A smoke identification method, a smoke identification device and a storage medium

Abstract

The present application provides a smoke identification method, a smoke identification device and a storage medium, wherein the smoke identification method includes: in response to a suspected smoke alarm being triggered by a current monitoring image, extracting the location information and image feature information corresponding to the current monitoring image as the current suspected smoke alarm information; performing image feature information comparison between the current suspected smoke alarm information and the historical suspected smoke alarm information with the same location information; and determining the accuracy of the current suspected smoke alarm information based on the result of the image feature information comparison. In the above manner, a secondary analysis of the suspected smoke alarm can be performed to achieve an accuracy evaluation of the suspected smoke alarm.

Description

Technical Field

[0001] The present application relates to image processing technology and fire protection fields, and in particular to a smoke recognition method, a smoke recognition device and a storage medium. Background Art

[0002] Among various disasters, fire is one of the major disasters that most frequently and commonly threatens public safety and social development.

[0003] On April 3, 2022, the Fire and Rescue Bureau of the Ministry of Emergency Management reported that in the first quarter of this year, a total of 219,000 fires were reported nationwide, with a total of 625 deaths and 397 injuries, and direct property losses of 1.52 billion yuan. Compared with the same period last year, the number of fires, casualties and losses decreased by 11.6%, 8.5% and 20.5% respectively, and the number of deaths increased by 4.9%; among them, there were 29 major fires, an increase of 6 over the same period last year.

[0004] In the prior art, fire monitoring is performed using a camera, which mainly uses the monitoring image / video taken by the camera and then uses image recognition technology to determine the smoke in the monitoring image / video to identify the fire. However, smoke recognition may have false alarms, which greatly affects the accuracy of fire monitoring. Summary of the invention

[0005] The main technical problem solved by the present application is to provide a smoke identification method, a smoke identification device and a storage medium, which can perform secondary analysis on suspicious smoke alarms and achieve accuracy evaluation of suspicious smoke alarms.

[0006] To solve the above technical problems, a technical solution adopted in the present application is: in response to the current monitoring image triggering a suspected smoke alarm, extracting the location information and image feature information corresponding to the current monitoring image as the current suspected smoke alarm information; comparing the image feature information of the current suspected smoke alarm information with the historical suspected smoke alarm information with the same location information; and determining the accuracy of the current suspected smoke alarm information based on the result of the image feature information comparison.

[0007] Among them, the image feature information of the location information in the current suspected smoke alarm information is compared with the historical suspected smoke alarm information with the same location information, including: comparing the location information in the current suspected smoke alarm information with the location information of the historical suspected smoke alarm information; in response to the current suspected smoke alarm information having the same location information as the historical suspected smoke alarm information, comparing the image feature information of the current suspected smoke alarm information with the image feature information of the historical suspected smoke alarm information.

[0008] The method further includes: in response to the current suspected smoke alarm information being different from the location information of the historical suspected smoke alarm information, using the current suspected smoke alarm information to update the historical suspected smoke alarm information.

[0009] The monitoring image is captured by a spherical camera, and the location information is the PTZ information and GPS information of the spherical camera.

[0010] The accuracy of the current suspected smoke alarm information is determined according to the result of the image feature information comparison, including: in response to the feature difference of the image feature information comparison being greater than a set threshold, determining that the current suspected smoke alarm information is a real smoke alarm.

[0011] In response to determining that the current suspected smoke alarm information is a real smoke alarm, the historical suspected smoke alarm information having the same location information as the current suspected smoke alarm information is deleted.

[0012] Among them, in response to the feature difference of the image feature information comparison being less than a set threshold, the suspected alarm times of the current suspected smoke alarm information and the historical suspected smoke alarm information with the same location information are obtained; and the accuracy of the current suspected smoke alarm information is determined based on the suspected alarm times.

[0013] The accuracy of the current suspected smoke alarm information is determined according to the number of suspected alarms, including: in response to the number of suspected alarms being greater than a set number, determining that the current suspected smoke alarm information is a false smoke alarm.

[0014] In response to determining that the current suspected smoke alarm information is a false smoke alarm, the historical suspected smoke alarm information having the same location information as the current suspected smoke alarm information is deleted.

[0015] In which, in response to the number of suspected alarms being less than the set number, the current suspected smoke alarm information is saved as historical suspected smoke alarm information.

[0016] The method further includes: establishing a smoke recognition model; obtaining suspected smoke alarm information as training samples, and labeling the suspected smoke alarm information based on a determined degree of accuracy; and inputting the training samples into the smoke recognition model to train the smoke recognition model.

[0017] In order to solve the above technical problems, another technical solution adopted by the present application is: providing a smoke identification device, including a memory for storing a computer software program;

[0018] The processor is used to read and execute the computer software program to implement the steps of the above-mentioned smoke identification method.

[0019] In order to solve the above technical problems, another technical solution adopted by the present application is: to provide a non-transitory computer scale storage medium to store a computer software program of the above smoke identification method steps.

[0020] The beneficial effects of the present application are as follows: the present application provides a smoke identification method, device and computer-readable storage medium, the method comprising: in response to a suspected smoke alarm being triggered by a current monitoring image, extracting the location information and image feature information corresponding to the current monitoring image as the current suspected smoke alarm information; performing image feature information comparison between the current suspected smoke alarm information and the historical suspected smoke alarm information with the same location information; and determining the accuracy of the current suspected smoke alarm information according to the result of the image feature information comparison. In the above manner, the suspected smoke alarm information is subjected to secondary analysis to achieve an accuracy evaluation of the suspected alarm, and to timely eliminate false smoke alarm information; at the same time, the scalability and incomplete consistency of smoke at different times at the same location are used to eliminate suspicious smoke false alarms, thereby reducing the probability of frequent false alarms at the same location during detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0022] Figure 1 This is a flow chart of a method for smoke identification provided in an embodiment of the present application;

[0023] Figure 2 yes Figure 1 Schematic diagram of the process of step S12;

[0024] Figure 3 It is a flow chart of a smoke identification method provided in an embodiment of the present application;

[0025] Figure 4 is a flow chart of another embodiment of the smoke identification method provided by the present application;

[0026] Figure 5 This is a structural schematic diagram of a smoke identification device of the present application;

[0027] Figure 6 It is a structural schematic diagram of the computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0029] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0030] The first aspect of the present application provides a smoke identification method, which, in response to a suspected smoke alarm being triggered by a current monitoring image, extracts location information and image feature information corresponding to the current monitoring image as current suspected smoke alarm information; compares the image feature information of the current suspected smoke alarm information with historical suspected smoke alarm information with the same location information; and determines the accuracy of the current suspected smoke alarm information based on the result of the image feature information comparison.

[0031] See also Figure 1 , Figure 1 : is a flow chart of a smoke identification method provided in an embodiment of the present application, the method comprising:

[0032] S11, in response to the current monitoring image triggering a suspected smoke alarm, extracting position information and image feature information corresponding to the current monitoring image as current suspected smoke alarm information.

[0033] Among them, the surveillance images are captured by surveillance cameras deployed in public places.

[0034] Optionally, in one embodiment, the monitoring camera may be a ball camera, and during the deployment of the ball camera, the ball camera is calibrated with the world coordinate position to achieve correspondence between the image coordinates of the ball camera and the physical geographic location information.

[0035] In a specific solution, the camera is installed horizontally when deployed, the horizontal angle of the gimbal is 0 degrees in the north direction, and supports horizontal rotation of [0,360] degrees; select a point on the ground as calibration point A, adjust the camera angle so that the calibration point is in the center of the picture, and record the camera angle (P0, T0) and the longitude and latitude of the calibration point (X0, Y0); select a point on the ground as calibration point B, adjust the camera angle so that the calibration point is in the center of the picture, and record the camera angle (P1, T1) and the longitude and latitude of the calibration point (X1, Y1); M0 is the horizontal unit angle corresponding to the longitude length, N0 is the vertical unit angle corresponding to the latitude length, where the calculation formula is , complete the calibration of the ball camera and world coordinates, and realize the correspondence between the ball camera image coordinates and the physical geographic location information.

[0036] Specifically, the surveillance camera obtains surveillance images of public places according to a preset shooting frequency, and the background smoke recognition device obtains the surveillance images taken by the surveillance camera and performs smoke recognition on the surveillance images to determine whether a suspected smoke alarm is triggered.

[0037] Optionally, the smoke recognition device is specifically used to extract feature information of the monitoring image, and then identify the special information to determine whether it is smoke, and if it is smoke, trigger a suspected smoke alarm.

[0038] In one embodiment, the ball camera performs smoke detection in cruise mode, obtains visible light image and thermal imaging image information, and analyzes the obtained image information through an AI fireworks recognition algorithm. When the image information obtained by the ball camera triggers a suspected smoke alarm, the current position information of the ball camera and the obtained image feature information are recorded to form a suspected smoke alarm base.

[0039] S12, comparing the image feature information of the current suspected smoke alarm information with the historical suspected smoke alarm information with the same location information.

[0040] In one embodiment, the current suspected smoke alarm information is compared with the historical suspected smoke alarm information in the suspected smoke alarm base library to determine whether the current suspected smoke alarm information and the historical suspected smoke alarm information in the base library are in the same area (i.e., the same position). If they are in the same area, a similarity comparison is performed between the current suspected smoke alarm image features and the image features corresponding to the same position in the base library; if they are not in the same area, the current suspected smoke alarm information is stored in the suspected smoke alarm base library.

[0041] Specifically, Figure 2 As shown, Figure 2 yes Figure 1 The schematic diagram of the process of step S12 in FIG. 1 , step S12 may specifically include:

[0042] Step S121: Compare the location information in the current suspected smoke alarm information with the location information of the historical suspected smoke alarm information.

[0043] Optionally, in an embodiment where the monitoring camera is a PTZ camera, the location information includes the PTZ information of the PTZ camera and the physical location information.

[0044] Among them, the PTZ information of the PTZ camera is the horizontal (left - right), vertical (up - down) information of the PTZ of the PTZ camera and the zoom information of the camera lens; the physical location information is the GPS information, that is, the global positioning information, generated after the corresponding conversion of the image coordinates and the physical geographical information when the PTZ camera acquires the smoke image.

[0045] In an embodiment, the location information of the suspected smoke alarm information includes the smoke PTZ information (P2, T2, Z2) and the geographical location information (X2, Y2), and is compared and discriminated with the corresponding smoke PTZ information (Pi, Ti, Zi) and the geographical location information (Xi, Yi) in the suspected smoke alarm database.

[0046] Optionally, set a threshold Z, and the location discriminant formula is pi = |P2 - Pi|, ti = |T2 - Ti|, xi = |X2 - Xi|, yi = |Y2 - Yi|. If pi < Z, ti < Z, xi < Z, and yi < Z are satisfied simultaneously, it is considered that there is a historical suspected smoke alarm at the same location in the suspected smoke alarm information database for the suspected smoke alarm.

[0047] Step S122: In response to the location information in the current suspected smoke alarm information being the same as the location information of the historical suspected smoke alarm information, compare the image feature information of the current suspected smoke alarm information with the image feature information of the historical suspected smoke alarm information.

[0048] Among them, comparing the image feature information of the current suspected smoke alarm information with the image feature information of the historical suspected smoke alarm information mainly compares the differences between the two image feature information. For example, this comparison of image feature differences includes the determination of feature similarities such as feature extraction by neural networks, smoke texture contours, and smoke area sizes.

[0049] In an embodiment, set a single or multiple feature difference thresholds. If the feature difference information is greater than the feature difference threshold, the credibility of the smoke alarm at this location is extremely high.

[0050] For example, cosine similarity calculation is performed on the suspected smoke image features and the historical suspected smoke image features corresponding to the same position. Among them, the cosine similarity uses the cosine value of the angle between two vectors in the vector space as a measure of the difference between two image individuals. The closer the cosine value is to 1, that is, the more similar the two vectors are, and the closer the cosine value is to 0, that is, the less similar the two vectors are, that is, the less similar the two images are. If the cosine similarity between the suspected smoke image features and the historical suspected smoke image features corresponding to the same position is greater than a specific threshold, it indicates that the suspected smoke features at this position have changed.

[0051] For another example, Canny edge detection is used to extract the smoke contour, the minimum external contour of the suspected smoke contour and the historical suspected smoke corresponding to the same position is obtained, the centers of the two contours are coincided, and the reciprocal of the intersection ratio of the two contours is calculated. The larger the intersection ratio of the contours is, the higher the similarity between the suspected smoke contour and the historical suspected smoke contour is. If the reciprocal of the intersection ratio is greater than the corresponding difference threshold N, it indicates that the similarity of the smoke contours is lower.

[0052] For example, similarity judgment can be performed based on multiple smoke feature information. For example, the value obtained by weighted summation of the similarities of each corresponding information is used as the final feature similarity and compared with the feature difference threshold M; for another example, for the multiple smoke feature information, difference thresholds for the multiple smoke feature information are set, the deviation between the similarity of each smoke feature information and the feature threshold is compared, and the accuracy of the suspected smoke alarm in this area is comprehensively determined based on the similarities of multiple information among several pieces of information.

[0053] Step S123: In response to the current suspected smoke alarm information being different from the position information of the historical suspected smoke alarm information, update the historical suspected smoke alarm information with the current suspected smoke alarm information.

[0054] In one embodiment, if the position discriminant formula cannot simultaneously satisfy pi < Z, ti < Z, xi < Z, yi < Z, it is considered that there is no historical suspected smoke alarm at the same position in the suspected smoke alarm information database for the current suspected smoke alarm. Therefore, the current suspected smoke alarm information is saved as historical suspected smoke alarm information and added to the suspected smoke alarm information database.

[0055] S13. Determine the accuracy of the current suspected smoke alarm information according to the result of the comparison of the image feature information.

[0056] Among them, the similarity between the current suspected smoke alarm image and the image features corresponding to the same position in the database is compared. If the feature difference value is greater than the set threshold, the current suspected smoke alarm is determined to be a real smoke.

[0057] In one embodiment, S13 may include:

[0058] In response to a feature difference in the image feature information comparison being greater than a set threshold, it is determined that the current suspected smoke alarm information is a real smoke alarm.

[0059] Specifically, if the feature difference value between the current suspected smoke alarm image and the historical suspected smoke alarm image at the same location is greater than a threshold value M, based on the characteristics of smoke extensibility and incomplete consistency at different times at the same location, the suspected smoke alarm information is confirmed to be real smoke alarm information, and all suspicious smoke alarm information at the same location is promptly eliminated, and the smoke alarm information is pushed in a timely manner.

[0060] Further, after determining that the current suspected smoke alarm information is a real smoke alarm, it can be considered that the location information indicates that a real fire has occurred, and the historical suspected smoke alarm information with the same location information as the current suspected smoke alarm information can be deleted.

[0061] In another embodiment, S13 may also include: in response to the feature difference of the image feature information comparison being less than the set threshold, obtaining the suspected alarm times of the current suspected smoke alarm information and the historical suspected smoke alarm information with the same location information; in response to the suspected alarm times being greater than the set times, determining that the current suspected smoke alarm information is a false smoke alarm, and deleting the historical suspected smoke alarm information with the same location information as the current suspected smoke alarm information.

[0062] It can be understood that when the number of historical suspicious smoke alarms is greater than the set alarm number threshold, based on the characteristics of smoke as a non-rigid object, the corresponding features at different times at the same location are different, and there are suspicious smoke alarms in the same area for a long time and the suspicious smoke image feature information has not changed significantly, then the suspicious smoke alarm in the area is considered to be an erroneous smoke alarm information. The historical suspected alarm information at the same location is determined to be an erroneous or non-smoke event, all suspicious smoke alarm information at the same location is promptly removed, and the location information and image feature information of the erroneous smoke alarm area are recorded.

[0063] In another embodiment, S13 may also include: in response to the feature difference of the image feature information comparison being less than the set threshold, obtaining the current suspected smoke alarm information and the number of suspected alarms of the historical suspected smoke alarm information with the same location information; in response to the feature difference of the image feature information comparison being less than the set threshold, obtaining the current suspected smoke alarm information and the number of suspected alarms of the historical suspected smoke alarm information with the same location information; in response to the number of suspected alarms being less than the set number, saving the current suspected smoke alarm information as the historical suspected smoke alarm information.

[0064] It can be understood that if the feature difference value between the current suspected smoke alarm image and the historical suspected smoke alarm image at the same location is not greater than the threshold, and the number of historical suspicious smoke alarms does not exceed the set alarm threshold number, then the suspicious smoke alarm information cannot be regarded as real smoke alarm information or false alarm information, and the suspicious smoke alarm information is regarded as historical suspected smoke alarm information and stored in the suspicious smoke alarm information base, and the number of historical suspicious smoke alarms at the current location is increased by 1.

[0065] Understandably, based on the fact that smoke is a non-rigid object, the image features corresponding to the same location at different times will change. If the image features of suspected smoke alarms at the same location at different times are quite different, it can be considered that the credibility of the suspected smoke alarm here is extremely high, and the suspected smoke alarm here is real smoke; if the image features of suspected smoke alarms at the same location at different times are not very different, and are less than the threshold of the number of suspicious alarms, it is considered that this is still a suspected smoke alarm; if the image features of suspected smoke alarms at the same location at different times are not very different, and reach the threshold of the number of suspicious alarms, it means that the suspected smoke alarms at the same location for a long time are false smoke alarms.

[0066] Combined with the above Figure 1 and Figure 2 For an example, see Figure 3 , Figure 3 : is a flow chart of a smoke identification method provided in an embodiment of the present application, the method comprising:

[0067] A: Smoke alarm detected;

[0068] B: location information, image feature extraction;

[0069] C: Is the position of the smoke alarm information in the suspected smoke alarm base database the same? If so, go to step F; otherwise, go to step D;

[0070] D: generate suspected smoke alarm data and store it in the suspected smoke alarm database;

[0071] E: Compare the image feature information with the historical suspected smoke alarm information at the same location;

[0072] F: Is the feature difference greater than the threshold? If so, execute step G; otherwise, execute step I;

[0073] G: The smoke alarm is real smoke;

[0074] H: Eliminate suspected smoke alarm information at the same location;

[0075] I: Record the number of suspected smoke alarms at the current location;

[0076] J: Whether the number of smoke alarms is greater than the threshold, if so, execute step K, otherwise execute step M;

[0077] K: The smoke alarm is real smoke;

[0078] L: Eliminate suspected smoke alarm information at the same location;

[0079] M: Add suspected smoke alarm information to the suspected smoke alarm database.

[0080] See also Figure 4 , Figure 4 : is a flow chart of another embodiment of the smoke identification method provided by the present application, the method comprising:

[0081] S41: Establish a smoke recognition model.

[0082] In one embodiment, the smoke recognition model may be a K-means clustering model, a support vector machine model, a convolutional neural network model, a deep reinforcement learning model, etc., to achieve smoke detection and accuracy determination of the input image.

[0083] S42: Obtain suspected smoke alarm information as a training sample, wherein the training sample labels the suspected smoke alarm information based on the determined accuracy.

[0084] In one embodiment, the smoke recognition model analyzes suspicious smoke alarm information detected in the actual environment, identifies real smoke alarm information and false smoke alarm information, and uses the real smoke alarm information and false smoke alarm information collected in the actual environment as training samples of the model to improve the reliability of the model.

[0085] S43: Inputting the training sample into the smoke recognition model to train the smoke recognition model.

[0086] It can be understood that inputting real smoke samples and negative samples collected at the actual scene into the smoke recognition model for training can effectively improve the robustness and generalization ability of the smoke recognition model. Regular model update deployment can greatly reduce the probability of frequent false alarms of the smoke recognition model at the same location.

[0087] In one embodiment, the real smoke alarm data and false smoke alarm data identified by the smoke recognition device are input into the smoke recognition model. The recognition model is trained with real smoke alarm samples and false smoke alarm samples collected from the scene environment, which can effectively improve the robustness and scene generalization ability of the smoke recognition model.

[0088] The second aspect of the present application provides a smoke identification device 20, see Figure 5 , Figure 5This is a structural schematic diagram of the smoke identification device of the present application.

[0089] The smoke identification device 20 includes a memory 21 and a processor 22 coupled to each other, and the processor 22 is used to execute program instructions stored in the memory 21 to implement the method provided by any of the above-mentioned embodiments of the present application and any non-conflicting combination.

[0090] In a specific implementation scenario, the smoke identification device 20 may include, but is not limited to: a microcomputer, a server, and in addition, the smoke identification device 20 may also include mobile devices such as a laptop computer and a tablet computer, which are not limited here.

[0091] Specifically, the processor 22 is used to control itself and the memory 21 to implement the steps of the above method embodiment. The processor 22 can also be called a CPU (Central Processing Unit). The processor 22 may be an integrated circuit chip with signal processing capabilities. The processor 22 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 22 can be implemented by an integrated circuit chip.

[0092] The third aspect of the present application also provides a computer-readable storage medium 30, see Figure 6 , Figure 6 It is a structural schematic diagram of the computer-readable storage medium of the present application.

[0093] The computer-readable storage medium 30 stores program instructions 31 , and when the program instructions 31 are executed by a processor, the method provided by any one of the first aspect or the second aspect is implemented.

[0094] The computer-readable storage medium 30 may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0095] Reference to an embodiment herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments in the absence of conflict.

[0096] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A smoke identification method, characterized in that: The method comprises: In response to the current monitoring image triggering a suspected smoke alarm, extracting position information and image feature information corresponding to the current monitoring image as current suspected smoke alarm information; Comparing the image feature information of the current suspected smoke alarm information with the historical suspected smoke alarm information having the same location information; Determining the accuracy of the current suspected smoke alarm information according to the result of the image feature information comparison, wherein determining the accuracy of the current suspected smoke alarm information according to the result of the image feature information comparison includes: In response to a feature difference in the image feature information comparison being greater than a set threshold, determining that the current suspected smoke alarm information is a real smoke alarm; In response to the feature difference of the image feature information comparison being less than the set threshold, obtaining the current suspected smoke alarm information and the suspected alarm times of the historical suspected smoke alarm information with the same location information; In response to the number of suspected alarms being greater than a set number, determining that the current suspected smoke alarm information is a false smoke alarm; In response to determining that the current suspected smoke alarm information is a false smoke alarm, the historical suspected smoke alarm information having the same location information as the current suspected smoke alarm information is deleted.

2. The method according to claim 1, characterized in that: The comparing the image feature information of the current suspected smoke alarm information with the historical suspected smoke alarm information having the same location information includes: Comparing the location information in the current suspected smoke alarm information with the location information in the historical suspected smoke alarm information; In response to the location information of the current suspected smoke alarm information being the same as the location information of the historical suspected smoke alarm information, the image feature information of the current suspected smoke alarm information is compared with the image feature information of the historical suspected smoke alarm information.

3. The method according to claim 2, characterized in that The method further comprises: In response to the location information of the current suspected smoke alarm information being different from the location information of the historical suspected smoke alarm information, the historical suspected smoke alarm information is updated using the current suspected smoke alarm information.

4. The method according to any one of claims 1 to 3, characterized in that: The monitoring image is captured by a spherical camera, and the location information is the PTZ information and GPS information of the spherical camera.

5. The method according to claim 1, characterized in that The method further comprises: In response to determining that the current suspected smoke alarm information is a real smoke alarm, the historical suspected smoke alarm information having the same location information as the current suspected smoke alarm information is deleted.

6. The method according to claim 1, characterized in that The method further comprises: In response to the suspected alarm number being less than the set number, the current suspected smoke alarm information is updated to historical suspected smoke alarm information.

7. The method according to claim 1, characterized in that The method further comprises: Building a smoke recognition model; and Acquire suspected smoke alarm information as a training sample, wherein the training sample labels the suspected smoke alarm information based on the determined accuracy; The training samples are input into the smoke recognition model to train the smoke recognition model.

8. A smoke identification device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program to implement the smoke identification method according to any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program for implementing the smoke identification method according to any one of claims 1 to 7.

Images