Smoke detection method and device, electronic equipment and storage medium

Abstract

The invention discloses a smoke detection method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a reference image corresponding to a to-be-detected area and at least three target images; determining center position information and size information of target areas in the at least three target images; determining whether the target area triggers a first preset condition according to the center position information of the target area in the at least three target images; determining whether the target area triggers a second preset condition according to the size information of the target area in the at least three target images; and if the target area simultaneously triggers the first preset condition and the second preset condition, triggering a smoke alarm event in the to-be-detected area. According to the method, the center position information and the size information of the target area are compared with the first preset condition and the second preset condition respectively for smoke detection, the indoor scene smoke detection precision is improved, the recall rate of the smoke target is increased through quantitative judgment, and therefore the problem that smoke in the early stage of a fire cannot be effectively recognized is solved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a smoke detection method, device, electronic equipment and storage medium. Background Art

[0002] With the popularization of image detection technology, cameras are replacing manual labor as the main patrol and defense tool in more and more scenarios, playing an important role in preventing theft and monitoring anomalies.

[0003] For the recent frequent fire incidents of electric vehicle charging piles, deploying corresponding behavior detection algorithms can better detect fires, thereby saving manpower and reducing losses. By reviewing relevant cases, it is found that these fires generally occur in indoor scenes. In the early stage of a fire, a large amount of smoke will be generated due to battery combustion and the spread area is wide. A larger flame will only appear after the thick smoke.

[0004] Related solutions are usually based on image detection and have high requirements on the environment. For example, the lighting conditions in the detection scene must be good and the background and the object to be detected must be distinguishable. However, indoor scenes such as parking lots rarely have lights on for a long time in each parking space. This causes some parking spaces to be dimly lit, which greatly reduces the recognition rate of image-based detection. In some cases, even if the smoke shape is good, only a small number of targets can be reported. It is impossible to report smoke as soon as it is generated. Smoke in the early stage of a fire cannot be effectively identified, and only the fire can be detected. Summary of the invention

[0005] The present invention provides a smoke detection method, device, electronic equipment and storage medium to solve the problem that smoke in the early stage of a fire cannot be effectively identified.

[0006] According to one aspect of the present invention, a smoke detection method is provided, the method comprising:

[0007] Determine a reference image and at least three target images corresponding to the area to be detected;

[0008] Determine the center position information and size information of the target area in the at least three target images, the target area being a suspected smoke area determined by the image difference between the target image and the reference image;

[0009] Determining whether the target area triggers a first preset condition based on the center position information of the target area in the at least three target images, wherein the first preset condition is used to determine whether a change trend of the center position of the target area over time matches a change trend of the center position of an actual smoke area;

[0010] Determining whether the target area triggers a second preset condition based on the size information of the target area in the at least three target images, wherein the second preset condition is used to determine whether a size change trend of the target area over time matches a size change trend of an actual smoke area;

[0011] If the target area triggers the first preset condition and the second preset condition at the same time, a smoke alarm event is triggered for the area to be detected.

[0012] According to another aspect of the present invention, there is provided a smoke detection device, the device comprising:

[0013] A detection module, used to determine a reference image and at least three target images corresponding to the area to be detected;

[0014] An information determination module, used to determine the center position information and size information of a target area in the at least three target images, wherein the target area is a suspected smoke area determined by the image difference between the target image and the reference image;

[0015] A first judgment module, used to determine whether the target area triggers a first preset condition based on the center position information of the target area in the at least three target images, wherein the first preset condition is used to determine whether the change trend of the center position of the target area over time matches the change trend of the center position of the actual smoke area;

[0016] A second judgment module is used to determine whether the target area triggers a second preset condition based on the size information of the target area in the at least three target images, and the second preset condition is used to determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area;

[0017] The early warning module is used to trigger a smoke alarm event for the target area if the target area triggers the first preset condition and the second preset condition at the same time.

[0018] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0019] at least one processor; and

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the smoke detection method described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the smoke detection method described in any embodiment of the present invention when executed.

[0023] The technical solution of the embodiment of the present invention determines a reference image and at least three target images corresponding to the area to be detected, and determines the target area as the suspected smoke area according to the image difference between the target image and the reference image, so as to facilitate the subsequent determination of the change trend of the center position information and size information of the target area in at least three target images, and match the change trend of the center position information and size information of the actual smoke area to determine whether the target area triggers the first preset condition and the second preset condition, thereby realizing accurate detection of smoke, and when the first preset condition and the second preset condition are triggered at the same time in the target area, a smoke alarm event is triggered for the area to be detected, that is, at this time, the target area is accurately determined to be smoke, thereby improving the smoke detection accuracy of indoor scenes, and improving the recall rate of smoke targets through quantitative judgment, thereby solving the problem that smoke in the early stage of a fire cannot be effectively identified.

[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, 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 invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1a is a flow chart of a smoke detection method provided according to an embodiment of the present invention;

[0027] Figure 1b are a reference image and a target image applicable to the embodiment of the present invention;

[0028] Figure 1c is a target image for which the embodiment of the present invention is applicable;

[0029] Figure 1d is a target circumscribed rectangle of a target area of ​​a target image to which the embodiment of the present invention is applicable;

[0030] Figure 2a is a flow chart of a smoke detection method provided according to an embodiment of the present invention;

[0031] Figure 2b is a center coordinate point diagram applicable to an embodiment of the present invention;

[0032] Figure 2c is a center position displacement diagram corresponding to the center coordinate point diagram applicable to an embodiment of the present invention;

[0033] Figure 2d is a vector diagram generated according to a center coordinate point diagram of a smoke target applicable to an embodiment of the present invention;

[0034] Figure 2e is a vector diagram generated from a center coordinate point diagram of a non-smoke target applicable to an embodiment of the present invention;

[0035] Figure 2f is an example diagram of first preset classifier training data applicable to an embodiment of the present invention;

[0036] Figure 3a is a flow chart of a smoke detection method provided according to an embodiment of the present invention;

[0037] Figure 3b is a dimensional change diagram applicable to an embodiment of the present invention;

[0038] Figure 3c is a diagram of the size change of smoke applicable to an embodiment of the present invention;

[0039] Figure 4 is a flow chart of a smoke detection method provided according to an embodiment of the present invention;

[0040] Figure 5 is a structural schematic diagram of a smoke detection device provided according to an embodiment of the present invention;

[0041] Figure 6 It is a structural schematic diagram of an electronic device for implementing the smoke detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0043] It should be noted that the terms "first", "second", "target" and "reference" in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] Figure 1a This is a flow chart of a smoke detection method provided by an embodiment of the present invention. This embodiment is applicable to the case of detecting smoke generated in indoor scenes (especially smoke that has just been generated). The method can be executed by a smoke detection device, which can be implemented in the form of hardware and / or software. The smoke detection device can be configured in any electronic device with network communication function. Figure 1a As shown, the method includes:

[0045] S110, determining a reference image and at least three target images corresponding to the area to be detected.

[0046] The reference image is the first frame image among multiple frame images continuously acquired at a preset interval of the detection area, and the at least three target images are the remaining images except the first frame image among the multiple frame images continuously acquired.

[0047] The present application is applicable to smoke detection in indoor scenes where smoke is initially generated, that is, detection of relatively small smoke that is not easily detected and may appear like bubbles.

[0048] Specifically, a plurality of frames of images are obtained by continuously capturing an area to be detected (the area to be detected is an indoor area with a light intensity less than a preset light intensity) at a preset interval, and then the first frame of the continuously captured multiple frames of images is determined as a reference image, and the remaining at least three frames of the continuously captured multiple frames of images except the first frame are determined as at least three target images.

[0049] For example, Figure 1b As shown, the preset interval can be set according to actual needs, for example, it can be 1 second, then 6 consecutive frames of images of the area to be detected are acquired at an interval of 1 second, and are recorded as frame 0, frame 1, frame 2, frame 3, frame 4 and frame 5 respectively, frame 0 is recorded as a reference image, and frame 1, frame 2, frame 3, frame 4 and frame 5 are recorded as target images.

[0050] S120, determining central position information and size information of a target area in at least three target images, where the target area is a suspected smoke area determined by an image difference between the target image and a reference image.

[0051] The center position information may be the position information of the center point of the target area, for example, if the target area is a rectangle, the center position information is the center of the rectangle. The size information is the boundary information of the target area, that is, if the target area is a rectangle, the size information is the length and width of the rectangle.

[0052] Specifically, the suspected smoke area determined by the image difference between the target image and the reference image is used as the target area, so that the smoke emitted indoors can be effectively detected, and then the center position information and size information of the target area in at least three target images can be quantitatively determined, so that the diffusion and movement of the smoke can be quantitatively determined based on the center position information and size information, so as to realize accurate detection of smoke, that is, it can be more accurately determined whether it is smoke.

[0053] Optionally, determining the center position information and size information of the target area in at least three target images may include steps A1-A3:

[0054] Step A1: performing grayscale transformation on a reference image and at least three target images.

[0055] Among them, grayscale transformation refers to a method of changing the grayscale value of each pixel in the source image point by point according to a certain target condition and a certain transformation relationship. The purpose is to improve the image quality and make the image display effect clearer. Image grayscale transformation processing is a very basic and direct spatial domain image processing method in image enhancement processing technology, and is also an important part of image digitization software and image display software.

[0056] For example, grayscale transformation is performed on the target images of 1 frame, 2 frames, 3 frames, 4 frames and 5 frames determined in the example of S110.

[0057] Step A2: for each target image in the at least three target images, determine a target area in the target image by performing grayscale subtraction between the target image after grayscale transformation and the reference image after grayscale transformation.

[0058] For example, the reference image of frame 0 in the S110 example is gray-scale transformed. Since the scenes in the images of frames 0 to 5 are the same, only the smoke is dynamically changing. Therefore, the gray-scale transformed target image is gray-scale subtracted from the gray-scale transformed reference image to obtain the transformed target images of frames 1, 2, 3, 4, and 5 (e.g., Figure 1c), the target area in the target image is determined by changing the prominent features in the target image.

[0059] Step A3: Determine the center position information and size information of the target area in the target image.

[0060] Specifically, a target bounding rectangle of a target area in a target image is determined, the target bounding rectangle being the maximum bounding rectangle of the target area, and then the center position information of the target bounding rectangle is determined as the center position information of the target area and the size information of the target bounding rectangle is determined as the size information of the target area.

[0061] For example, taking the 4th frame as an example, the target area in the target image is determined by grayscale subtraction of the target image (4th frame) after grayscale transformation and the reference image (0th frame) after grayscale transformation, and then the target circumscribed rectangle of the target area in the target image is determined (such as Figure 1d As shown), the coordinate information of the center point of the target circumscribed rectangle is further used as the center position information of the target area, and the height and width of the target circumscribed rectangle are used as the size information of the target area.

[0062] The present technical solution realizes accurate determination of the suspected smoke area by performing grayscale transformation on a reference image and at least three target images, and then determining the target area in the target image by performing grayscale subtraction between the target image after grayscale transformation and the reference image after grayscale transformation, thereby further determining the center position information and size information of the target area in the target image, so as to facilitate subsequent quantitative and accurate detection of smoke.

[0063] S130. Determine whether the target area triggers a first preset condition according to central position information of the target area in at least three target images.

[0064] The first preset condition is used to determine whether the changing trend of the center position of the target area over time matches the changing trend of the center position of the actual smoke area.

[0065] Specifically, the center position information of the target area in each target image is obtained, and then the center position information of the target area in the target image is mapped to the same coordinate system in order of the frame numbers of the target image, from small to large, to obtain a change graph of the center position information, and then the change trend of the smoke over time is determined based on the change graph of the center position information, and then compared with the change trend of the actual smoke. If they meet the requirements, the first preset condition is triggered, that is, it may be smoke. Otherwise, the first preset condition is not triggered, that is, it is not smoke.

[0066] In a feasible embodiment, optionally, determining whether the target area triggers the first preset condition based on the center position information of the target area in at least three target images may include steps A11-A12:

[0067] Step A11, determining a center position displacement map of the target area based on the center position information of the target area in at least three target images, wherein the center position displacement map of the target area is a displacement map characterizing the change of the center position of the target area formed by mapping the center positions of the target areas in each target image to images and connecting them.

[0068] Specifically, the center position information of the target area in at least three target images is mapped to an image of a preset size to obtain a center coordinate point map, and then each center position information in the center coordinate point map is connected to form a center position displacement map that represents the change in the center position of the target area.

[0069] Optionally, the center of the smoke area moves at a nearly uniform speed in an indoor windless environment, and the moving direction is fixed, and usually there is no sudden displacement phenomenon. This phenomenon can be used to filter out some "dynamic" non-smoke targets, such as pedestrians. Therefore, before determining the center position displacement map of the target area based on the center position information of the target area in at least three target images, the following steps A111-A114 need to be performed:

[0070] A111. Determine, based on the center position information of the target area in at least three target images, at least two center position vectors formed from the center position of the target area in the third image to the center positions of the target areas in at least two fourth images, respectively, where the third image is the first frame image in the at least three target images, and the at least two fourth images are the remaining images in the at least three target images except the third image;

[0071] A112: if the minimum angle formed by at least two center position vectors is greater than a first preset angle value, it is determined that the target area does not trigger the first preset condition;

[0072] A113, if the minimum angle formed by at least two center position vectors is between the second preset angle value and the first preset angle, determining a center position displacement map of the target area based on the center position information of the target area in at least three target images;

[0073] A114. If the minimum angle formed by at least two center position vectors is smaller than a second preset angle, it is determined that the target area triggers a first preset condition, and the first preset angle is greater than the second preset angle.

[0074] This technical solution, before determining the center position displacement map of the target area based on the center position information of the target area in at least three target images, excludes non-smoke targets with "dynamicity" by judging the minimum angle formed by at least two center position vectors and the first preset angle value and the second preset angle value, thereby achieving more accurate detection of smoke.

[0075] Step A12: performing classification matching according to the center position displacement map of the target area to determine whether the target area triggers the first preset condition.

[0076] Specifically, the present application can classify and match the center position displacement map of the target area through various neural networks to determine whether the center position change trend of the target area over time matches the center position change trend of the actual smoke area. It can also be assigned and matched through a first preset classifier, and the first preset classifier is used to classify and determine whether the center position change trend of the target area over time matches the center position change trend of the actual smoke area.

[0077] Taking the allocation and matching of the first preset classifier as an example, specifically: the center position displacement map of the target area is input into the first preset classifier, and the first preset classifier outputs whether the target area triggers the first preset condition.

[0078] The technical solution of the present application determines the center position displacement map of the target area based on the center position information of the target area in at least three target images, so as to avoid the inaccuracy of the determined center position displacement map due to too little image information, thereby facilitating classification and matching based on the center position displacement map of the target area to determine whether the target area triggers the first preset condition, that is, to accurately judge the "dynamics" of the smoke.

[0079] S140: Determine whether the target area triggers a second preset condition according to size information of the target area in at least three target images.

[0080] The second preset condition is used to determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area.

[0081] Specifically, the size information of the target area in each target image is obtained, and the area map of the target area in each target image is determined based on the size information, and the changing trend of the area map is compared with the changing trend of the actual smoke area to accurately determine whether there is smoke in the target image. If the changing trend of the area map matches the changing trend of the actual smoke area, the second preset condition is triggered, indicating that it may be smoke. Otherwise, the second preset condition is not triggered, that is, it is not smoke.

[0082] S150: If the target area triggers the first preset condition and the second preset condition at the same time, a smoke alarm event is triggered for the area to be detected.

[0083] Specifically, if only one of the first preset condition and the second preset condition is triggered, it is considered that the target area is not filled with smoke at this time and no early warning is needed. If the target area triggers both the first preset condition and the second preset condition at the same time, it means that smoke has been generated indoors at this time, and an early warning is required immediately, that is, a smoke alarm event is triggered in the detection area to remind personnel to deal with it in time to avoid greater disasters.

[0084] The technical solution of the embodiment of the present invention determines a reference image and at least three target images corresponding to the area to be detected, and determines the target area as the suspected smoke area according to the image difference between the target image and the reference image, so as to facilitate the subsequent determination of the change trend of the center position information and size information of the target area in at least three target images, and match the change trend of the center position information and size information of the actual smoke area to determine whether the target area triggers the first preset condition and the second preset condition, thereby realizing accurate detection of smoke, and when the first preset condition and the second preset condition are triggered at the same time in the target area, a smoke alarm event is triggered for the area to be detected, that is, at this time, the target area is accurately determined to be smoke, thereby improving the smoke detection accuracy of indoor scenes, and improving the recall rate of smoke targets through quantitative judgment, thereby solving the problem that smoke cannot be effectively identified in the early stage of a fire.

[0085] Figure 2a 4 is a flow chart of a smoke detection method provided by an embodiment of the present invention. This embodiment is a detailed description of the above embodiment S130. Figure 2a As shown, the method includes:

[0086] S210, determining a reference image and at least three target images corresponding to the area to be detected, and determining central position information and size information of the target area in the at least three target images.

[0087] S220: Determine, based on the center position information of the target area in at least three target images, the center position measurement values ​​of the target area in at least two first images and the center position measurement value of the target area in the second image.

[0088] The second image is the last frame of at least three target images, and the at least two first images are the remaining images of at least three target images except the second image. The center position measurement value is the center coordinate value of the target area and the predicted distance information to the next frame.

[0089] For example, the center position measurement values ​​of frames 1, 2, 3, 4, and 5 are determined, where frames 1 to 4 are the first image and frame 5 is the second image.

[0090] S230, based on the center position measurement values ​​of the target area in at least two first images, simulate and predict the center position prediction value of the target area in the second image according to the position change trend of the actual smoke area.

[0091] Specifically, according to the order of image frames of the first image and based on the principle of Kalman filtering, the center position measurement value of the target area in the first frame of the first image is used to predict the center position prediction value of the next frame, and then the center position prediction value of the next frame is corrected according to the center position measurement value of the next frame, and the center position prediction value of the next frame is continued to be determined using the corrected center position prediction value, and it is repeated to accurately determine the center position prediction value of the target area in the second image.

[0092] For example, Figure 4 As shown, according to the principle of Kalman filtering, the center position measurement value of the target area of ​​1 frame image is predicted by the center position prediction value of 2 frames image, and then the center position measurement value of 2 frames image is used to correct the center position prediction value of 2 frames image, and the corrected center position prediction value is used to continue to determine the center position prediction value of 3 frames image, and it is repeatedly iterated to accurately determine the center position prediction value of the target area in 5 frames image.

[0093] S240: If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is less than a preset difference, determine that the target area triggers a first preset condition.

[0094] Specifically, the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image may be the distance between the predicted value of the center position and the measured value of the center position, which may be calculated according to the distance formula between two points and may be expressed as:

[0095]

[0096] Among them, (x′ center5 , y′ center5 ) is the predicted value of the center position of the target area in the second image, (x center5 ,y center5 ) is the measured value of the center position of the target area in the second image.

[0097] If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is less than the preset difference, it means that the target in the target area may be smoke, and it is determined that the target area triggers the first preset condition; otherwise, it means that the target in the target area is not smoke, and the target area does not trigger the first preset condition.

[0098] If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is not less than the preset difference, it is necessary to determine whether the target area triggers the first preset condition by other methods, including steps B1-B2:

[0099] Step B1, determining a center position displacement map of the target area based on the center position information of the target area in at least three target images, wherein the center position displacement map of the target area is a displacement map characterizing the change of the center position of the target area formed by mapping the center positions of the target areas in each target image to images and connecting them.

[0100] Specifically, the center position information of the target area in at least three target images is mapped to an image of a preset size to obtain a center coordinate point map, and then each center position information in the center coordinate point map is connected to form a center position displacement map representing the change of the center position of the target area. For example, Figure 4 As shown, at least three target images are 1 frame, 2 frames, 3 frames, 4 frames and 5 frames, and then the 1 frame to 5 frame images are mapped to an image of a preset size (which can be 64*64) to obtain a center coordinate point map (such as Figure 2b ), then Figure 2b The points on the center are connected to obtain the center position displacement map (such as Figure 2c ).

[0101] In a feasible embodiment, optionally, the center of the smoke area moves at a nearly uniform speed in an indoor windless environment, and the moving direction is fixed, and usually there is no sudden displacement phenomenon. Through this phenomenon, some "dynamic" non-smoke targets, such as pedestrians, can be filtered out. Therefore, before determining the center position displacement map of the target area based on the center position information of the target area in at least three target images, the following steps need to be performed, including steps C1-C4:

[0102] Step C1, based on the center position information of the target area in at least three target images, determine at least two center position vectors formed from the center position of the target area in the third image to the center position of the target area in at least two fourth images, where the third image is the first frame image of the at least three target images, and the at least two fourth images are the remaining images of the at least three target images except the third image.

[0103] For example, at least three target images are 1 frame, 2 frames, 3 frames, 4 frames and 5 frames, and 1 frame image is used as the third image. Figure 2bThe center coordinate point diagram is obtained. Then, draw straight lines from the center position of the 1st frame image to the center positions of the 2nd, 3rd, 4th and 5th frames of images to form 4 center position vectors. Then, use the vector angle formula to calculate the angle between the 4 center position vectors. The vector angle formula can be expressed as:

[0104]

[0105] Step C2: If the minimum angle formed by at least two center position vectors is greater than a first preset angle value, it is determined that the target area does not trigger the first preset condition.

[0106] Specifically, the first preset angle is set according to the actual situation, and is used to determine the maximum value that the minimum angle formed by the center position vector can reach. When it exceeds the maximum value, it means that the target in the target area is not smoke (such as Figure 2e The vectors shown are generated by the center coordinate point diagram of non-smoke targets. The angle between the vectors is too large), that is, when the minimum angle formed by at least two center position vectors is greater than the first preset angle value, it means that all points fluctuate greatly and suddenly move. It is determined that the dynamics of the target in the target area is not caused by smoke, that is, it is determined that the target area has not triggered the first preset condition.

[0107] Step C3: if the minimum angle formed by at least two center position vectors is between the second preset angle value and the first preset angle, determine the center position displacement map of the target area based on the center position information of the target area in at least three target images.

[0108] Specifically, the minimum angle formed by at least two center position vectors is between the second preset angle value and the first preset angle, which means that it is impossible to determine whether the target area triggers the first preset condition by the angle at this time. Therefore, it is necessary to determine the center position displacement map of the target area based on the center position information of the target area in at least three target images, so as to subsequently determine whether the target area triggers the first preset condition based on the center position displacement map of the target area.

[0109] Step C4: if the minimum angle formed by at least two center position vectors is smaller than the second preset angle, it is determined that the target area triggers the first preset condition, and the first preset angle is greater than the second preset angle.

[0110] Specifically, the second preset angle is set according to the actual situation, and when the minimum angle formed by at least two center position vectors is less than the second preset angle, it means that all points do not fluctuate significantly at this time, and no sudden displacement occurs, and it is determined that the dynamics of the target in the target area is caused by smoke (such as Figure 2d The vectors generated by the center coordinate point diagram of the smoke target are shown, and the angles between the vectors are relatively small), that is, the target area is determined to trigger the first preset condition.

[0111] This technical solution, before determining the center position displacement map of the target area based on the center position information of the target area in at least three target images, excludes non-smoke targets with "dynamicity" by judging the minimum angle formed by at least two center position vectors and the first preset angle value and the second preset angle value, thereby achieving more accurate detection of smoke.

[0112] Step B2: performing classification matching according to the center position displacement map of the target area to determine whether the target area triggers the first preset condition.

[0113] Specifically, the present application can classify and match the center position displacement map of the target area through various neural networks to determine whether the center position change trend of the target area over time matches the center position change trend of the actual smoke area. It can also be assigned and matched through a first preset classifier, and the first preset classifier is used to classify and determine whether the center position change trend of the target area over time matches the center position change trend of the actual smoke area.

[0114] For example, Figure 4 As shown, classification matching is performed through a first preset classifier, the center position displacement map of the target area is input into the first preset classifier, and the first preset classifier outputs whether the target area triggers the first preset condition.

[0115] The main structure of the first preset classifier in this case adopts the ResNet18 network, and the training data of the first preset classifier adopts the center coordinate displacement map generated by the actual smoke video, such as Figure 2f Shown is an example diagram of the first preset classifier training data.

[0116] Optionally, in order to more accurately determine whether the target area triggers the first preset condition when the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is not less than a preset difference, it is also possible to combine the minimum angle θ formed by at least two center position vectors and the classification confidence c of the first preset classifier to determine whether the target area triggers the first preset condition, specifically:

[0117] The minimum angle θ formed by at least two center position vectors and the classification confidence c of the first preset classifier are input into the judgment formula to obtain a judgment value. When the judgment value is greater than the preset judgment value (the preset judgment value is set according to the actual situation, for example, it can be selected in the range of [0.7,1]), it is considered whether the target area triggers the first preset condition.

[0118] Among them, the judgment formula can be expressed as:

[0119] z=β1*cosθ+β2*c;

[0120] Among them, the value range of β1 and β2 is between (0,1), and the specific value is selected according to the actual situation.

[0121] S250: Determine whether the target area triggers a second preset condition according to size information of the target area in at least three target images.

[0122] S260: If the target area triggers the first preset condition and the second preset condition at the same time, a smoke alarm event is triggered for the area to be detected.

[0123] The technical solution of the embodiment of the present invention determines a reference image and at least three target images corresponding to the area to be detected, and determines the target area as the suspected smoke area according to the image difference between the target image and the reference image, and simultaneously determines the center position information and size information of the target area in the at least three target images, and then determines the center position measurement values ​​of the target area in at least two first images and the center position measurement values ​​of the target area in the second image according to the center position information of the target area in the at least three target images, and further, according to the center position measurement values ​​of the target area in the at least two first images, simulates and predicts the center position prediction value of the target area in the second image according to the position change trend of the actual smoke area, so as to facilitate the prediction of the center position of the target area in the second image according to the center position measurement value of the target area. The relationship between the measured value and the measured value of the center position of the target area in the second image can accurately determine whether the target area triggers the first preset condition, that is, accurately judge the "dynamicity" of the smoke. At the same time, by determining the changing trend of the size information of the target area in at least three target images, the changing trend of the size information of the actual smoke area is matched to determine whether the target area triggers the second preset condition, thereby realizing accurate detection of smoke. If the first preset condition and the second preset condition are triggered in the target area at the same time, a smoke alarm event is triggered for the detection area, that is, the target area is accurately determined to be smoke at this time, which improves the smoke detection accuracy of indoor scenes, and improves the recall rate of smoke targets through quantitative judgment, thereby solving the problem of ineffective identification of smoke in the early stage of a fire.

[0124] Figure 3a 4 is a flowchart of a smoke detection method provided by an embodiment of the present invention. This embodiment is a detailed description of the above embodiment S140. Figure 3a As shown, the method includes:

[0125] S310, determining a reference image and at least three target images corresponding to the area to be detected, and determining central position information and size information of the target area in the at least three target images.

[0126] S320: Determine, based on the center position information of the target area in at least three target images, the center position measurement value of the target area in at least two first images and the center position measurement value of the target area in the second image.

[0127] S330: Determine whether the target area triggers a first preset condition according to central position information of the target area in at least three target images.

[0128] S340: Determine a size change map of the target area based on size information of the target area in at least three target images.

[0129] The size change map of the target area is a size change map that is formed by mapping the size of the target area in each target image to an image and represents the size change of the target area.

[0130] Specifically, the judgment of whether the size change trend of the target area over time matches the size change trend of the actual smoke area is the judgment of the "diffusivity" of the target area. Through observation, it is known that the area occupied by indoor smoke in the video will expand rapidly in a short period of time. From the time series, the area of ​​the smoke target tends to increase. It is this feature that distinguishes it from other objects. This feature can be used to judge whether the target in the target area is smoke, and some moving non-smoke targets can also be filtered out.

[0131] Therefore, the size information of the target area in at least three target images is used to determine the area value of the target area in at least three target images, and then the area value is sequentially mapped to images of a preset size according to the number of frames to obtain a size change map of the target area. Figure 4 As shown, at least three target images are 1 frame, 2 frames, 3 frames, 4 frames and 5 frames. The target area values ​​of these 5 frames are calculated respectively, and mapped to an image of a preset size (the preset size can be 64*64) in a height manner. Each vertical line in the image represents the relative area of ​​a frame target area, thereby obtaining a size change diagram (such as Figure 3b ).

[0132] S350 , performing classification matching according to the size change graph of the target area to determine whether the target area triggers a second preset condition.

[0133] This application quantifies the "diffusion" of the target area mainly to determine whether the target is in a state of area expansion. Through observation of the actual smoke diffusion process and the generated size change graph, it is found that it is not possible to simply judge whether the target is "diffusion" by the area of ​​the next frame being larger than the area of ​​the previous frame. Figure 3cAs shown in the figure, all of them are size change diagrams generated by the operations mentioned in this case for normal smoke. It can be seen from the figure that not every frame has an area larger than the previous frame. On the one hand, this is due to the uncertainty of the smoke itself, and on the other hand, it is due to the fact that the image processing method proposed in this case failed to successfully obtain part of the smoke area during the grayscale transformation and image subtraction operations. However, it can be seen that the area occupied by the smoke is expanding. In order to measure the trend of "diffusion" of smoke in continuous frames, this case proposes to use a neural network to train an image classifier, and its main structure adopts the ResNet18 network. The positive samples of the training data select smoke with obvious features, clear edges, and obvious diffusion; the negative samples select other moving objects but fixed shapes, such as vehicles, pedestrians, etc. The size change diagram is generated by the training data and sent to the image classifier for training to obtain the second preset classifier of this application.

[0134] Specifically, the present application can classify and match the center position displacement map of the target area through various neural networks to classify and determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area. It can also perform classification and matching through a second preset classifier, and the second preset classifier is used to classify and determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area.

[0135] For example, Figure 4 As shown, the second preset classifier is used for allocation matching, specifically: the size change graph of the target area is input into the second preset classifier, and the second preset classifier outputs whether the target area triggers the second preset condition. The size change graph is sent to the trained second preset classifier for judgment, and finally the judgment result is output. The advantage of using this method is that the neural network can learn the characteristic that the area is increasing through data training, which is more robust than simply using mathematical calculations and logical judgments.

[0136] S360: If the target area triggers the first preset condition and the second preset condition at the same time, a smoke alarm event is triggered for the area to be detected.

[0137] The technical solution of the embodiment of the present invention is to determine a reference image and at least three target images corresponding to the area to be detected, and determine the target area as the suspected smoke area according to the image difference between the target image and the reference image, and at the same time determine the center position information and size information of the target area in at least three target images, and then match the change trend of the center position information of the target area in at least three target images with the change trend of the center position information of the actual smoke area to determine whether the target area triggers the first preset condition, thereby achieving accurate detection of smoke. At the same time, based on the size information of the target area in at least three target images, a size change graph of the target area is determined, and according to the size change graph of the target area, it is determined whether the target area triggers the second preset condition, that is, the "diffusivity" of the smoke is accurately judged, and the first preset condition and the second preset condition are triggered in the target area at the same time, then a smoke alarm event is triggered for the area to be detected, that is, at this time, the target area is accurately determined to be smoke, thereby improving the smoke detection accuracy of indoor scenes, and improving the recall rate of smoke targets through quantitative judgment, thereby solving the problem that smoke cannot be effectively identified in the early stage of a fire.

[0138] Figure 5 This is a schematic diagram of the structure of a smoke detection device provided by an embodiment of the present invention. This embodiment is applicable to the case of detecting smoke generated in indoor scenes (especially smoke that has just been generated). The smoke detection device can be implemented in the form of hardware and / or software. The smoke detection device can be configured in any electronic device with network communication function. Figure 5 As shown, the device comprises:

[0139] A detection module 410 is used to determine a reference image and at least three target images corresponding to the area to be detected, wherein the reference image is a first frame image in a plurality of frames of images continuously acquired at a preset interval length in the area to be detected, and the at least three target images are the remaining images in the plurality of frames of images continuously acquired except the first frame image;

[0140] An information determination module 420, configured to determine central position information and size information of a target area in the at least three target images, wherein the target area is a suspected smoke area determined by an image difference between the target image and a reference image;

[0141] A first judgment module 430, used to determine whether the target area triggers a first preset condition based on the center position information of the target area in the at least three target images, wherein the first preset condition is used to determine whether the change trend of the center position of the target area over time matches the change trend of the center position of the actual smoke area;

[0142] A second judgment module 440 is used to determine whether the target area triggers a second preset condition based on the size information of the target area in the at least three target images, and the second preset condition is used to determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area;

[0143] The early warning module 450 is used to trigger a smoke alarm event for the target area if the target area triggers the first preset condition and the second preset condition at the same time.

[0144] Optional, detection module for:

[0145] The detection area is continuously captured at a preset interval to obtain multiple frames of images, wherein the detection area is an indoor area with a light intensity less than a preset light intensity;

[0146] The first frame image among the continuously acquired multiple frame images is determined as a reference image, and the remaining at least three frame images except the first frame image among the continuously acquired multiple frame images are determined as at least three target images.

[0147] Optionally, the information determination module includes:

[0148] A grayscale conversion unit, used for performing grayscale conversion on the reference image and the at least three target images;

[0149] A target region determining unit, configured to determine a target region in each of the at least three target images by performing grayscale subtraction between the target image after grayscale transformation and the reference image after grayscale transformation;

[0150] The target information determination unit is used to determine the center position information and size information of the target area in the target image.

[0151] Optionally, the target information determination unit is used to:

[0152] Determine a target bounding rectangle of a target area in the target image, wherein the target bounding rectangle is a maximum bounding rectangle of the target area;

[0153] The center position information of the target circumscribed rectangle is determined as the center position information of the target area, and the size information of the target circumscribed rectangle is determined as the size information of the target area.

[0154] Optionally, the first judgment module includes a first judgment unit, configured to:

[0155] Determining whether the target area triggers a first preset condition according to the center position information of the target area in the at least three target images includes:

[0156] Determine, according to the center position information of the target area in the at least three target images, a center position measurement value of the target area in at least two first images and a center position measurement value of the target area in a second image, wherein the second image is a last frame image of the at least three target images, and the at least two first images are the remaining images of the at least three target images except the second image;

[0157] Based on the center position measurement values ​​of the target area in at least two first images, the center position prediction value of the target area in the second image is simulated and predicted according to the position change trend of the actual smoke area;

[0158] If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is less than a preset difference, it is determined that the target area triggers the first preset condition.

[0159] Optionally, the first judging unit further includes:

[0160] A first displacement map determining unit is configured to determine a displacement map of a center position of the target area based on the center position information of the target area in the at least three target images if the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is not less than a preset difference, wherein the displacement map of the center position of the target area is a displacement map representing a change in the center position of the target area formed by mapping the center positions of the target areas in the respective target images to the images and connecting them;

[0161] The first trigger unit is used to determine whether the target area triggers a first preset condition according to the center position displacement map of the target area.

[0162] The first judgment module includes a second judgment unit, including:

[0163] A second displacement map determining unit is used to determine a center position displacement map of the target area based on the center position information of the target area in the at least three target images, wherein the center position displacement map of the target area is a displacement map representing the change of the center position of the target area formed by mapping the center positions of the target areas in the respective target images to images and connecting them;

[0164] The second trigger unit is used to perform classification matching according to the center position displacement map of the target area to determine whether the target area triggers the first preset condition.

[0165] Optionally, the first trigger unit or the second trigger unit is used to:

[0166] The center position displacement map of the target area is input into a first preset classifier, and the first preset classifier outputs whether the target area triggers a first preset condition. The first preset classifier is used to classify and determine whether the center position change trend of the target area over time matches the center position change trend of the actual smoke area.

[0167] Optionally, the first displacement map determining unit or the first displacement map determining unit further includes a third triggering unit, configured to:

[0168] Determine, based on the center position information of the target area in the at least three target images, at least two center position vectors formed from the center position of the target area in the third image to the center positions of the target areas in at least two fourth images, respectively, wherein the third image is a first frame image in the at least three target images, and the at least two fourth images are the remaining images in the at least three target images except the third image;

[0169] If the minimum angle formed by the at least two center position vectors is greater than a first preset angle value, it is determined that the target area does not trigger the first preset condition;

[0170] If the minimum angle formed by the at least two center position vectors is between the second preset angle value and the first preset angle, determining a center position displacement map of the target area based on the center position information of the target area in the at least three target images;

[0171] If the minimum angle formed by the at least two center position vectors is smaller than a second preset angle, it is determined that the target area triggers a first preset condition, and the first preset angle is greater than the second preset angle.

[0172] Optionally, the second judgment module includes:

[0173] A size change map determining unit, configured to determine a size change map of a target area based on size information of the target area in the at least three target images, wherein the size change map of the target area is a size change map representing a size change of the target area formed by mapping the size of the target area in each target image to an image;

[0174] The fourth triggering unit is used to determine whether the target area triggers a second preset condition according to the size change diagram of the target area.

[0175] Optionally, a fourth trigger unit is used to:

[0176] The size change graph of the target area is input into a second preset classifier, and the second preset classifier outputs whether the target area triggers a second preset condition. The second preset classifier is used to classify and determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area.

[0177] The smoke detection device provided in the embodiment of the present invention can execute the smoke detection method provided in any embodiment of the present invention, and has the corresponding functions and beneficial effects of executing the smoke detection method. For detailed process, please refer to the relevant operations of the smoke detection method in the above embodiment.

[0178] Figure 6 A schematic diagram of the structure of an electronic device that can be used to implement the smoke detection method of an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0179] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0180] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0181] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a smoke detection method.

[0182] In some embodiments, the smoke detection method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the smoke detection method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the smoke detection method in any other appropriate manner (e.g., by means of firmware).

[0183] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0184] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0185] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0186] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0187] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0188] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0189] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0190] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A smoke detection method, characterized in that: include: Determine a reference image and at least three target images corresponding to the area to be detected; Determine the center position information and size information of the target area in the at least three target images, the target area being a suspected smoke area determined by the image difference between the target image and the reference image; Determining whether the target area triggers a first preset condition based on the center position information of the target area in the at least three target images, wherein the first preset condition is used to determine whether a change trend of the center position of the target area over time matches a change trend of the center position of an actual smoke area; Determining whether the target area triggers a second preset condition based on the size information of the target area in the at least three target images, wherein the second preset condition is used to determine whether a size change trend of the target area over time matches a size change trend of an actual smoke area; If the target area triggers the first preset condition and the second preset condition at the same time, a smoke alarm event is triggered for the area to be detected.

2. The method according to claim 1, characterized in that Determining the center position information and size information of the target area in the at least three target images includes: Performing grayscale transformation on the reference image and the at least three target images; For each of the at least three target images, determining a target area in the target image by performing grayscale subtraction between the target image after grayscale transformation and the reference image after grayscale transformation; Determine the center position information and size information of the target area in the target image.

3. The method according to claim 1, characterized in that Determining whether the target area triggers a first preset condition according to the center position information of the target area in the at least three target images includes: Determine, according to the center position information of the target area in the at least three target images, a center position measurement value of the target area in at least two first images and a center position measurement value of the target area in a second image, wherein the second image is a last frame image of the at least three target images, and the at least two first images are the remaining images of the at least three target images except the second image; Based on the center position measurement values ​​of the target area in at least two first images, the center position prediction value of the target area in the second image is simulated and predicted according to the position change trend of the actual smoke area; If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is less than a preset difference, it is determined that the target area triggers the first preset condition.

4. The method according to claim 3, characterized in that The method further comprises: If the difference between the predicted value of the center position of the target area in the second image and the measured value of the center position of the target area in the second image is not less than a preset difference, a center position displacement map of the target area is determined based on the center position information of the target area in the at least three target images, wherein the center position displacement map of the target area is a displacement map characterizing the change of the center position of the target area formed by mapping the center positions of the target areas in the respective target images to the images and connecting them; Classification matching is performed according to the center position displacement map of the target area to determine whether the target area triggers a first preset condition.

5. The method according to claim 1, characterized in that Determining whether the target area triggers a first preset condition according to the center position information of the target area in the at least three target images includes: Determine a center position displacement map of the target area based on the center position information of the target area in the at least three target images, wherein the center position displacement map of the target area is a displacement map representing the change of the center position of the target area formed by mapping the center positions of the target areas in the respective target images to the images and connecting them; Classification matching is performed according to the center position displacement map of the target area to determine whether the target area triggers a first preset condition.

6. The method according to claim 4 or 5, characterized in that: Before determining the center position displacement map of the target area based on the center position information of the target area in the at least three target images, the method further includes: Determine, based on the center position information of the target area in the at least three target images, at least two center position vectors formed from the center position of the target area in the third image to the center positions of the target areas in at least two fourth images, respectively, wherein the third image is a first frame image in the at least three target images, and the at least two fourth images are the remaining images in the at least three target images except the third image; If the minimum angle formed by the at least two center position vectors is greater than a first preset angle value, it is determined that the target area does not trigger the first preset condition; If the minimum angle formed by the at least two center position vectors is between the second preset angle value and the first preset angle, determining a center position displacement map of the target area based on the center position information of the target area in the at least three target images; If the minimum angle formed by the at least two center position vectors is smaller than a second preset angle, it is determined that the target area triggers a first preset condition, and the first preset angle is greater than the second preset angle.

7. The method according to claim 1, characterized in that Determining whether the target area triggers a second preset condition according to size information of the target area in the at least three target images includes: Determine a size change map of the target area based on the size information of the target area in the at least three target images, wherein the size change map of the target area is a size change map representing the size change of the target area formed by mapping the size of the target area in each target image to an image; Classification matching is performed according to the size change graph of the target area to determine whether the target area triggers a second preset condition.

8. A smoke detection device, characterized in that: include: A detection module, used to determine a reference image and at least three target images corresponding to the area to be detected; An information determination module, used to determine the center position information and size information of a target area in the at least three target images, wherein the target area is a suspected smoke area determined by the image difference between the target image and the reference image; A first judgment module, used to determine whether the target area triggers a first preset condition based on the center position information of the target area in the at least three target images, wherein the first preset condition is used to determine whether the change trend of the center position of the target area over time matches the change trend of the center position of the actual smoke area; A second judgment module is used to determine whether the target area triggers a second preset condition based on the size information of the target area in the at least three target images, and the second preset condition is used to determine whether the size change trend of the target area over time matches the size change trend of the actual smoke area; The early warning module is used to trigger a smoke alarm event for the target area if the target area triggers the first preset condition and the second preset condition at the same time.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the smoke detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the smoke detection method according to any one of claims 1 to 7 when executed.

Images