A fire monitoring method and system based on Internet of Things

By calculating the correlation between flame and background and combustion risk values, the misjudgment problem of relying solely on open fire in the prior art is solved, and the accuracy and robustness of fire risk monitoring is improved.

CN118470656BActive Publication Date: 2024-11-05SHAANXI ZHIZHUO IND CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202410926365.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-05
Estimated Expiration
2044-07-11

AI Technical Summary

Technical Problem

In existing fire monitoring technology, only open fires are used to judge fire risks, and it is impossible to effectively distinguish the combustion capacity and fire risks of different objects, resulting in inaccurate misjudgment and monitoring.

Method used

By collecting images, the correlation between flame and background and combustion risk values ​​are calculated, a two-way graph is constructed, the area of ​​the maximum connected sub-graph is calculated, and the fire risk value is calculated in combination with the combustion direction coefficient, thereby improving the accuracy of fire fire risk monitoring.

Benefits of technology

It improves the accuracy and robustness of fire risk monitoring, reduces misjudgment, and enhances the effectiveness of fire safety monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118470656B_ABST
    Figure CN118470656B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of fire monitoring, and more specifically, the present invention relates to a fire monitoring method and system based on the Internet of Things, including: collecting images of the area to be monitored, and obtaining the flame area in the image; segmenting the image without the flame area to obtain the background area of ​​each object; calculating the correlation between any flame and any background, constructing a bidirectional graph with each background as a node, and taking the node corresponding to the background connected to any flame as the starting point, calculating the maximum connected subgraph corresponding to the starting point, thereby obtaining the total area of ​​the background corresponding to each node in the subgraph. When calculating the fire risk level of the flame in the current image, the present invention takes into account the fire risk that may be caused by the flame and the surrounding background, and does not judge the fire risk level only by the size of the flame of the open flame, thereby improving the robustness during fire monitoring and improving the accuracy of fire risk monitoring.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of fire monitoring, and more specifically, to a fire monitoring method and system based on the Internet of Things. Background Art

[0002] In the related technology, for example, the Chinese patent document with publication number CN112036411A discloses a method for cyclic error correction of an intelligent fire monitoring and early warning system. By obtaining the edge irregularity characteristics of the smoke pattern of the smoke image, the frequency characteristics of pixel changes in the internal area of ​​the smoke, the diffusion characteristics of the smoke, and the gray consistency characteristics of the smoke image, feature fusion judgment is performed, which can comprehensively judge the smoke situation, improve the accuracy of the judgment, and reduce the error of the early warning information. This invention analyzes the smoke characteristics to warn of fires. The early warning method is single, and the accuracy of the early warning through smoke characteristics is lower than that of the open fire early warning.

[0003] There are also image-based open flame monitoring models, but in actual use, only open flames are monitored. For example, the fire is judged by the size of the flame. If the flame is large, the risk is high, and if the flame is small, the risk is low. However, since there can be many different problems with the combustion source and the combustion capacity of different objects is different, the corresponding fire risks are also different. At present, judging the risk only by open flames may lead to misjudgment of the fire, affecting the accuracy of fire risk monitoring. Summary of the invention

[0004] The present invention provides a fire monitoring method and system based on the Internet of Things, aiming to solve the problem in the related art that because different objects have different combustion capabilities and corresponding fire risks are also different, judging the risk only by open flames may lead to misjudgment of fire, thus affecting the accuracy of fire risk monitoring.

[0005] In a first aspect, the present invention provides a fire monitoring method based on the Internet of Things, comprising: collecting an image of an area to be monitored and obtaining a flame area in the image; segmenting the image without the flame area to obtain a background area of ​​each object; and calculating the correlation between any flame and any background, the calculation formula being: ; In the formula, The current image The flame and The relevance of the background, The current image The minimum Euclidean distance between the boundary and the ridge line corresponding to the background, is the ridge length, For the historical image The flame is located at The boundary between the background and the current image The flame is located at The distance between the borders on the background, The current image Background and The length of the boundary between flames, is the combustion direction coefficient; a bidirectional graph is constructed with each background as a node, and the node corresponding to the background connected to any flame is taken as the starting point, and the maximum connected subgraph corresponding to the starting point is calculated, so as to obtain the total area of ​​the background corresponding to each node in the subgraph, which is used as the combustion risk value of the background connected to the flame; the product of the correlation between any flame and each background and the combustion risk value is calculated, the maximum value is selected as the fire risk value of the flame, and fire safety is monitored according to the size of the fire risk value of the flame.

[0006] When calculating the fire risk level of the flame in the current image, the correlation between any flame and any background is calculated first, and the combustion risk value of the flame and the connected background is calculated. When monitoring fire safety, the fire risk that may be caused by the flame and the surrounding background is combined, and the accuracy of fire risk monitoring is improved.

[0007] In one embodiment, fire safety is monitored according to the size of the fire risk value of the flame, including: selecting the fire risk value of the largest flame as the fire risk level of the current image; and in response to the fire risk level of the current image being greater than a first threshold, issuing an alarm.

[0008] The fire risk value of the maximum flame is used as the fire risk level of the current image. Compared with the average value, the alarm level is higher, which makes the fire monitoring safer.

[0009] In one embodiment, fire safety is monitored according to the size of the fire risk value of the flame, including: calculating the average value of the fire risk value of each flame as the fire risk level of the current image; in response to the fire risk level of the current image being greater than a second threshold, an alarm is issued.

[0010] In one embodiment, an image of the area to be monitored is collected, and a flame area in the image is obtained: the image is input into a pre-trained flame recognition model to obtain a mask image; a connected domain analysis is performed on the mask image to obtain a valued area in the mask image, wherein the valued area is the area where the flame is located.

[0011] In one embodiment, the image with the flame portion removed is segmented to obtain the background region of each object, including: using a SAM segmentation network to segment the image with the flame portion removed to obtain the background region of each object.

[0012] By removing the flame part, other objects in the background can be seen more clearly, thereby improving the accuracy of background object recognition.

[0013] In one embodiment, a method for calculating a combustion direction coefficient includes: obtaining the first The flame is located The current background area of ​​the upper boundary of the background is equal to the background area of ​​the historical image. The flame is located In response to the difference between the current background area and the historical background area being greater than a third threshold, the combustion direction coefficient is 1; in response to the difference between the current background area and the historical background area being less than the third threshold, the combustion direction coefficient is 0.

[0014] According to the above method, it can be determined whether the burning direction of the flame spreads toward the background. If it spreads toward the background, it means that the background is more likely to be the combustion source, so the calculated correlation is more accurate.

[0015] In one embodiment, For the historical image The flame is located at The boundary between the background and the current image The flame is located at The distance value between the boundaries on the background includes: using the DTW algorithm to calculate the distance between the first Any point on the boundary of a flame is The shortest distance value to any point on the boundary of a flame.

[0016] In one embodiment, the method further includes: taking the intersection of the current image and the flame region in the previous frame image to obtain the overlapped position of the two images, and extracting the ridge line of the overlapped position to obtain the ridge line.

[0017] In one embodiment, a bidirectional graph is constructed with each background as a node, including: obtaining the ignition point of each node; judging whether to retain the bidirectional edge based on calculating the difference in the ignition points between two nodes; in response to the ignition point difference being greater than a fourth threshold, replacing the bidirectional edge with a unidirectional edge from a high ignition point to a low ignition point; in response to the ignition point difference being less than the fourth threshold, retaining it as a bidirectional edge.

[0018] The significance of replacing the bidirectional edges with burning point difference greater than the fourth threshold with unidirectional edges (from high temperature nodes to low temperature nodes) is to accurately reflect the heat transfer direction and gradient in physical phenomena. This means that the burning point may flow from the node with higher temperature to the node with lower temperature. This unidirectional edge is more consistent with the actual heat transfer direction.

[0019] In a second aspect of the present invention, there is also provided a fire monitoring system based on the Internet of Things, comprising a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the above-mentioned fire monitoring methods based on the Internet of Things.

[0020] Beneficial effect: When calculating the fire risk level of the flame in the current image, the fire risk that may be caused by the flame and the surrounding background is taken into account. The fire risk level is not judged solely by the size of the open flame, which improves the robustness of fire monitoring and improves the accuracy of fire risk monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0022] Figure 1 is a flowchart schematically illustrating calculation of a fire risk value according to an embodiment of the present invention;

[0023] Figure 2 is a schematic diagram schematically showing flame and background according to an embodiment of the present invention;

[0024] Figure 3 FIG. 1 is a schematic diagram schematically showing a system structure according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0026] like Figure 1 As shown, the specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings.

[0027] Step S101: collecting an image of the area to be monitored, and obtaining a flame area in the image.

[0028] In one embodiment, an RGB camera is installed above the area to be monitored, and an image of the area to be monitored is collected by the RGB camera. Specifically, when the image of the area to be monitored is collected by the RGB camera, the collection can be performed at a preset frequency, for example, the preset frequency is 1 Hz, 3 Hz or 5 Hz, etc., which can be adjusted according to actual conditions. Then, the image of the area to be monitored is sent to the Internet of Things center for analysis and processing to obtain the flame area in the image.

[0029] In one embodiment, the image of the area to be monitored is sent to the Internet of Things Center for analysis and processing to obtain the flame area in the image, including: inputting the image into a trained flame recognition model to obtain a mask image in the flame recognition result. For training the flame recognition model, it can be Mask-RCNN (a deep learning model for target detection and instance segmentation) or other semantic segmentation networks. Specifically, various types of pictures about flames and smoke are collected on the Internet, and the LabelMe annotation tool is used to annotate the pixels belonging to the flame in each picture, where the pixels belonging to the flame are labeled as 1, and the pixels belonging to the background are labeled as 0.

[0030] Exemplarily, an image is input into a trained flame recognition model, and a second image is output. The second image is a mask image, so the second image is a 01 binary image. Then, a connected domain analysis is performed on the second image to obtain a valued area in the second image, wherein the valued area is the area where a single flame is located. At this point, the area where each flame in the current image is located can be obtained.

[0031] Step S102: segmenting the image with the flame area removed to obtain the background area of ​​each object.

[0032] In one embodiment, after obtaining the area of ​​a single flame, in order to obtain the risks that may be caused by open flames, the SAM (Segment Anything Model) segmentation network or its network model is used for segmentation. It should be noted that during segmentation, the image belonging to the flame part needs to be removed from the acquired image through a mask operation, so as to obtain the background area of ​​each object corresponding to the SAM model. The reason for removing the flame part is that when there is a flame in the image, it may interfere with or cover up the recognition of other objects in the background. By removing the flame part, other objects in the background can be seen more clearly, thereby improving the accuracy of background object recognition. At this point, the boundary of each background area is extracted to obtain the background boundary corresponding to each object.

[0033] Step S103: Calculate the correlation between any flame and any background.

[0034] In one embodiment, for any flame and any background, if the flame is formed by the combustion of an object corresponding to the background, then the flame and the background must have a partial overlap in the boundary. It should be noted that if the flame burns unstably, the area that causes the flame to be unstable is not the location of the combustion source, because the combustion source is often at the bottom of the flame.

[0035] According to the above principle, the overlapping parts of the flames between the two frames are obtained. The more overlapping parts there are, the more likely the flame at that position is to be the combustion source. The closer the background is to the overlapping parts, the more likely it is to be the combustion source. If it is the combustion source, the boundary between the combustion source and the flame should change and expand toward the position of the combustion source. Therefore, the correlation between any flame and any background can be calculated.

[0036] In one embodiment, the correlation between any flame and any background is calculated using the following formula:

[0037] .

[0038] In the formula, The current image The flame and The relevance of the background, The current image The minimum Euclidean distance between the boundary and the ridge line corresponding to the background, is the ridge length, For the historical image The flame is located at The boundary between the background and the current image The flame is located at The distance between the borders on the background, The current image Background and The length of the boundary between flames, is the combustion direction coefficient. The flame is located at The boundary between the background and the current image The flame is located at The shortest distance between the boundaries on the background can be calculated by the DTW algorithm.

[0039] In one embodiment, the method for obtaining the ridge line includes: taking the intersection of the current image and the flame area in the previous frame image to obtain the overlapping position of the two images, and extracting the ridge line of the overlapping position to obtain the ridge line. It should be noted that the ridge line extraction method can adopt existing technologies, such as distance transformation method, etc.

[0040] In one embodiment, a method for calculating a combustion direction coefficient includes: obtaining the first The flame is located The current background area of ​​the upper boundary of the background is equal to the background area of ​​the historical image. The flame is located The historical background area on the upper boundary of the background, wherein the historical image can be the 10th frame image before the current image or the 15th frame image before the current image, which can be set manually according to the actual situation. If the difference between the current background area and the historical background area is greater than the third threshold, the burning direction coefficient is 1; if the difference between the current background area and the historical background area is less than the third threshold, the burning direction coefficient is 0, wherein the difference in area can be calculated to calculate the number of pixels in the background area, and the difference between the number of pixels in the current background and the number of pixels in the historical background is compared with the third threshold, so as to determine the burning direction coefficient, and the size of the third threshold can be set manually, and the value range of the third threshold is a constant greater than 0 and less than 100. In this embodiment, the third threshold is 50. For example, when the value after the historical background area minus the current background area is greater than 50, the burning direction coefficient is 1. In general, check whether the direction of the flame burning is spreading towards the background. If it spreads towards the background, it means that the background is more likely to be a burning source, and the burning direction coefficient is 1. If the flame does not spread towards the background, the background is less likely to be a burning source, and the burning direction coefficient is 0. At this point, the correlation between any flame and each background can be obtained.

[0041] Step S104: Calculate the combustion risk value of any flame and the background connected to the flame.

[0042] Specifically, since any flame may be connected to multiple backgrounds (e.g. Figure 2 As shown, Figure 2The H1 flame is connected to the Z1 background, Z4 background, Z7 background and Z8 background; the H2 flame is connected to the Z7 background and Z6 background), and the combustion risks between different backgrounds are different. After obtaining each background, it is necessary to calculate the combustion risk value of the flame and the background connected to the flame. Specific steps: construct graph structure information, construct a bidirectional graph with each background as a node, and then obtain the ignition point of each node; determine whether to retain the bidirectional edge based on the difference in ignition points between the two nodes, and determine whether to retain the bidirectional edge. If the difference in ignition points between any two nodes is greater than the fourth threshold, the bidirectional edge is replaced with a unidirectional edge from the high ignition point to the low ignition point; if the difference in ignition points between any two nodes is less than the fourth threshold, it is retained as a bidirectional edge, wherein the size of the fourth threshold can be adjusted by the implementer according to the specific implementation scenario, and the fourth threshold can be set to 5, 8 or 10, etc. Exemplarily, when the fourth threshold is 5, if the burning point of node J1 is R1 and the burning point of node J2 is R2, then the difference between the burning points of node J1 and node J2 is calculated to obtain R1-R2. If the value of R1-R2 is less than 5, it is retained as a bidirectional edge. Then, taking the node corresponding to the background connected to any flame as the starting point, the maximum connected subgraph corresponding to the starting point is calculated, so as to obtain the total area of ​​the background corresponding to each node in the subgraph as the combustion risk value of the background connected to the flame.

[0043] Step S105: Calculate the fire risk value of each flame to monitor fire safety.

[0044] In one embodiment, the product of the correlation between any flame and each background and the combustion risk value is calculated, the maximum value is selected as the fire risk value of the flame, and fire safety is monitored according to the size of the fire risk value of the flame.

[0045] In one embodiment, fire safety is monitored according to the fire risk values ​​of flames, including: after obtaining the fire risk values ​​of each flame, selecting the fire risk value of the flame with the largest fire risk value as the fire risk level of the current image.

[0046] For example, there are flames A, B, and D in the current image, and the backgrounds connected to flame A in the current image include the first object, the second object, the third object, and the fourth object. The correlations between flame A and all the backgrounds are calculated, respectively. , , , , the combustion risk values ​​of flame A and all backgrounds are calculated as , , , Calculate the fire risk value of flame A. The calculation method includes: calculating the product of the correlation between flame A and each background and the combustion risk value, and obtaining , , and If the maximum value in the product is , then The value is taken as the fire risk value of flame A. According to the above method, the fire risk values ​​of flames B and D are calculated, and the largest fire risk value is selected from the fire risk values ​​of flame A, flame B and flame D. For example, the fire risk value of flame B is the largest, and the fire risk value of flame B is taken as the fire risk degree of the current image.

[0047] In another embodiment, fire safety is monitored according to the size of the fire risk value of the flame, including: calculating the average value of the fire risk value of each flame as the fire risk level of the current image, and in response to the fire risk level of the current image being greater than a second threshold, an alarm is issued, wherein the second threshold is set to 80 in this scheme, and in other embodiments, it can be adjusted by the implementer according to the specific implementation scenario.

[0048] In one embodiment, if the fire risk level of the current image is greater than a first threshold, an alarm is triggered and relevant personnel are notified through an IoT device; if the fire risk level of the current image is less than the first threshold, it means that the current fire risk level will not cause a large fire, and no processing is performed. In this solution, the first threshold is set to 100, and in other embodiments, the implementer may adjust it according to the specific implementation scenario.

[0049] In another embodiment, if the fire risk level of the current image is less than the first threshold, an alarm may be issued, but the alarm level is lower than the alarm level when the fire risk level is greater than the first threshold, so as to remind relevant personnel.

[0050] For example, when the fire risk level is greater than the first threshold, the alarm level is level 3, and for level 3 alarms, relevant personnel and firefighters can be notified. If the fire risk level is less than the first threshold, the alarm level is reduced to level 1, and for level 1 alarms, only relevant personnel can be notified, and firefighters are not notified.

[0051] The present invention also provides a fire monitoring system based on the Internet of Things. Figure 3 As shown, the system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a fire monitoring method of the Internet of Things according to the first aspect of the present invention is implemented.

[0052] In one embodiment, the present invention provides a computer device whose internal structure can be as follows: Figure 3As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities, and various varieties such as CPU, single-chip microcomputer, DSP or FPGA can be selected. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. When the computer program is executed, the steps described in the above method embodiment can be completed, such as S101-S105. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a fire monitoring method based on the Internet of Things is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0053] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present invention, and does not constitute a limitation on the computer device of the present invention. The specific computer device may include more or less components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0054] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface, whose configuration and functions are known in the art and thus will not be described in detail here.

[0055] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0056] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0057] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0058] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent application. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A fire monitoring method based on the Internet of Things, characterized in that: include: Collecting an image of the area to be monitored and obtaining a flame area in the image; Segment the image after removing the flame area to obtain the background area of ​​each object; Calculate the correlation between any flame and any background, the calculation formula is: ; In the formula, The current image The flame and The relevance of the background, The current image The minimum Euclidean distance between the boundary and the ridge line corresponding to the background, is the ridge length, For the historical image The flame is located at The boundary between the background and the current image The flame is located at The distance between the borders on the background, The current image Background and The length of the boundary between the flames, is the combustion direction coefficient, wherein the ridge line acquisition method includes: taking the intersection of the flame area in the current image and the previous frame image to obtain the overlap position of the two images, and extracting the ridge line of the overlap position to obtain the ridge line; The calculation method of the combustion direction coefficient includes: Get the current image The flame is located The current background area of ​​the upper boundary of the background is equal to the background area of ​​the historical image. The flame is located The historical background area of ​​the border on the background; In response to the difference between the current background area and the historical background area being greater than a third threshold, the combustion direction coefficient is 1; in response to the difference between the current background area and the historical background area being less than the third threshold, the combustion direction coefficient is 0; A bidirectional graph is constructed with each background as a node, and the node corresponding to the background connected to any flame is taken as the starting point, and the maximum connected subgraph corresponding to the starting point is calculated, so as to obtain the total area of ​​the background corresponding to each node in the subgraph as the combustion risk value of the background connected to the flame; Calculate the product of the correlation between any flame and each background and the combustion risk value, select the maximum value as the fire risk value of the flame, and monitor fire safety according to the size of the fire risk value of the flame.

2. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: Fire safety is monitored based on the fire risk value of the flame, including: Select the fire risk value of the largest flame as the fire risk level of the current image; In response to the fire risk level of the current image being greater than a first threshold, an alarm is issued, wherein the first threshold is 100.

3. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: Fire safety is monitored based on the fire risk value of the flame, including: Calculate the average of the fire risk values ​​of each flame as the fire risk level of the current image; In response to the fire risk level of the current image being greater than a second threshold, an alarm is issued, wherein the second threshold is 80.

4. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: Collect an image of the area to be monitored and obtain the flame area in the image: Inputting the image into a pre-trained flame recognition model to obtain a mask image; Connected domain analysis is performed on the mask image to obtain a valued region in the mask image, wherein the valued region is the region where the flame is located.

5. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: The image without the flame is segmented to obtain the background area of ​​each object, including: The SAM segmentation network is used to segment the image with the flame removed to obtain the background area of ​​each object.

6. The fire monitoring method based on the Internet of Things according to any one of claims 1 to 5, characterized in that: The third threshold is 50.

7. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: For the historical image The flame is located at The boundary between the background and the current image The flame is located at The distance between the borders on the background includes: The DTW algorithm is used to calculate the Any point on the boundary of a flame is The shortest distance value to any point on the boundary of a flame.

8. The fire monitoring method based on the Internet of Things according to claim 1 is characterized in that: Construct a bidirectional graph with each background as a node, including: Get the burning point of each node; Determine whether to keep the bidirectional edge based on the difference in the burning points between the two nodes; In response to the difference being greater than a fourth threshold, replacing the bidirectional edge with a unidirectional edge from a high ignition point to a low ignition point; In response to the burning point difference being less than the fourth threshold, it is retained as a bidirectional edge.

9. A fire monitoring system based on the Internet of Things, characterized in that: It comprises a processor and a memory, the memory stores a computer program, and the processor executes the computer program to implement the fire monitoring method based on the Internet of Things as described in any one of claims 1 to 8.

Citation Information

Patent Citations

  • Cyclic error correction method for intelligent fire-fighting monitoring and early warning system

    CN112036411A

  • Daylight fire disaster video detection method

    CN105976398A

  • Flame detection and identification method based on infrared video image

    CN108108695A

  • Safety monitoring method and equipment, and computer readable storage medium

    CN110501914A

  • Comprehensive flame detection method based on ultraviolet and binocular vision

    CN111179279A