

[0012] Further, the step of inputting the monitoring images of the normal sub-regions adjacent to the abnormal sub-region into the preset risk identification model and issuing a warning signal according to the output result of the risk identification model includes:

[0013] Input the monitoring images of the normal sub-regions adjacent to the abnormal sub-region into the preset risk identification model. If the output result of the risk identification model indicates that a certain normal sub-region is at risk, a warning signal of the first priority that both the certain normal sub-region and the abnormal sub-region are abnormal is issued;

[0014] Input the monitoring images of the normal sub-regions adjacent to the abnormal sub-region at the second monitoring moment into the preset risk identification model. If the output result of the risk identification model indicates that none of the adjacent normal sub-regions is at risk, a warning signal of the second priority that the abnormal sub-region is abnormal is issued.

[0015] Further, the risk identification model includes a high-altitude falling object identification model and a smoke identification model. Among them, the specific process of constructing the smoke identification model is as follows:

[0016] Receive multiple smoke color templates, where the smoke color templates record the color value regions of the smoke colors; obtain a monitoring image of a sub-region, extract the pixel region that matches the color value region of the smoke color, and denote it as the feature region; continuously obtain multiple monitoring images at the same position at a preset period, and extract the corresponding feature regions; if the pixel area of the feature region exceeds the threshold and the change range of the pixel positions of the feature region exceeds the preset threshold, it is determined that there is smoke in the sub-region;

[0017] The specific process of constructing the high-altitude falling object recognition model is as follows:

[0018] Receive the monitoring images of the obliquely downward-looking sub-region, receive the manually marked building boundary pixel region, denoted as the marked region; receive the calibration signal sent manually. When the calibration signal is received, obtain the latest monitoring image of the obliquely downward-looking sub-region, and fix the pixels of the latest monitoring image within the marked region as the reference marked region image; read and obtain the monitoring image of the obliquely downward-looking sub-region. If the reference marked region image is blocked, an alarm for the risk of high-altitude falling objects is issued. Otherwise, this step is re-executed in the next cycle.

[0019] In a second aspect, the present invention provides a construction safety risk early warning system, including: a division module configured to divide a construction site into several sub-regions and obtain first monitoring images of each sub-region according to a first time interval; a first judgment module configured to judge whether several first monitoring images are abnormal images; an obtaining module configured to, if a certain first monitoring image is an abnormal image, define the sub-region corresponding to the certain first monitoring image as an abnormal sub-region and obtain at least one second monitoring image of the abnormal sub-region based on a second time interval, where the second time interval is less than the first time interval; a second judgment module configured to obtain second monitoring images with a similarity less than a preset threshold to the certain first monitoring image and judge whether the time length during which the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image exist is greater than a preset time threshold; an identification module configured to, if the time length during which the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image exist is greater than the preset time threshold, input the monitoring images of the normal sub-regions adjacent to the abnormal sub-region into a preset risk identification model and issue a warning signal according to the output result of the risk identification model.

[0020] In a third aspect, there is provided an electronic device, which includes: at least one processor, and a memory communicatively connected to the at least one processor, where the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the steps of the construction safety risk early warning method according to any embodiment of the present invention.

[0021] In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the program instructions are executed by a processor, the processor is enabled to execute the steps of the construction safety risk early warning method according to any embodiment of the present invention.

[0022] An engineering construction safety risk early warning method and system of the present application can separately process abnormal images by determining whether the first monitoring image of each sub-region is an abnormal image, thereby improving the efficiency of troubleshooting. Moreover, by determining whether the duration of the second monitoring image whose similarity to a certain first monitoring image is less than a preset threshold in the abnormal sub-region corresponding to the abnormal image is greater than a preset time threshold, it can avoid the phenomenon that the construction site cannot be timely warned of risks due to the presence of an occlusion object for too long. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0024] Figure 1 FIG. is a flowchart of an engineering construction safety risk early warning method provided by an embodiment of the present invention;

[0025] Figure 2 FIG. is a structural block diagram of an engineering construction safety risk early warning system provided by an embodiment of the present invention;

[0026] Figure 3 FIG. is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0028] Please refer to Figure 1 , which shows a flowchart of an engineering construction safety risk early warning method of the present application.

[0029] As Figure 1 shown, the engineering construction safety risk early warning method specifically includes the following steps:

[0030] Step S101, divide the construction site into several sub-regions, and obtain the first monitoring image of each sub-region according to the first time interval.

[0031] In this embodiment, the construction site is the ground position where construction operations need to be carried out as involved in the project construction design plan. By dividing the construction site into several sub-regions and obtaining the first monitoring images of each sub-region according to the first time interval, it is possible to achieve regional management of the construction site.

[0032] For example, by dividing according to functional areas, multiple monitoring devices such as cameras and infrared sensors are set up at a construction site, and the monitoring images of each sub-region are obtained at an interval of 4 hours.

[0033] Step S102: Determine whether several first monitoring images are abnormal images.

[0034] In this embodiment, due to the complex environment of the construction site, it is often possible that construction equipment blocks the monitoring equipment, resulting in incomplete monitoring images. For the above possible actual problems, by determining whether several first monitoring images are abnormal images, it is possible to separately process abnormal images, thereby improving the efficiency of fault troubleshooting.

[0035] Specifically, if none of the several first monitoring images are abnormal images, the first monitoring images of each sub-region are input into a preset risk identification model to obtain the output of the risk identification model; if the output obtained by inputting the first monitoring image in a certain sub-region into the risk identification model indicates the existence of a risk, a warning signal of the first priority level is issued.

[0036] Among them, the risk identification model includes a high-altitude falling object identification model and a smoke identification model. The specific process of constructing the smoke identification model is as follows: Receive multiple smoke color templates, and the smoke color templates record the color value areas of smoke colors; Obtain a monitoring image of a sub-region, and extract the pixel area that matches the color value area of the smoke color, which is recorded as the feature area; Continuously obtain multiple monitoring images of the same position at a preset period, and extract the corresponding feature areas; If the pixel area of the feature area exceeds the threshold and the change range of the pixel position of the feature area exceeds the preset threshold, it is determined that there is smoke in the sub-region.

[0037] The specific process of constructing the high-altitude falling object identification model is as follows: Receive the monitoring image of the sub-region from an oblique top view, receive the manually marked building boundary pixel area, which is recorded as the marked area; Receive the calibration signal sent manually. When the calibration signal is received, obtain the latest monitoring image of the sub-region from an oblique top view, and fix the pixels of the latest monitoring image in the marked area as the reference marked area image; Read the monitoring image obtained from the oblique top view of the sub-region. If the reference marked area image is blocked, a high-altitude falling object risk alarm is issued. Otherwise, this step is re-executed in the next cycle.

[0038] It should be noted that the warning signals include multiple types, and different types of warning signals correspond to different priorities. For example, if there is no occlusion in the acquired first monitoring image, that is, none of them are abnormal images, then the first monitoring images of each sub-region are input into a preset risk identification model to obtain the output of the risk identification model; if the output obtained by inputting the first monitoring image in a certain sub-region into the risk identification model indicates a risk, a warning signal of sounding a horn or flashing a light is issued.

[0039] Step S103, if a certain first monitoring image is an abnormal image, then define the sub-region corresponding to the certain first monitoring image as an abnormal sub-region, and obtain at least one second monitoring image of the abnormal sub-region based on a second time interval, where the second time interval is less than the first time interval.

[0040] In this embodiment, if there is an occlusion in a certain first monitoring image, that is, it is an abnormal image, then define the sub-region corresponding to the certain first monitoring image as an abnormal sub-region. Then, at least one second monitoring image of the abnormal sub-region is obtained based on the second time interval for the abnormal sub-region. This can quickly check for potential safety risks that may exist in the abnormal sub-region.

[0041] In a specific application scenario, the construction site A is divided into 9 sub-regions according to functional areas. In the 5th sub-region, due to tower crane construction, at this time, the tower crane equipment occludes some areas in the 5th sub-region, resulting in blind spots in the monitoring image. During the tower crane construction process, adjust the time interval for obtaining the monitoring image. For example, when there is no tower crane, the time interval for obtaining the monitoring image is 4 hours, and when there is a tower crane, the time interval for obtaining the monitoring image is 2 hours. In this way, when the tower crane changes from being present to absent, a complete monitoring image without blind spots can be quickly obtained, effectively improving the accurate risk identification of the construction site.

[0042] Step S104, obtain a second monitoring image whose similarity to the certain first monitoring image is less than a preset threshold, and determine whether the time length during which the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than a preset time threshold.

[0043] In this embodiment, by obtaining a second monitoring image whose similarity to a certain first monitoring image is less than a preset threshold, a monitoring image with an occlusion in the sub-region can be obtained. By determining whether the time length during which the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than a preset time threshold, it is possible to avoid the phenomenon that the occlusion time is too long and the risk warning of the construction site cannot be carried out.

[0044] Specifically, if the duration for which a second surveillance image with a similarity less than a preset threshold to a certain first surveillance image exists is not greater than a preset time threshold, the second surveillance image is input into a preset risk recognition model to obtain the output of the risk recognition model; if the output obtained by inputting the second surveillance image into the risk recognition model indicates a risk, a warning signal of the first priority is issued.

[0045] Those skilled in the art should understand that the similarity between a certain first surveillance image and a second surveillance image can be calculated through different algorithms in combination with actual business requirements. For example, the Euclidean distance between the binary images corresponding to the similarity between a certain first surveillance image and a second surveillance image can be calculated to ultimately calculate the similarity between the corresponding first surveillance image and the second surveillance image.

[0046] In an implementable application scenario, the construction site A is divided into 9 sub - regions according to functional areas. In the 5th sub - region, since tower crane construction is required at 10 am, and the tower crane equipment blocks some areas in the 5th sub - region, resulting in a blind area in the surveillance image for a duration of 3 hours. During the tower crane construction, the time interval for obtaining the surveillance image is adjusted. When the tower crane is present, the time interval for obtaining the surveillance image is 1 hour. In this way, the similarity between the surveillance image with a blind area obtained at 10 am and the surveillance image without a blind area obtained at 6 am is less than the threshold, and since the duration of the blind area is 3 hours, which is less than the preset time threshold (4 hours). After continuously obtaining the surveillance images with blind areas at 11 am and 12 pm, a surveillance image without a blind area is obtained at 13 pm. At this time, the surveillance image obtained at 13 pm is input into the preset risk recognition model to obtain the output of the risk recognition model.

[0047] Step S105, if the duration for which a second surveillance image with a similarity less than a preset threshold to the certain first surveillance image exists is greater than a preset time threshold, the surveillance image of the normal sub - region adjacent to the abnormal sub - region is input into the preset risk recognition model, and a warning signal is issued according to the output result of the risk recognition model.

[0048] In this embodiment, if the duration for which a second surveillance image with a similarity less than a preset threshold to a certain first surveillance image exists is greater than a preset time threshold, the surveillance image of the normal sub - region adjacent to the abnormal sub - region is input into the preset risk recognition model. If the output result of the risk recognition model indicates that a certain normal sub - region is at risk, a warning signal of the first priority indicating that both the certain normal sub - region and the abnormal sub - region are abnormal is issued;

[0049] If the duration for which a second surveillance image with a similarity less than a preset threshold to a certain first surveillance image exists is greater than a preset time threshold, then the surveillance image of the normal sub-region adjacent to the abnormal sub-region at the second surveillance moment is input into a preset risk recognition model. If the output result of the risk recognition model is that there is no risk in the adjacent normal sub-regions, a warning signal of the second priority for the abnormal sub-region being abnormal is sent.

[0050] It should be noted that the warning signal of the second priority can be to send a prompt message in the communication device of the staff.

[0051] In summary, the method of the present application can separately process abnormal images by determining whether the first surveillance image of each sub-region is an abnormal image, thereby improving the efficiency of fault troubleshooting. Moreover, by determining whether the duration for which a second surveillance image with a similarity less than a preset threshold to the abnormal sub-region corresponding to the abnormal image exists is greater than a preset time threshold, it can avoid the phenomenon that the risk warning for the construction site cannot be given in a timely manner due to the long existence of the occluder.

[0052] Please refer to Figure 2 , which shows the structural block diagram of an engineering construction safety risk warning system of the present application.

[0053] As Figure 2 shown, the engineering construction safety risk warning system 200 includes a division module 210, a first judgment module 220, an acquisition module 230, a second judgment module 240, and an identification module 250.

[0054] Among them, the division module 210 is configured to divide the construction site into several sub-regions and obtain the first surveillance image of each sub-region according to the first time interval; the first judgment module 220 is configured to judge whether several first surveillance images are abnormal images; the acquisition module 230 is configured to, if a certain first surveillance image is an abnormal image, define the corresponding sub-region of the certain first surveillance image as an abnormal sub-region and obtain at least one second surveillance image of the abnormal sub-region based on the second time interval, where the second time interval is less than the first time interval; the second judgment module 240 is configured to obtain the second surveillance image with a similarity less than a preset threshold to the certain first surveillance image and judge whether the duration for which the second surveillance image with a similarity less than a preset threshold to the certain first surveillance image exists is greater than a preset time threshold; the identification module 250 is configured to, if the duration for which the second surveillance image with a similarity less than a preset threshold to the certain first surveillance image exists is greater than a preset time threshold, input the surveillance image of the normal sub-region adjacent to the abnormal sub-region into a preset risk recognition model and send a warning signal according to the output result of the risk recognition model.

[0055] It should be understood that Figure 2 the various modules described in Figure 1 correspond to the respective steps in the method described in the reference Figure 2 Therefore, the operations, features, and corresponding technical effects described above for the method also apply to

[0056] In some other embodiments, the embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored. When the program instructions are executed by a processor, the processor is caused to execute the engineering construction safety risk early warning method in any of the above method embodiments;

[0057] As an implementation manner, the computer-readable storage medium of the present invention stores computer-executable instructions, and the computer-executable instructions are set as:

[0058] Divide the construction site into several sub-regions, and obtain the first monitoring image of each sub-region according to the first time interval;

[0059] Determine whether several first monitoring images are abnormal images;

[0060] If a certain first monitoring image is an abnormal image, then define the sub-region corresponding to the certain first monitoring image as an abnormal sub-region, and obtain at least one second monitoring image of the abnormal sub-region based on the second time interval, where the second time interval is less than the first time interval;

[0061] Obtain the second monitoring images whose similarity to the certain first monitoring image is less than the preset threshold, and determine whether the time length of the second monitoring images whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than the preset time threshold;

[0062] If the time length of the second monitoring images whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than the preset time threshold, then input the monitoring images of the normal sub-regions adjacent to the abnormal sub-region into a preset risk identification model, and issue a warning signal according to the output result of the risk identification model.

[0063] A computer-readable storage medium may include a storage program area and a storage data area. Among them, the storage program area can store an operating system and application programs required for at least one function; the storage data area can store data created according to the use of the engineering construction safety risk warning system, etc. In addition, the computer-readable storage medium may include a high-speed random access memory, and may also include a memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the computer-readable storage medium may optionally include a memory remotely provided with respect to the processor, and these remote memories can be connected to the engineering construction safety risk warning system through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0064] Figure 3 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention, as Figure 3 shown, the device includes: a processor 310 and a memory 320. The electronic device may further include: an input device 330 and an output device 340. The processor 310, the memory 320, the input device 330, and the output device 340 can be connected through a bus or other means, Figure 3 taking the connection through the bus as an example. The memory 320 is the above-mentioned computer-readable storage medium. The processor 310 executes various functional applications and data processing of the server by running non-volatile software programs, instructions, and modules stored in the memory 320, that is, implements the engineering construction safety risk warning method in the above method embodiment. The input device 330 can receive input digital or character information, and generate key signal inputs related to the user settings and function controls of the engineering construction safety risk warning system. The output device 340 may include a display device such as a display screen.

[0065] The above electronic device can execute the method provided by the embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method. For technical details not described in detail in this embodiment, reference can be made to the method provided by the embodiment of the present invention.

[0066] As an implementation manner, the above electronic device is applied to an engineering construction safety risk warning system and is used for a client, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to:

[0067] Divide the construction site into several sub-areas, and obtain the first monitoring image of each sub-area according to the first time interval;

[0068] Determine whether several first monitoring images are abnormal images;

[0069] If a certain first monitoring image is an abnormal image, a certain sub-region corresponding to the certain first monitoring image is defined as an abnormal sub-region, and at least one second monitoring image of the abnormal sub-region is obtained based on a second time interval, where the second time interval is less than the first time interval;

[0070] Obtain a second monitoring image whose similarity to the certain first monitoring image is less than a preset threshold, and determine whether the time length during which the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than a preset time threshold;

[0071] If the time length during which the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold exists is greater than the preset time threshold, input the monitoring image of the normal sub-region adjacent to the abnormal sub-region into a preset risk recognition model, and send out a warning signal according to the output result of the risk recognition model.

[0072] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution, in essence, or the part that contributes to the prior art can be embodied in the form of a software product, and this computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiments.

[0073] 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; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. An engineering construction safety risk warning method, characterized in that, it includes: Dividing the construction site into several sub-areas, and obtaining the first monitoring images of each sub-area according to the first time interval; Judging whether several first monitoring images are abnormal images; If a certain first monitoring image is an abnormal image, then defining the sub-area corresponding to the certain first monitoring image as an abnormal sub-area, and obtaining at least one second monitoring image of the abnormal sub-area based on the second time interval, where the second time interval is less than the first time interval; Obtaining second monitoring images with a similarity less than a preset threshold to the certain first monitoring image, and judging whether the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image is greater than a preset time threshold; If the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image is greater than the preset time threshold, then inputting the monitoring images of the normal sub-areas adjacent to the abnormal sub-area into a preset risk identification model, and sending a warning signal according to the output result of the risk identification model; After judging whether several first monitoring images are abnormal images, the method further includes: If several first monitoring images are not abnormal images, then inputting the first monitoring images of each sub-area into a preset risk identification model to obtain the output of the risk identification model; If there is a risk in the output obtained by inputting the first monitoring image of a certain sub-area into the risk identification model, then send a warning signal of the first priority; After judging whether the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image is greater than the preset time threshold, the method further includes: If the time length of the second monitoring images with a similarity less than the preset threshold to the certain first monitoring image is not greater than the preset time threshold, then inputting the second monitoring images into a preset risk identification model to obtain the output of the risk identification model; If the output obtained by inputting the second monitoring image into the risk identification model is a risk, then send a warning signal of the first priority.

2. The engineering construction safety risk warning method according to claim 1, characterized in that, The step of inputting the monitoring images of the normal sub-areas adjacent to the abnormal sub-area into a preset risk identification model and sending a warning signal according to the output result of the risk identification model includes: Inputting the monitoring images of the normal sub-areas adjacent to the abnormal sub-area into a preset risk identification model, and if the output result of the risk identification model is that there is a risk in a certain normal sub-area, then send a warning signal of the first priority that both the certain normal sub-area and the abnormal sub-area are abnormal; Inputting the monitoring images of the normal sub-areas adjacent to the abnormal sub-area at the second monitoring time into a preset risk identification model, and if the output result of the risk identification model is that there is no risk in the adjacent normal sub-areas, then send a warning signal of the second priority that the abnormal sub-area is abnormal.

3. A method for early warning of engineering construction safety risks according to any one of claims 1 to 2, characterized in that, the risk identification model includes a falling object identification model and a smoke identification model. Among them, the specific process of constructing the smoke identification model is as follows: Receive multiple smoke color templates, and the smoke color templates record the color value areas of smoke colors; obtain a monitoring image of a sub-region, extract the pixel region that matches the color value area of the smoke color, and record it as the feature region; continuously obtain multiple monitoring images at the same position at a preset period, and extract the corresponding feature regions; if the pixel area of the feature region exceeds the threshold and the change range of the pixel position of the feature region exceeds the preset threshold, it is determined that there is smoke in the sub-region; The specific process of constructing the falling object identification model is as follows: Receive the monitoring image of the sub-region from an oblique top view, receive the pixel region of the building boundary manually marked, and record it as the marked area; receive the calibration signal sent manually. When the calibration signal is received, obtain the latest monitoring image of the sub-region from an oblique top view, and fix the pixels of the latest monitoring image in the marked area as the reference marked area image; read and obtain the monitoring image of the sub-region from an oblique top view. If the reference marked area image is blocked, an alarm for the risk of falling objects is issued. Otherwise, this step is re-executed in the next cycle.

4. An engineering construction safety risk early warning system that adopts the method according to any one of claims 1 to 3, characterized in that, it includes: A division module configured to divide the construction site into several sub-regions and obtain the first monitoring image of each sub-region according to the first time interval; A first judgment module configured to judge whether several first monitoring images are abnormal images; An acquisition module configured to, if a certain first monitoring image is an abnormal image, define the corresponding sub-region of the certain first monitoring image as an abnormal sub-region, and obtain at least one second monitoring image of the abnormal sub-region based on the second time interval, where the second time interval is less than the first time interval; A second judgment module configured to obtain the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold, and judge whether the existence time length of the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold is greater than the preset time threshold; An identification module configured to, if the existence time length of the second monitoring image whose similarity to the certain first monitoring image is less than the preset threshold is greater than the preset time threshold, input the monitoring image of the normal sub-region adjacent to the abnormal sub-region into the preset risk identification model, and issue a warning signal according to the output result of the risk identification model.

5. An electronic device, characterized in that, it includes: At least one processor, and a memory communicatively connected to the at least one processor, where the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 3.

6. A computer-readable storage medium, on which a computer program is stored, characterized in that, When the described program is executed by a processor, it implements the method described in any one of claims 1 to 3.

Images