A fire alarm information generation method based on multi-element data fusion, medium and equipment

Abstract

The present application relates to the field of information generation, in particular to a fire alarm information generation method, medium and equipment based on multi-element data fusion, comprising: generating a scene type identifier and a scene confidence P corresponding to a sub-warning area; obtaining a plurality of second real scene image information according to the scene type identifier and P; generating a first information sequence of the sub-warning area according to the plurality of second real scene image information; generating first abnormal information of the sub-warning area according to the first information sequence; and generating fire alarm information of the sub-warning area according to the first abnormal information. In the present application, information of different sensors is considered simultaneously to determine whether a fire occurs, and whether a fire occurs can be determined more accurately through the fusion of multi-element information. In addition, the first information sequence reflecting the area change trend of the fire starting area is generated, and the accuracy of the fire alarm information of the sub-warning area is further improved by the information, thereby preventing false alarms.

Description

Technical Field

[0001] This invention relates to the field of information generation, and in particular to a method, medium, and device for generating fire alarm information based on multi-source data fusion. Background Technology

[0002] With the development of technology, fire prevention is becoming increasingly intelligent. Typically, fire detection sensors are installed in areas that require special protection to trigger fire alarms in those areas.

[0003] Existing fire detection sensors typically determine the presence of a fire based on a single signal, thereby triggering the corresponding alarm. For example, smoke sensors generate an alarm by detecting whether smoke concentrations in a target area exceed a certain level. Temperature sensors generate an alarm by detecting whether temperatures in a target area exceed a certain level. These methods infer fire occurrences based on a single change in information, and because the reference dimension is relatively limited, false alarms are prone to occur. Summary of the Invention

[0004] To address the aforementioned technical problems, the technical solution adopted by this invention is as follows:

[0005] According to one aspect of the present invention, a method for generating fire alarm information based on multi-source data fusion is provided, the method comprising the following steps:

[0006] In response to receiving primary anomaly information from any sub-warning area, the system acquires the first real-scene image information of the sub-warning area; the primary anomaly information is either smoke concentration anomaly information or target temperature anomaly information.

[0007] Input the first real-scene image information corresponding to the sub-early warning area into the first target recognition model to generate the scene type identifier and scene confidence P corresponding to the sub-early warning area;

[0008] The continuous monitoring period for the sub-early warning area is determined based on the scene type identifier;

[0009] Based on P, the monitoring image acquisition frequency Q of the sub-early warning area is determined; Q satisfies the following condition: Where K1 and K2 are the first and second adjustment coefficients, respectively; S is the upper limit of the number of data collections.

[0010] During the continuous monitoring period, multiple second-real-scene image information of the sub-early warning area are collected according to Q;

[0011] Multiple second real-scene image information are input into the second target recognition model to generate the first information sequence (A1, A2, ..., A1) of the sub-early warning area. iA f(Q) ); where A i Let f(Q) be the ratio between the area of ​​the fire zone in the i-th second real-scene image and the total area of ​​the entire image; f(Q) is the total number of second real-scene images collected according to Q, i = 1, 2, ..., f(Q);

[0012] If the similarity between the first information sequence of the sub-earning area and any first type of preset abnormal sequence is greater than the first similarity threshold, then the abnormal information corresponding to the first type of preset abnormal sequence will be used as the first abnormal information of the sub-earning area.

[0013] Based on the first abnormal information, fire alarm information for the sub-early warning area is generated.

[0014] Furthermore, before acquiring the first real-world image information of any sub-warning area in response to receiving primary anomaly information of any sub-warning area, the method further includes:

[0015] According to the first preset acquisition frequency, the thermal imaging information of each sub-early warning area in the target area is acquired cyclically.

[0016] If the target temperature in the thermal imaging information of any sub-warning area is greater than the temperature threshold, then the target temperature anomaly information corresponding to the sub-warning area is generated.

[0017] Furthermore, each sub-warning area is equipped with a corresponding smoke sensor to collect smoke concentration information in the sub-warning area;

[0018] Before acquiring the first real-world image information of any sub-warning area in response to receiving primary anomaly information of any sub-warning area, the method further includes:

[0019] If the smoke concentration information in any sub-warning area is greater than the smoke concentration threshold, then abnormal smoke concentration information corresponding to the sub-warning area is generated.

[0020] Furthermore, based on the scenario type identifier, the continuous monitoring period for the sub-early warning area is determined to include:

[0021] Based on the scenario type identifier, obtain the continuous monitoring period of the sub-early warning area from the first mapping table;

[0022] The first mapping table includes the mapping relationship between different scene type identifiers and their corresponding preset continuous monitoring periods.

[0023] Furthermore, multiple second real-scene image information are input into the second target recognition model to generate a first information sequence for the sub-warning area, including:

[0024] Multiple second real-scene image information are input into the second target recognition model to generate a target recognition box corresponding to the fire area in each second real-scene image information;

[0025] The ratio between the total area of ​​all target recognition boxes in each second real-scene image information and the total area of ​​the second real-scene image information is obtained to generate the first information sequence of the sub-warning area.

[0026] Furthermore, after determining the continuous monitoring period for the sub-early warning area based on the scene type identifier, the method also includes:

[0027] According to the second preset acquisition frequency W, the smoke concentration information of the sub-early warning area during the continuous monitoring period is obtained, and the second information sequence (B1, B2, ..., B) of the sub-early warning area is generated. n A f(W) ); where B n f(W) represents the nth smoke concentration information collected in the sub-warning area; f(W) represents the total number of smoke concentration information collected in the sub-warning area according to W, where n = 1, 2, ..., f(W);

[0028] If the similarity between the second information sequence of the sub-earning area and any second type of preset abnormal sequence is greater than the second similarity threshold, then the abnormal information corresponding to the second type of preset abnormal sequence will be used as the second abnormal information of the sub-earning area.

[0029] Based on the first and second abnormal information, fire alarm information for the sub-early warning area is generated.

[0030] Furthermore, after determining the monitoring image acquisition frequency Q of the sub-early warning area based on P, the method also includes:

[0031] During the continuous monitoring period, multiple thermal imaging images of the sub-early warning area are collected according to Q.

[0032] Multiple thermal imaging images are input into the second target recognition model to generate a third information sequence (C1, C2, ..., C...) for the sub-warning areas. i ..., C f(Q) ); where C i If the similarity between the three information sequences of the area of ​​the fire zone and the total area of ​​the entire image in the i-th thermal imaging image information and any third-class preset anomaly sequence is greater than the first similarity threshold, then the anomaly information corresponding to the third-class preset anomaly sequence is taken as the third anomaly information of the sub-warning area.

[0033] If at least two of the first, second, and third abnormal information represent the same fire information, then fire alarm information for the sub-warning area is generated based on the same fire information.

[0034] Furthermore, the first abnormal information includes the first fire sign and the fire type sign; the second abnormal information includes the second fire sign and the smoke type sign.

[0035] Based on the first and second anomaly information, the fire alarm information for the sub-early warning area is generated as follows:

[0036] If both the first fire indicator and the second fire indicator are the first preset indicators, then fire information is generated based on the first abnormal information and the second abnormal information; the fire information includes the location information of the sub-warning area, the fire type indicator, and the smoke type indicator;

[0037] Based on the fire type identifier and smoke type identifier, information on the type of extinguishing material to be deployed by the corresponding sprinkler extinguishing device in the sub-early warning area is generated.

[0038] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the above-described method for generating fire alarm information based on multi-source data fusion.

[0039] According to a third aspect of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for generating fire alarm information based on multi-source data fusion.

[0040] The present invention has at least the following beneficial effects:

[0041] In this invention, upon receiving initial anomaly information from any sub-warning area, image information within that sub-warning area is acquired, and the current scene type within the sub-warning area is determined based on the image information. The scene type can be specifically set according to the actual usage scenario, such as including no-fire scenarios, fire scenarios, and suspected fire scenarios. Furthermore, the length of the subsequent continuous monitoring period can be specifically set according to the different scenarios, thereby determining whether subsequent continuous monitoring is necessary based on the specific type of scenario.

[0042] Furthermore, since the first target recognition model infers and predicts corresponding scene information based on image information, it is also susceptible to scene misjudgment. Generally, the lower the scene confidence P obtained by the first target recognition model, the higher the likelihood of misjudgment. In this invention, through... The sampling frequency is determined during the subsequent continuous detection period for each scene. As shown in the algorithm above, within the upper limit of the sampling count, a lower scene confidence P corresponds to a higher sampling frequency, meaning a higher sampling frequency for image information in the sub-warning area. This allows for the acquisition of a denser first information sequence, forming more detailed and dense information features. Therefore, when calculating the similarity with the first type of preset abnormal sequence, the increased number of features improves the accuracy of the final generated fire alarm information, enabling a more accurate reflection of whether a fire has occurred in the area.

[0043] In this invention, after receiving initial anomaly information from any sub-warning area, image information of the sub-warning area during the corresponding continuous detection period is acquired, and a corresponding first information sequence is generated. This first information sequence is then used to determine whether a fire has occurred, and a corresponding fire alarm is generated. This approach simultaneously considers information from different sensors to jointly determine whether a fire has occurred, allowing for more accurate fire detection through the fusion of multi-source information and preventing false alarms. Furthermore, by using image information from the sub-warning area during the corresponding continuous detection period, a first information sequence reflecting the changing trend of the fire area is generated, further improving the accuracy of the generated fire alarm information and preventing false alarms. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A flowchart illustrating a method for generating fire alarm information based on multi-source data fusion, provided in an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] As one possible embodiment of the present invention, such as Figure 1 As shown, a method for generating fire alarm information based on multi-source data fusion is provided, which includes the following steps:

[0048] S100: In response to receiving primary anomaly information from any sub-warning area, acquire the first real-view image information of the sub-warning area. The primary anomaly information is either smoke concentration anomaly information or target temperature anomaly information.

[0049] In S100: before obtaining the first real-scene image information of the sub-warning area in response to receiving the primary anomaly information of any sub-warning area, the method further includes a method for generating the primary anomaly information.

[0050] On the one hand, when the primary anomaly information is the target temperature anomaly information, the primary anomaly information is generated according to the following method:

[0051] S110: According to the first preset acquisition frequency, the thermal imaging information of each sub-warning area in the target area is acquired cyclically.

[0052] The target area may be a large space, such as a stadium, warehouse, farm, or enclosed space like a corridor. Because the target area is large, monitoring equipment cannot completely cover it. Therefore, it is necessary to divide the target area into multiple sub-warning zones and use an infrared thermal imaging PTZ camera to scan each sub-warning zone in a loop to obtain thermal imaging information, such as a thermal image, for each sub-warning zone.

[0053] S120: If the target temperature in the thermal imaging information of any sub-warning area is greater than the temperature threshold, then generate the target temperature anomaly information corresponding to the sub-warning area.

[0054] If the temperature of any region or target in the thermal imaging image is greater than the temperature threshold at the time of ignition, target temperature anomaly information is generated.

[0055] On the other hand, when the primary anomaly information is smoke concentration anomaly information, the primary anomaly information is generated according to the following method:

[0056] S130: If the smoke concentration in any sub-warning area exceeds the smoke concentration threshold, an abnormal smoke concentration information for that sub-warning area is generated. Each sub-warning area is equipped with a corresponding smoke sensor to collect smoke concentration information.

[0057] Similarly, the generation method of abnormal smoke concentration information is basically the same as that of abnormal target temperature information, mainly by judging through the corresponding threshold.

[0058] S200: Input the first real-scene image information corresponding to the sub-warning area into the first target recognition model to generate the scene type identifier and scene confidence level P corresponding to the sub-warning area.

[0059] The first target recognition model can be an existing neural network model. It is trained by using real-world images of various preset scenarios within the sub-warning area as training samples, enabling it to recognize corresponding scenarios. Specific scenario types can be set according to actual usage scenarios, such as scenarios without fire, fire scenarios, and suspected fire scenarios. Suspected fire scenarios can include various situations that could potentially cause a fire, such as someone smoking in the corresponding area, or electric vehicles charging in the corridor. It can also include scenarios prone to misjudgment, such as dust and sand rising rapidly due to cleaning, or the presence of non-flammable items that continuously emit a large amount of heat (such as self-heating pots or high-calorie takeout food).

[0060] At the same time, since real-scene images can more accurately represent various behavioral information of the current scene, using real-scene images can more accurately reflect the type of the current scene.

[0061] S300: Determine the continuous monitoring period for the sub-early warning area based on the scene type identifier.

[0062] Specifically, the S300 includes:

[0063] S301: Based on the scene type identifier, obtain the continuous monitoring period of the sub-early warning area from the first mapping table. The first mapping table includes the mapping relationship between different scene type identifiers and their corresponding preset continuous monitoring periods.

[0064] Specifically, the continuous monitoring period for a fire-free scenario can be 0 minutes; the continuous monitoring period for a fire scenario can be 5-10 minutes. This continuous monitoring period is used to collect data to determine the changes in the fire situation, so as to more rationally dispatch firefighting forces. Of course, if a fire scenario is determined and the corresponding scenario confidence level P is greater than 0.8, a fire alarm will be issued immediately.

[0065] The continuous monitoring period for suspected fire scenarios needs to be set based on the specific scenario. Generally, this continuous monitoring period corresponds to the time interval between the occurrence of a fire caused by a potentially flammable object in the scenario, or the typical time it takes for the flammable object to extinguish. For example, when an electric bicycle is charging, it typically catches fire within 3-5 minutes once the charger or temperature reaches the ignition threshold; therefore, the corresponding continuous monitoring period for this scenario could be 3 minutes. Smoking typically dissipates within 5-7 minutes, so the corresponding continuous monitoring period for this scenario could be 7 minutes.

[0066] After the above segmentation of the scenarios, different continuous monitoring periods will be set because the circumstances under which fires may occur in different scenarios. This will facilitate further monitoring and judgment, and determine the specific fire situation based on subsequent monitoring data.

[0067] S400: Based on P, determine the monitoring image acquisition frequency Q of the sub-early warning area. Q satisfies the following condition: Wherein, K1 and K2 are the first and second adjustment coefficients, respectively, which can be set as needed, mainly for adjustment. The value range of S. S is the upper limit of the number of data collections, which can be 30. Q mentioned in this step is the number of data collections per minute.

[0068] Since the first-object recognition model is a predictive model that infers and predicts corresponding scene information from image information, it is susceptible to scene misjudgments. Generally, the lower the scene confidence score P obtained by the first-object recognition model, the higher the likelihood of misjudgments. Because the first-object recognition model is trained using pre-prepared image information from various scenes, its judgment of these existing scenes is more accurate, but its judgment of unfamiliar scenes will be lower. Furthermore, human behavior in real-world scenarios is more diverse, inevitably leading to some scenes that are relatively unfamiliar to the first-object recognition model. Therefore, in these unfamiliar scenes, feature information needs to be collected more frequently.

[0069] In this invention, by The sampling frequency during the subsequent continuous detection period for each scene is determined. As shown in the algorithm above, within the upper limit of the sampling count, a lower scene confidence P corresponds to a higher sampling frequency, meaning a higher sampling frequency for image information in the sub-warning area. This allows for the acquisition of a denser first information sequence, forming more detailed and dense information features. Consequently, the changing trend of the fire area in the scene can be determined earlier, facilitating more timely fire alarm information generation.

[0070] S500: During the continuous monitoring period, collect multiple second real-scene image information of the sub-early warning area according to Q.

[0071] S600: Input multiple second real-scene image information into the second target recognition model respectively to generate the first information sequence of the sub-warning area (A1, A2, ..., A...). i A f(Q) ), where A iLet f(Q) be the ratio between the area of ​​the fire zone in the i-th second real-scene image and the total area of ​​the entire image. Let f(Q) be the total number of second real-scene images collected according to Q, where i = 1, 2, ..., f(Q).

[0072] Specifically, the S600 includes:

[0073] S601: Input multiple second real-scene image information into the second target recognition model respectively to generate a target recognition box corresponding to the fire area in each second real-scene image information.

[0074] S602: Obtain the ratio between the total area value of all target recognition boxes in each second real-scene image information and the total area value of the second real-scene image information, and generate the first information sequence of the sub-warning area.

[0075] The second target recognition model can also be an existing neural network model. This model can identify the name of the burning target in the image information and configure a rectangular target recognition box of corresponding size for the image area of ​​the fire, such as marking a burning cigarette butt in the image. This yields the first information sequence for generating sub-warning areas. Furthermore, each feature value in the first information sequence can represent the changing trend of the fire area.

[0076] S700: If the similarity between the first information sequence of the sub-early warning area and any first type of preset abnormal sequence is greater than the first similarity threshold, then the abnormal information corresponding to the first type of preset abnormal sequence shall be taken as the first abnormal information of the sub-early warning area.

[0077] Specifically, existing vector similarity calculation methods can be used to determine the similarity between any two information sequences.

[0078] S800: Based on the first abnormal information, generate fire alarm information for the sub-early warning area.

[0079] The first type of preset abnormal sequence is a sequence of information on the change in the area of ​​the fire zone after ignition. Typically, as the fire continues to develop, the corresponding fire zone area gradually increases, exceeding a certain threshold, and the rate of increase accelerates later. Therefore, multiple preset abnormal sequences can be formed based on historical fire situations in different fire scenarios. Furthermore, since this invention also monitors suspected fire scenarios, multiple preset sequences of information on the change in the fire area in suspected fire scenarios are also created. For example, in a smoking scenario, the fire zone area may remain constant for a period of time, then gradually decrease until it disappears. Or, in some spontaneous combustion scenarios, the fire zone area may continuously increase for a period of time, then remain constant, and then gradually decrease until it disappears, with the fire area not exceeding the threshold throughout the entire process.

[0080] Therefore, based on the specific fire scenario corresponding to the first type of preset abnormal sequence, it is possible not only to determine whether a fire has occurred, but also to determine the specific circumstances of the fire. The generated fire alarm information is more detailed, and thus, rescue forces can be allocated more accurately based on this information.

[0081] In this invention, after receiving initial anomaly information from any sub-warning area, image information of the sub-warning area during the corresponding continuous detection period is acquired, and a corresponding first information sequence is generated. This first information sequence is then used to determine whether a fire has occurred, and a corresponding fire alarm is generated. This approach simultaneously considers information from different sensors to jointly determine whether a fire has occurred, allowing for more accurate fire detection through the fusion of multi-source information and preventing false alarms. Furthermore, by using image information from the sub-warning area during the corresponding continuous detection period, a first information sequence reflecting the changing trend of the fire area is generated, further improving the accuracy of the generated fire alarm information and preventing false alarms.

[0082] As another embodiment of the present invention, after S300: determining the continuous monitoring period of the sub-early warning area based on the scene type identifier, the method further includes:

[0083] S310: Obtain smoke concentration information of the sub-early warning area during the continuous monitoring period according to the second preset acquisition frequency W, and generate the second information sequence (B1, B2, ..., B...) of the sub-early warning area. n A f(W) ), among which, B n Let f(W) be the nth smoke concentration information collected in the sub-warning area. f(W) is the total number of smoke concentration information collected in the sub-warning area according to W, where n = 1, 2, ..., f(W).

[0084] S320: If the similarity between the second information sequence of the sub-earning area and any second type of preset abnormal sequence is greater than the second similarity threshold, then the abnormal information corresponding to the second type of preset abnormal sequence shall be taken as the second abnormal information of the sub-earning area.

[0085] S330: Generate fire alarm information for the sub-early warning area based on the first and second abnormal information.

[0086] Specifically, the first anomaly information includes the first fire indicator and the fire type indicator. The second anomaly information includes the second fire indicator and the smoke type indicator.

[0087] The S330 includes:

[0088] S331: If both the first fire indicator and the second fire indicator are first preset indicators, then fire information is generated based on the first and second abnormal information. The fire information includes the location information of the sub-warning area, the fire type indicator, and the smoke type indicator.

[0089] Normally, a fire will inevitably produce a large amount of smoke, so changes in smoke levels can help determine if a fire has occurred. However, in scenarios prone to false alarms, such as when dust and sand are stirred up during cleaning, causing a rapid increase in smoke concentration, or when there are non-flammable items emitting significant heat (such as self-heating pots or high-calorie takeout food) in the area, relying solely on smoke concentration can trigger false alarms. Therefore, to avoid false alarms, a fire alarm is only generated when both the first and second fire indicators are the first preset indicators signifying a fire.

[0090] In addition, to provide a more detailed picture of the fire, fire type and smoke type indicators will be added to the fire information. The fire type indicator is generated based on the type of burning material to indicate the specific type of fire, such as electrical fires, gas fires, solid material fires, etc. The smoke type indicator indicates the concentration of smoke. This allows rescue personnel to make more thorough preparations for rescue operations.

[0091] S332: Based on the fire type identifier and smoke type identifier, generate information on the type of extinguishing material to be delivered by the corresponding sprinkler extinguishing device in the sub-early warning area.

[0092] In addition, based on the fire type and smoke type labels, it is possible to determine what kind of extinguishing material to use for the corresponding fire sprinkler system, so as to extinguish the fire more promptly and effectively.

[0093] As another embodiment of the present invention, after S400: determining the monitoring image acquisition frequency Q of the sub-early warning area according to P, the method further includes:

[0094] S410: During the continuous monitoring period, collect multiple thermal imaging image information of the sub-early warning area according to Q.

[0095] S420: Input multiple thermal imaging image information into the second target recognition model to generate a third information sequence (C1, C2, ..., C...) for the sub-warning area. i ..., C f(Q) ), where C i If the similarity between the three information sequences of the area of ​​the fire region and the total area of ​​the entire image in the i-th thermal imaging image information and any third-class preset anomaly sequence is greater than the first similarity threshold, then the anomaly information corresponding to the third-class preset anomaly sequence is taken as the third anomaly information of the sub-warning region.

[0096] Simultaneously, based on the temperature information of the fire area represented in multiple thermal imaging images, a fire temperature corresponding to each thermal imaging image can be generated, thereby generating a fourth information sequence for the sub-warning area. The fourth information sequence is used to represent the temperature change pattern during the fire process in the sub-warning area, and the presence of this pattern also determines whether a fire has occurred.

[0097] Due to the influence of light or obstruction from other objects, the fire area may be unclear or completely obscured in real-world images captured by conventional cameras. This reduces the accuracy of the information obtained in S600. Therefore, S420 in this embodiment can be used to obtain the corresponding information more accurately, thereby improving the accuracy of the judgment. Correspondingly, S410 and S420 can be used to replace S500 and S600, or S410 and S420 can be used as supplementary information.

[0098] S430: If at least two of the first, second, and third abnormal information represent the same fire information, then fire alarm information for the sub-warning area is generated based on the same fire information.

[0099] Since any single piece of information may result in a false alarm, this embodiment combines the three different dimensions of information mentioned above to improve the accuracy of fire alarm judgment and further reduce the probability of false alarms.

[0100] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0101] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0102] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided.

[0103] Those skilled in the art will understand that various aspects of the present invention can be implemented as systems, methods, or program products. Therefore, various aspects of the present invention can be specifically implemented in the following forms: entirely hardware implementations, entirely software implementations (including firmware, microcode, etc.), or implementations combining hardware and software aspects, collectively referred to herein as “circuits,” “modules,” or “systems.”

[0104] An electronic device according to this embodiment of the invention. The electronic device is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the invention.

[0105] Electronic devices are manifested in the form of general-purpose computing devices. Components of an electronic device may include, but are not limited to: at least one processor, at least one memory, and buses connecting different system components (including memory and processor).

[0106] The memory stores program code that can be executed by a processor, causing the processor to perform the steps described in the "Exemplary Methods" section above, according to various exemplary embodiments of the present invention.

[0107] The storage may include readable media in the form of volatile storage, such as random access memory (RAM) and / or cache memory, and may further include read-only memory (ROM).

[0108] The storage may also include programs / utilities having a set (at least one) of program modules, including but not limited to: an operating system, one or more applications, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0109] A bus can represent one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus that uses any of the various bus architectures.

[0110] The electronic device can also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication can be performed via input / output (I / O) interfaces. Furthermore, the electronic device can communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter. The network adapter communicates with other modules of the electronic device via a bus. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0111] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0112] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of the present invention may also be implemented as a program product comprising program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps of the various exemplary embodiments of the present invention described in the "Exemplary Methods" section above.

[0113] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0114] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0115] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0116] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0117] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0118] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0119] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for generating fire alarm information based on multi-source data fusion, characterized in that, The method includes the following steps: In response to receiving primary anomaly information from any sub-warning area, the system acquires first real-scene image information of the sub-warning area; the primary anomaly information is either smoke concentration anomaly information or target temperature anomaly information. The first real-scene image information corresponding to the sub-early warning area is input into the first target recognition model to generate the scene type identifier and scene confidence P corresponding to the sub-early warning area; The continuous monitoring period for the sub-early warning area is determined based on the scenario type identifier; Based on P, the monitoring image acquisition frequency Q of the sub-early warning area is determined; Q satisfies the following condition: Where K1 and K2 are the first and second adjustment coefficients, respectively; S is the upper limit of the number of data collections. During the continuous monitoring period, multiple second real-scene image information of the sub-early warning area are collected according to Q; Multiple second real-scene image information are respectively input into the second target recognition model to generate the first information sequence (A1, A2, ..., A1) of the sub-early warning area. i A f(Q) ); where A i Let f(Q) be the ratio between the area of ​​the fire zone in the i-th second real-scene image and the total area of ​​the entire image; f(Q) is the total number of second real-scene image information collected according to Q, i=1, 2, ..., f(Q); If the similarity between the first information sequence of the sub-early warning area and any first type of preset abnormal sequence is greater than the first similarity threshold, then the abnormal information corresponding to the first type of preset abnormal sequence is taken as the first abnormal information of the sub-early warning area. Based on the first abnormal information, fire alarm information for the sub-early warning area is generated; Based on the scenario type identifier, the continuous monitoring period for the sub-early warning area includes: Based on the scenario type identifier, the continuous monitoring period of the sub-early warning area is obtained from the first mapping table; The first mapping table includes the mapping relationship between different scene type identifiers and corresponding preset continuous monitoring periods; After determining the continuous monitoring period of the sub-early warning area based on the scenario type identifier, the method further includes: According to the second preset acquisition frequency W, the smoke concentration information of the sub-early warning area during the continuous monitoring period is acquired, and a second information sequence (B1, B2, ..., B1, ..., B2) of the sub-early warning area is generated. n A f(W) ); where B n f(W) represents the nth smoke concentration information collected in the sub-warning area; f(W) represents the total number of smoke concentration information collected in the sub-warning area according to W, where n = 1, 2, ..., f(W); If the similarity between the second information sequence of the sub-early warning region and any second type of preset abnormal sequence is greater than the second similarity threshold, then the abnormal information corresponding to the second type of preset abnormal sequence is taken as the second abnormal information of the sub-early warning region. Based on the first and second abnormal information, fire alarm information for the sub-early warning area is generated.

2. The method according to claim 1, characterized in that, Before acquiring the first real-scene image information of the sub-warning area in response to receiving primary anomaly information of any sub-warning area, the method further includes: According to the first preset acquisition frequency, the thermal imaging information of each sub-early warning area in the target area is acquired cyclically. If the target temperature in the thermal imaging information of any of the sub-warning areas is greater than the temperature threshold, then target temperature anomaly information corresponding to the sub-warning area is generated.

3. The method according to claim 1, characterized in that, Each of the aforementioned sub-warning areas is equipped with a corresponding smoke sensor to collect smoke concentration information within the sub-warning area; Before acquiring the first real-scene image information of the sub-warning area in response to receiving primary anomaly information of any sub-warning area, the method further includes: If the smoke concentration information in any of the sub-warning areas is greater than the smoke concentration threshold, then abnormal smoke concentration information corresponding to the sub-warning area is generated.

4. The method according to claim 1, characterized in that, Multiple sets of second real-scene image information are respectively input into a second target recognition model to generate a first information sequence of the sub-warning area, including: Multiple second real-scene image information are respectively input into the second target recognition model to generate a target recognition box corresponding to the fire area in each second real-scene image information; The ratio between the total area of ​​all target recognition boxes in each of the second real-scene image information and the total area of ​​the second real-scene image information is obtained to generate the first information sequence of the sub-early warning area.

5. The method according to claim 1, characterized in that, After determining the monitoring image acquisition frequency Q of the sub-early warning area based on P, the method further includes: During the continuous monitoring period, multiple thermal imaging image information of the sub-early warning area are collected according to Q; The thermal imaging image information is input into the second target recognition model to generate the third information sequence (C1, C2, ..., C...) of the sub-early warning area. i ..., C f(Q) ); where C i If the similarity between the three information sequences of the area of ​​the fire region and the total area of ​​the entire image in the i-th thermal imaging image information and any third type of preset abnormal sequence is greater than the first similarity threshold, then the abnormal information corresponding to the third type of preset abnormal sequence is taken as the third abnormal information of the sub-early warning area. If at least two of the first, second, and third abnormal information represent the same fire information, then fire alarm information for the sub-early warning area is generated based on the same fire information.

6. The method according to claim 5, characterized in that, The first abnormal information includes a first fire indicator and a fire type indicator; the second abnormal information includes a second fire indicator and a smoke type indicator. Based on the first and second anomaly information, the fire alarm information for the sub-early warning area is generated as follows: If both the first fire indicator and the second fire indicator are first preset indicators, then fire information is generated based on the first abnormal information and the second abnormal information; the fire information includes the location information of the sub-warning area, the fire type indicator, and the smoke type indicator; Based on the fire type identifier and smoke type identifier, the type information of the fire extinguishing material to be delivered by the corresponding sprinkler fire extinguishing device in the sub-early warning area is generated.

7. A non-transitory computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a fire alarm information generation method based on multi-source data fusion as described in any one of claims 1 to 6.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a fire alarm information generation method based on multi-source data fusion as described in any one of claims 1 to 6.

Images