A data processing system for monitoring the safety status of users in a building area

Abstract

The present invention relates to the field of data processing, and in particular to a data processing system for monitoring the safety status of users in a building area. When a computer program is executed by a processor, the following steps are implemented: sampling according to a preset sampling interval to obtain a sampled image data set, and combining a target building area information set to obtain a user safety information set, a first user safety probability set, and a user safety status corresponding to each target building area, and monitoring the user safety status by screening out s sampled images from each target video data, thereby reducing the amount of calculation and improving the monitoring efficiency, and extracting a first image area corresponding to a non-user and a second image area corresponding to a user, as well as the distribution of non-user pixel points and user pixel points in each image area, and combining the weight corresponding to each pixel point to obtain the center of gravity corresponding to each area, so as to characterize the safety status corresponding to the non-user and the user as a whole, thereby improving the accuracy of the monitoring result of the user safety status.

Description

Technical Field

[0001] The invention relates to the field of data processing, and in particular to a data processing system for monitoring the safety status of users in a building area. Background Art

[0002] Monitoring the safety status of users in the building area and taking corresponding safety measures based on the monitoring results can effectively reduce the probability of safety accidents. For example, when a user carries a heavy object on the stairs in the building area, the video acquisition device collects and analyzes the images of the heavy object and the user during the interaction process to determine the probability of the user falling on the stairs. The user's safety status on the stairs can be monitored, and the user's action behavior when carrying a heavy object can be guided based on the monitoring results, thereby reducing the probability of the user falling on the stairs and improving the safety of the user in the building area.

[0003] In the prior art, the images captured by the video acquisition device can be input into a trained model and the corresponding safety status monitoring results can be output. However, due to the complex interaction between the user and the heavy object and the user's posture, there is mutual occlusion between the heavy object and the user in the image captured by the video acquisition device, making it difficult to capture complete images of the heavy object and the user, resulting in low quality of the input image of the model, and further resulting in low accuracy of the output safety status monitoring results, which is difficult to meet the safety requirements in the building area.

[0004] Therefore, how to monitor the user's safety status based on the images collected by the video acquisition device and improve the accuracy of the monitoring results has become an urgent problem to be solved. Summary of the invention

[0005] In view of the above technical problems, the technical solution adopted by the present invention is a data processing system for monitoring the safety status of users in a building area, wherein the system comprises a monitoring device and a plurality of target video acquisition devices, wherein the monitoring device comprises a processor and a memory storing a computer program, and the memory also stores a target building area information set A={A1, A2, ..., A j , ..., A n} and target video dataset C = {C1, C2, ..., C j , ..., C n}, where A j ={A j1 , A j2 , ..., A jk , ..., A je(j)}, A jk is the inclination angle of the kth step in the jth target building area, C jIt refers to the target video data of the user and non-user interaction collected by the target video collection device corresponding to the j-th target building area, j=1, 2, ..., n, n is the number of target building areas, k=1, 2, ..., e(j), e(j) is the number of steps in the j-th target building area, when the computer program is executed by the processor, the following steps are implemented:

[0006] S100, sampling C according to a preset sampling interval T, and obtaining a sampled image data set D = {D1, D2, ..., D j , ..., D n}, where D j ={D j1 , D j2 , ..., D jr , ..., D js}, D jr It means according to T to C j The rth image obtained by sampling, r=1, 2, ..., s, s is the number of images sampled from each target video data.

[0007] S200, according to A, D, and the preset non-user weight W 1 and the preset user weight W 2 , obtain the user security information set U corresponding to D = {U1, U2, ..., U j , ..., U n}, where U j = {U j1 , U j2 , ..., U jr , ..., U js}, D jr Corresponding user security information U jr Obtain it by following the steps below:

[0008] S210, according to A and D jr , obtain the first image area information H 1 jr =(H 1 jr1 , H 1 jr2 , H 1 jr3 ) and the second image area information H 2 jr =(H 2 jr1 , H 2 jr2 , H 2 jr3 ), where H 1 jr1 Djr The user pixel point set in the first image area corresponding to the non-user, H 1 jr2 is the set of non-user pixels in the first image area, H 1 jr3 is the inclination angle of the first target step corresponding to the first image area, H 2 jr1 D jr The user pixel point set in the second image area corresponding to the user, H 2 jr2 is the set of non-user pixels in the second image area, H 2 jr3 is the inclination angle of the second target step corresponding to the second image area.

[0009] S220, according to H 1 jr2 , H 2 jr2 and W 1 , get D jr The first weight w corresponding to each non-user pixel in 1 =W 1 / L 1 1, where L 1 1 refers to H 1 jr2 The number of non-user pixels and H 2 jr2 The sum of the number of non-user pixels in .

[0010] S230, according to H 1 jr1 , H 2 jr1 and W 2 , get D jr The second weight w corresponding to each user pixel in 2 =W 2 / L 2 2, where L 2 2 refers to H 1 jr1 The number of user pixels and H 2 jr1 The sum of the number of user pixels in .

[0011] S240, according to H 1 jr1 , H 1 jr2 、w 1 and w 2 , get D jr The first centroid coordinates G of the first image area in1 jr .

[0012] S250, according to H 2 jr1 , H 2 jr2 、w 1 and w 2 , get D jr The second centroid coordinates G of the second image area 2 jr .

[0013] S260, according to G 1 jr , G 2 jr , H 1 jr3 and H 2 jr3 , get D jr Corresponding user security information U jr = {G 1 jr , G 2 jr , H 1 jr3 , H 2 jr3}.

[0014] S300, according to U, obtain the first user safety probability set F corresponding to D = {F1, F2, ..., F j , ..., F n}, where F j ={F j1 , F j2 , ..., F jr , ..., F js}, F jr It means D jr The corresponding first user safety probability refers to the probability that the user safety status of the corresponding user is safe.

[0015] S400: According to F, determine the user safety status corresponding to each target building area.

[0016] Compared with the prior art, the present invention has obvious beneficial effects. By means of the above technical scheme, the data processing system for monitoring the safety status of users in a building area provided by the present invention can achieve considerable technical advancement and practicality, and has wide industrial utilization value. It has at least the following beneficial effects: sampling C according to a preset sampling interval T, screening out s sampled images from each target video data as the basis for judging the user's safety status, reducing the amount of calculation in the user's safety status monitoring process, and improving the efficiency of the user's safety status monitoring, and then from D jr The first image area and the user pixel point set, the non-user pixel point set and the inclination angle of the first target step corresponding to the first image area are extracted to characterize the weight distribution in the first image area corresponding to the non-user, and the second image area and the user pixel point set, the non-user pixel point set and the inclination angle of the second target step corresponding to the second image area are extracted to characterize the weight distribution in the second image area corresponding to the user, thereby combining the weight corresponding to each pixel point and each pixel point in D jr Get the coordinates in D jr The first centroid coordinates and the second centroid coordinates in the image are obtained, and the first user safety probability corresponding to the non-user and the user as a whole in each sampled image is further obtained, so as to obtain the user safety status corresponding to each target building area, thereby improving the accuracy of the monitoring results of the user safety status. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. 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.

[0018] Figure 1 A flowchart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention.

[0019] Figure 2 Another flow chart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention.

[0020] Figure 3 Another flow chart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention.

[0021] Figure 4 Another flow chart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention.

[0022] Figure 5 Another flow chart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention.

[0023] Figure 6 Another flow chart of executing a computer program of a data processing system for monitoring the safety status of users in a building area provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] It should be noted that the terms "first", "second", etc. 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 interchangeable 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 are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] The first embodiment provides a data processing system for monitoring the safety status of users in a building area. The system includes a monitoring device and a plurality of target video acquisition devices. The monitoring device includes a processor and a memory storing a computer program. The memory also stores a target building area information set A={A1, A2, ..., A j , ..., A n} and target video dataset C = {C1, C2, ..., C j , ..., C n}, where A j ={A j1 , A j2 , ..., A jk , ..., A je(j)}, A jk is the inclination angle of the kth step in the jth target building area, C jIt refers to the target video data of the user and non-user interaction collected by the target video collection device corresponding to the j-th target building area, j=1, 2, ..., n, n is the number of target building areas, k=1, 2, ..., e(j), e(j) is the number of steps in the j-th target building area, when the computer program is executed by the processor, the following steps are implemented, such as Figure 1 As shown:

[0027] S100, sampling C according to a preset sampling interval T, and obtaining a sampled image data set D = {D1, D2, ..., D j , ..., D n}, where D j ={D j1 , D j2 , ..., D jr , ..., D js}, D jr It means according to T to C j The rth image obtained by sampling, r=1, 2, ..., s, s is the number of images sampled from each target video data.

[0028] The target building area may refer to the step area of ​​any floor in the target building that needs to monitor the user safety status, and the video acquisition device is a device that is set in the target building and performs video acquisition. When the target building area corresponds to only one video acquisition device that can acquire images in the target building area, the video acquisition device is determined as the target video acquisition device corresponding to the target building area; when multiple video acquisition devices can acquire images in the same target building area, one of the video acquisition devices is selected as the target video acquisition device corresponding to the target building area.

[0029] The target video acquisition device is used to acquire images in the corresponding target building area, and filter out video data in which users interact with non-users as corresponding target video data. The user may refer to a construction worker, the non-user may refer to a heavy object that needs to be carried by the construction worker from the bottom steps to the upper steps, and the interaction between the user and the non-user may refer to the behavior of the user carrying the non-user on the steps.

[0030] In order to obtain images of user interaction with non-users, this embodiment samples each target video data according to a preset sampling interval T. T can be set according to the average action step time of the user to obtain a sampled image corresponding to each action step of the user, and obtain a corresponding sampled image set as a basis for judging the user's safety status.

[0031] The target building area information includes the inclination angle of each step in the target building area, and the inclination angle may refer to the angle between the line between the lower left corner target point of the sampled image and the edge corner point of the corresponding step in the sampled image, and the bottom edge of the sampled image. The edge corner point may refer to the vertex in the upper rectangular surface of the step that does not touch other steps or walls.

[0032] In the above, C is sampled according to the preset sampling interval T, and s sampled images are selected from each target video data as the basis for judging the user's safety status, which reduces the amount of calculation in the user's safety status monitoring process and improves the efficiency of user safety status monitoring.

[0033] In a specific implementation manner, the memory also stores a sample user's action step time set t={t1, t2, ..., t g ,……,t z}, where t g It refers to the action step time of the gth sample user. The action step time refers to the time interval from the left foot being lifted to the left foot being put down, or from the right foot being lifted to the right foot being put down when the sample user walks on flat ground. g = 1, 2, ..., z, where z refers to the number of sample users. T meets the following conditions:

[0034] T=W0*∑ g＝1 z (t g / z), W0 refers to the preset step length weight.

[0035] In this embodiment, the average value of the action step duration of z sample users is obtained as the basis for obtaining the preset sampling interval T.

[0036] When the sample user walks on the stairs, the time interval from lifting the left foot to putting the left foot down, or from lifting the right foot to putting the right foot down, will be longer than when walking on flat ground. Therefore, this embodiment combines the preset step weight W0 to further obtain the preset sampling interval T, so as to sample the target video data according to T and obtain the sampled image corresponding to each step of the user.

[0037] As described above, by combining the preset step length weight W0 and the average of the action step times of the z sample users, the preset sampling interval T is obtained, which improves the accuracy of obtaining T and further improves the reliability of the sampled image in characterizing the user's walking state.

[0038] S200, according to A, D, and the preset non-user weight W 1 and the preset user weight W 2 , obtain the user security information set U corresponding to D = {U1, U2, ..., U j , ..., U n}, where U j = {U j1 , U j2 , ..., U jr , ..., U js}, D jr Corresponding user security information U jr Obtain it through the following steps, such as Figure 2 As shown:

[0039] S210, according to A and D jr , obtain the first image area information H 1 jr =(H 1 jr1 , H 1 jr2 , H 1 jr3 ) and the second image area information H 2 jr =(H 2 jr1 , H 2 jr2 , H 2 jr3 ), where H 1 jr1 D jr The user pixel point set in the first image area corresponding to the non-user, H 1 jr2 is the set of non-user pixels in the first image area, H 1 jr3 is the inclination angle of the first target step corresponding to the first image area, H 2 jr1 D jr The user pixel point set in the second image area corresponding to the user, H 2 jr2 is the set of non-user pixels in the second image area, H 2 jr3 is the inclination angle of the second target step corresponding to the second image area.

[0040] Among them, during the process of the user interacting with the non-user, if the center of gravity of the user and the non-user as a whole is high, the user may be at risk of falling, which may lead to a dangerous accident on the steps. Therefore, in order to monitor the safety status of users in the building area, this embodiment needs to analyze the center of gravity of the user and the non-user as a whole to obtain the detection result of the user's safety status, so as to guide the user's action behavior when interacting with the non-user according to the monitoring result, thereby reducing the probability of the user falling on the steps and improving the safety of the user in the building area.

[0041] In a specific implementation, S210 further includes the following steps: Figure 3 As shown:

[0042] S211, for D jr Perform semantic segmentation and transform D jr The pixels in are classified into user pixels, non-user pixels and background pixels.

[0043] S212, to D jr The bounding box detection is performed, and the image area corresponding to the first bounding box not corresponding to the user is determined as the first image area.

[0044] S213, according to D jr The user pixel point and the first image area in H 1 jr1 .

[0045] S214, according to D jr The non-user pixel points and the first image area in H 1 jr2 .

[0046] S215, based on the first bounding box and the e(j) steps in the j-th target building area, obtain D jr The corresponding first projection overlap length set Y jr 1 = {Y jr 11 , Y jr 12 , ..., Y jr 1k , ..., Y jr 1e(j)},Y jr 1k It means that the first bounding box is in D jr The first projection in the bottom direction of D jr The overlapping length between the second projections in the direction of the base edge.

[0047] S216, max(Y jr 1 ) is determined as the first target step corresponding to the first image area, and H is obtained according to the inclination angle of the first target step. 1 jr3 , where max() is the maximum value function.

[0048] S217, according to H 1 jr1 , H 1 jr2 and H1 jr3 , obtain the first image area information H 1 jr =(H 1 jr1 , H 1 jr2 , H 1 jr3 ).

[0049] Among them, D jr Input into the trained semantic segmentation model for semantic segmentation, D jr The pixels in the image are classified into three categories: user pixels, non-user pixels, or background pixels. User pixels refer to pixels corresponding to users, non-user pixels refer to pixels corresponding to non-users, and background pixels refer to pixels corresponding to D jr Other pixels except user pixels and non-user pixels. Among them, those skilled in the art know that any semantic segmentation model in the prior art falls within the protection scope of the present invention, and will not be described in detail here.

[0050] D jr Input into the trained bounding box detection model for bounding box detection, D jr The image region corresponding to the first bounding box corresponding to the non-user is determined as the first image region. Those skilled in the art know that any bounding box detection model in the prior art falls within the protection scope of the present invention, which will not be described in detail here.

[0051] Since the direction of gravity is vertically downward, in order to determine the target step corresponding to the majority of the weight of the non-user, the first bounding box is calculated in D jr The first projection in the bottom direction of D jr The overlap length between the second projection in the bottom direction and the max(Y jr 1 ) is determined as the first target step corresponding to the first image area.

[0052] The first image area is the image area corresponding to the first bounding box corresponding to the non-user, and the second image area is the image area corresponding to the second bounding box corresponding to the user. Due to the complex interaction between the user and the non-user and the user's posture, in D jr There is a situation where users and non-users block each other. Therefore, the first image area includes most non-user pixels and a small number of user pixels, and the second image area includes most user pixels and a small number of non-user pixels. Then, H can be obtained based on the user pixels and the first image area. 1 jr1, according to the non-user pixel points and the first image area, obtain H 1 jr2 .

[0053] As mentioned above, through D jr Semantic segmentation is performed to obtain user pixel points corresponding to the user and non-user pixel points corresponding to the non-user. Combined with the first image area corresponding to the first bounding box corresponding to the non-user, the user pixel point set and the non-user pixel point set in the first image area are obtained, the detection of the area where the non-user is located is completed, the distribution of user and non-user pixel points in the area where the non-user is located is clarified, and the pixel points of the user and the non-user are determined according to the first bounding box in D jr The first projection in the bottom direction of D jr The overlapping length between the second projections in the bottom edge direction is determined to determine the first target step corresponding to the first image area, providing a data basis for analyzing the center of gravity in the first image area.

[0054] In a specific implementation, S210 further includes the following steps: Figure 4 As shown:

[0055] S218, to D jr Performing bounding box detection, determining the image area corresponding to the second bounding box corresponding to the user as the second image area, and acquiring the third bounding box corresponding to the user's supporting leg;

[0056] S219, according to D jr The user pixel point and the second image area in H 2 jr1 ;

[0057] S2110, according to D jr The non-user pixels and the second image area in H 2 jr2 ;

[0058] S2111, based on the third bounding box and the e(j) steps in the j-th target building area, obtain D jr The corresponding second projection overlap length set Y jr 2 = {Y jr 21 , Y jr 22 , ..., Y jr 2k , ..., Y jr 2e(j)},Y jr 2k It means that the third bounding box is in D jrThe third projection in the direction of the bottom edge of the j-th target building area is the k-th step in D jr The overlapping length between the second projections in the bottom direction;

[0059] S2112, max(Y jr 2 ) is determined as the second target step corresponding to the second image area, and H is obtained according to the inclination angle of the second target step. 2 jr3 ;

[0060] S2113, according to H 2 jr1 , H 2 jr2 and H 2 jr3 , obtain the second image area information H 2 jr =(H 2 jr1 , H 2 jr2 , H 2 jr3 ).

[0061] Among them, D jr The input is input into the trained bounding box detection model for bounding box detection. The image area corresponding to the second bounding box corresponding to the user can be determined as the second image area, and the third bounding box corresponding to the user's supporting leg is obtained. Then, according to the third bounding box, the D jr The third projection in the direction of the bottom edge of the j-th target building area is the k-th step in D jr The overlapping length between the second projections in the bottom direction of the image is used to determine the step corresponding to most of the user's weight and use it as the second target step corresponding to the second image area.

[0062] In the above, based on the user pixel points corresponding to the user and the non-user pixel points corresponding to the non-user, in combination with the second image area corresponding to the second bounding box corresponding to the user, the user pixel point set and the non-user pixel point set in the second image area are obtained, the detection of the area where the user is located is completed, the distribution of the user and non-user pixel points in the area where the user is located is clarified, and the pixel points of the user and non-user in the area where the user is located are determined according to the third bounding box in D jr The third projection in the direction of the bottom edge of the j-th target building area is the k-th step in D jr The overlapping length between the second projections in the bottom edge direction is determined to determine the second target step corresponding to the second image area, providing a data basis for analyzing the center of gravity in the second image area.

[0063] S220, according to H 1jr2 , H 2 jr2 and W 1 , get D jr The first weight w corresponding to each non-user pixel in 1 =W 1 / L 1 1, where L 1 1 refers to H 1 jr2 The number of non-user pixels and H 2 jr2 The sum of the number of non-user pixels in .

[0064] In one embodiment, W 1 Obtain it by following the steps below:

[0065] S01, identifying a first image region, and obtaining a predicted category of non-user objects in the first image region;

[0066] S02, determining the preset object weight corresponding to the predicted object category as W 1 .

[0067] The predicted object category may refer to a first type of heavy object, a second type of heavy object, a third type of heavy object, etc., and the memory further stores a preset object weight corresponding to each object category. The preset object weight corresponding to the predicted object category of the non-user in the first image area may be determined as W. 1 .

[0068] In the above, by classifying the non-users in the first image area and combining the preset object weight corresponding to the predicted object category, W is obtained. 1 It is used to calculate the weight corresponding to each non-user pixel point, and then used to calculate the first center of gravity, thereby improving the calculation accuracy of the first center of gravity.

[0069] S230, according to H 1 jr1 , H 2 jr1 and W 2 , get D jr The second weight w corresponding to each user pixel in 2 =W 2 / L 2 2, where L 2 2 refers to H 1 jr1 The number of user pixels and H 2 jr1 The sum of the number of user pixels in .

[0070] In one embodiment, W 2 Obtain it by following the steps below:

[0071] S03, identifying the second image area, and obtaining the predicted height of the user in the second image area;

[0072] S04: Determine the preset user weight corresponding to the predicted height as W 2 .

[0073] The memory also stores a preset user weight corresponding to each user's height, and the preset user weight corresponding to the predicted height of the user in the second image area can be determined as W. 2 .

[0074] In the above, by predicting the height of the user in the second image area and combining the preset user weight corresponding to the predicted height, W is obtained. 2 It is used to calculate the weight corresponding to each user pixel point, and then used to calculate the second center of gravity, thereby improving the calculation accuracy of the second center of gravity.

[0075] S240, according to H 1 jr1 , H 1 jr2 、w 1 and w 2 , get D jr The first centroid coordinates G of the first image area in 1 jr .

[0076] In one embodiment, G 1 jr =(x 1 jr ,y 1 jr ), x 1 jr refers to the abscissa of the first centroid of the first image region, y 1 jr is the ordinate of the first centroid of the first image region, G 1 jr Meet the following conditions:

[0077] x 1 jr =(Σ α＝1 γ (w 2 *x 2 jr1 α )+Σ β＝1 δ (w 1 *x 1jr2 β )) / (γ*w 2 +δ*w 1 ), where x 2 jr1 α It means H 1 jr1 The horizontal coordinate of the αth user pixel point in x 1 jr2 β It means H 1 jr2 The horizontal coordinate of the βth non-user pixel in H, α = 1, 2, ..., γ, γ refers to 1 jr1 The number of user pixels in H, β = 1, 2, ..., δ, δ refers to 1 jr2 The number of non-user pixels in the

[0078] y 1 jr =(Σ α＝1 γ (w 2 *y 2 jr1 α )+Σ β＝1 δ (w 1 *y 1 jr2 β )) / (γ*w 2 +δ*w 1 ), where y 2 jr1 α It means H 1 jr1 The ordinate of the αth user pixel point in y 1 jr2 β It means H 1 jr2 The ordinate of the βth non-user pixel in .

[0079] Among them, in D jr In the example, the lower left corner is regarded as the origin, the left side is regarded as the y-axis, and the bottom side is regarded as the x-axis, then the coordinates of each pixel in D can be obtained. jr The coordinates in .

[0080] The first image region is regarded as a region with uneven quality. Then, according to the weight corresponding to each pixel in the first image region, the weight of each pixel in D jr Get the coordinates in D jr The first centroid coordinates G of the first image area in1 jr =(x 1 jr ,y 1 jr ). Among them, the weight corresponding to the background pixel is regarded as 0.

[0081] In the above, the first weight corresponding to each non-user pixel and the second weight corresponding to each user pixel are obtained according to the preset non-user weight and the preset user weight, and the weight of each pixel in D jr Get the coordinates in D jr The first centroid coordinates of the first image area in the image provide a data basis for analyzing the centroid situation of the user and the non-user as a whole.

[0082] S250, according to H 2 jr1 , H 2 jr2 、w 1 and w 2 , get D jr The second centroid coordinates G of the second image area 2 jr .

[0083] In one embodiment, G 2 jr =(x 2 jr ,y 2 jr ), x 2 jr refers to the abscissa of the second centroid of the second image area, y 2 jr is the ordinate of the second centroid of the second image area, G 2 jr Meet the following conditions:

[0084] x 2 jr =(Σ ζ＝1 λ (w 2 *x 2 jr1 ζ )+Σ η＝1 μ (w 1 *x 2 jr2 η )) / (λ*w 2 +μ*w 1 ), where x 2 jr1 ζ It means H2 jr1 The horizontal coordinate of the ζ-th user pixel point in 2 jr2 η It means H 2 jr2 The horizontal coordinate of the nth non-user pixel in H, ζ = 1, 2, ..., λ, λ refers to 2 jr1 The number of user pixels in H, η = 1, 2, ..., μ, μ refers to 2 jr2 The number of non-user pixels in the

[0085] y 2 jr =(Σ ζ＝1 λ (w 2 *y 2 jr1 ζ )+Σ η＝1 μ (w 1 *y 2 jr2 η )) / (λ*w 2 +μ*w 1 ), where y 2 jr1 ζ It means H 2 jr1 The ordinate of the ζ-th user pixel point in y 2 jr2 η It means H 2 jr2 The vertical coordinate of the nth non-user pixel in .

[0086] The second image region is regarded as a region with uneven quality. Then, according to the weight corresponding to each pixel point in the second image region, the weight of each pixel point in D jr Get the coordinates in D jr The second centroid coordinates G of the second image area 2 jr =(x 2 jr ,y 2 jr ), which provides a data basis for analyzing the overall center of gravity of users and non-users.

[0087] S260, according to G 1 jr , G 2 jr , H 1jr3 and H 2 jr3 , get D jr Corresponding user security information U jr = {G 1 jr , G 2 jr , H 1 jr3 , H 2 jr3}.

[0088] Above, by jr The first image area and the user pixel point set, the non-user pixel point set, and the inclination angle of the first target step corresponding to the first image area are extracted, and the weight distribution in the first image area corresponding to the non-user is characterized. The second image area and the user pixel point set, the non-user pixel point set, and the inclination angle of the second target step corresponding to the second image area are extracted, and the weight distribution in the second image area corresponding to the user is characterized, so as to combine the weight corresponding to each pixel point and each pixel point in D jr Get the coordinates in D jr The first and second center of gravity coordinates in the image provide a data basis for analyzing the center of gravity of users and non-users as a whole.

[0089] S300, according to U, obtain the first user safety probability set F corresponding to D = {F1, F2, ..., F j , ..., F n}, where F j ={F j1 , F j2 , ..., F jr , ..., F js}, F jr It means D jr The corresponding first user safety probability refers to the probability that the user safety status of the corresponding user is safe.

[0090] In one specific embodiment, the inclination angle of the first target step is D jr The first line between the lower left corner target point and the edge corner point of the first target step and D jr The angle between the bottom edges of the second target step is D jr The second line between the lower left corner target point and the edge corner point of the second target step and D jr The angle between the bottom edges of , S300 specifically includes the following steps: Figure 5 As shown:

[0091] S310, get G1 jr and G 2 jr The third link between

[0092] S320, the first intersection point between the first link and the third link, the second intersection point between the second link and the third link, G 1 jr and G 1 jr The highest point in is determined as the first target point, and the first intersection point, the second intersection point, G 1 jr and G 1 jr The lowest point in is determined as the second target point;

[0093] S330, set the first target point at D jr The projection point on the bottom edge of is determined as the third target point;

[0094] S340, acquiring a third image area according to the first target point, the second target point, and the third target point, where the third image area is an area enclosed by a first target line segment between the first target point and the second target point, a second target line segment between the first target point and the third target point, and a third target line segment between the second target point and the third target point;

[0095] S350, according to the category and coordinates of each pixel point in the third image area, w 1 and w 2 , obtain the third centroid coordinates G of the third image area 3 jr =(x 3 jr ,y 3 jr ), where x 3 jr refers to the abscissa of the third centroid of the third image region, y 3 jr is the ordinate of the third centroid of the third image region, G 3 jr Meet the following conditions:

[0096] x 3 jr =(Σ ε＝1 σ (w 2 *x 3 jr1 ε )+Σ ρ＝1 τ (w 1 *x 3 jr2ρ )) / (σ*w 2 +τ*w 1 ), where x 3 jr1 ε refers to the horizontal coordinate of the εth user pixel in the third image area, x 3 jr2 ρ refers to the horizontal coordinate of the ρth non-user pixel in the third image area, ε=1, 2, ..., σ, σ refers to the number of user pixels in the third image area, ρ=1, 2, ..., τ, τ refers to the number of non-user pixels in the third image area;

[0097] y 3 jr =(Σ ε＝1 σ (w 2 *y 3 jr1 ε )+Σ ρ＝1 τ (w 1 *y 3 jr2 ρ )) / (σ*w 2 +τ*w 1 ), where y 3 jr1 ε refers to the ordinate of the εth user pixel in the third image area, y 3 jr2 ρ refers to the ordinate of the ρth non-user pixel in the third image area;

[0098] S360, acquiring fourth centroid coordinates G corresponding to the third image area according to the coordinates of each pixel point in the third image area 4 jr =(x 4 jr ,y 4 jr ), where x 4 jr refers to the abscissa of the fourth centroid of the third image region, y 4 jr is the ordinate of the fourth centroid of the third image region, G 4 jr Meet the following conditions:

[0099] x 4 jr =(Σ ε＝1 σ (x3 jr1 ε )+Σ ρ＝1 τ (x 3 jr2 ρ )) / (σ+τ);

[0100] y 4 jr =(Σ ε＝1 σ (y 3 jr1 ε )+Σ ρ＝1 τ (y 3 jr2 ρ )) / (σ+τ);

[0101] S370, according to G 3 jr and G 4 jr , get F jr , where F jr Meet the following conditions:

[0102] When A j(k-1) <A jk When F jr =e^(-(x 4 jr -x 3 jr )-(y 3 jr -y 4 jr ));

[0103] When A j(k-1) >A jk When F jr =e^(-(x 3 jr -x 4 jr )-(y 3 jr -y 4 jr )).

[0104] The third image area includes user pixels, non-user pixels and background pixels. The third image area is regarded as an area with uneven quality. According to the weight corresponding to each pixel, the weight of each pixel in D jr Get the coordinates in D jr The third centroid coordinate G of the third image area in 3jr =(x 3 jr ,y 3 jr ).

[0105] At the same time, if the weight difference between user pixels, non-user pixels and background pixels is not considered, the weight of each pixel in D jr Get the coordinates in D jr The fourth centroid coordinate G corresponding to the third image area in 4 jr =(x 4 jr ,y 4 jr ).

[0106] Therefore, by comparing the positions of the third center of gravity coordinate and the fourth center of gravity coordinate, it is possible to determine the changes in the center of gravity of the user and non-user as a whole under the influence of the weight of the non-user and the weight of the user, and then to characterize the user's safety status based on the changes in the center of gravity, and obtain the monitoring results of the user's safety status.

[0107] Specifically, when A j(k-1) <A jk When, according to D jr From the left to the right, the steps in the jth target building area are raised successively. Compared with the fourth centroid coordinate, the abscissa x of the third centroid coordinate is 3 jr Less than the abscissa x of the fourth barycentric coordinate 4 jr , the ordinate y of the third barycentric coordinate 3 jr The ordinate y is greater than the fourth centroid coordinate 4 jr When , the user is more likely to dump, and x 4 jr -x 3 jr The larger the y 3 jr -y 4 jr The larger the value, the higher the probability of user dumping. Therefore, when A j(k-1) <A jk When F jr =e^(-(x 4 jr -x 3 jr )-(y 3 jr -y 4 jr )).

[0108] When A j(k-1) >A jk When, according to D jr From the left to the right, the steps in the jth target building area decrease successively. Then, compared with the fourth centroid coordinate, the abscissa x of the third centroid coordinate is 3 jr Greater than the abscissa x of the fourth barycentric coordinate 4 jr , the ordinate y of the third barycentric coordinate 3 jr The ordinate y is greater than the fourth centroid coordinate 4 jr When , the user is more likely to dump, and x 3 jr -x 4 jr The larger the y 3 jr -y 4 jr The larger the value, the higher the probability of user dumping. Therefore, when A j(k-1) <A jk When F jr =e^(-(x 3 jr -x 4 jr )-(y 3 jr -y 4 jr )).

[0109] In the above, considering the weight difference between user pixels, non-user pixels and background pixels, D jr The third centroid coordinates of the third image area in D are obtained without considering the weight difference between the user pixel points, the non-user pixel points and the background pixel points. jr The fourth center of gravity coordinate corresponding to the third image area in the step can be analyzed and compared with the positions of the third center of gravity coordinate and the fourth center of gravity coordinate in combination with the direction of the rise or fall of the step to characterize the overall center of gravity change of the user and the non-user under the influence of the weight of the non-user and the weight of the user, and then to characterize the safety status of the user, thereby improving the accuracy of the monitoring results of the user's safety status.

[0110] S400: According to F, determine the user safety status corresponding to each target building area.

[0111] Among them, the safety status includes two states: safe and unsafe. When the user safety status corresponding to the target building area is unsafe, it means that the corresponding user is more likely to fall on the steps when interacting with the corresponding non-user. In this case, it is necessary to guide the corresponding user's action behavior to reduce the probability of the user falling on the steps and improve the user's safety in the building area.

[0112] F jr It means D jr The probability that the corresponding user's user security status is safe, therefore, if F jr ≥E0, then determine D jr The corresponding user's user security status is safe, if F jr <E0, then determine D jr The user security status of the corresponding user is non-secure, and E0 refers to a preset security probability threshold.

[0113] In a specific implementation, S400 specifically includes the following steps: Figure 6 As shown:

[0114] S410, according to F, obtain the second user safety probability set E = {E1, E2, ..., E j ,……,E n}, the second user safety probability E corresponding to the jth target building area j Meet the following conditions:

[0115] E j =Σ r＝1 s ((K jr1 +K jr2 )*0.5*F jr ) / s, where K jr1 Refers to the first preset priority corresponding to the first target step, K jr2 refers to the second preset priority corresponding to the second target step;

[0116] S420, if E j ≥E0, the user safety status corresponding to the jth target building area is determined to be safe, and E0 refers to the preset safety probability threshold;

[0117] S430, if E j <E0, then the user safety status corresponding to the jth target building area is determined to be unsafe.

[0118] The risk of a user falling on a higher step is higher than that on a lower step, and the corresponding safety is lower. Therefore, in this embodiment, the steps are prioritized according to the cosine of the step height, and correspondingly, the higher the step, the lower the corresponding preset priority.

[0119] Combined with the first preset priority corresponding to the first target step and the second preset priority corresponding to the second target step, the second user safety probability E corresponding to the jth target building area is obtained. j , and in E j ≥E0, the user safety status corresponding to the jth target building area is monitored to be safe. j When <E0, the user safety status corresponding to the jth target building area is monitored to be unsafe.

[0120] In the above, priorities are set for the steps according to their heights, and combined with the first preset priority corresponding to the first target step and the second preset priority corresponding to the second target step, the second user safety probability corresponding to the j-th target building area is obtained to determine the user safety status corresponding to the j-th target building area, thereby improving the accuracy of the monitoring results of the user safety status.

[0121] This embodiment samples C according to a preset sampling interval T, and selects s ​​sampled images from each target video data as the basis for judging the user's safety status, which reduces the amount of calculation in the user's safety status monitoring process and improves the efficiency of user safety status monitoring. jr The first image area and the user pixel point set, the non-user pixel point set and the inclination angle of the first target step corresponding to the first image area are extracted to characterize the weight distribution in the first image area corresponding to the non-user, and the second image area and the user pixel point set, the non-user pixel point set and the inclination angle of the second target step corresponding to the second image area are extracted to characterize the weight distribution in the second image area corresponding to the user, thereby combining the weight corresponding to each pixel point and each pixel point in D jr Get the coordinates in D jr The first centroid coordinates and the second centroid coordinates in the image are obtained, and the first user safety probability corresponding to the non-user and the user as a whole in each sampled image is further obtained, so as to obtain the user safety status corresponding to each target building area, thereby improving the accuracy of the monitoring results of the user safety status.

[0122] Although some specific embodiments of the present invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It will also be appreciated by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A data processing system for monitoring the safety status of users in a building area, characterized in that: The system comprises a monitoring device and a plurality of target video acquisition devices, wherein the monitoring device comprises a processor and a memory storing a computer program, and the memory also stores a target building area information set A={A1, A2, ..., A j , ..., A n } and the target video dataset C = {C1, C2, ..., C j , ..., C n }, where A j ={A j1 , A j2 , ..., A jk , ..., A je(j) }, A jk is the inclination angle of the kth step in the jth target building area, C j It refers to the target video data of the user and non-user interaction collected by the target video acquisition device corresponding to the j-th target building area, j=1, 2, ..., n, n is the number of target building areas, k=1, 2, ..., e(j), e(j) is the number of steps in the j-th target building area, when the computer program is executed by the processor, the following steps are implemented: S100, sampling C according to a preset sampling interval T, and obtaining a sampled image data set D={D1, D2, ..., D j , ..., D n }, where D j ={D j1 , D j2 , ..., D jr , ..., D js }, D jr It means according to T to C j The rth image obtained by sampling, r=1, 2, ..., s, s is the number of images sampled from each target video data; S200, according to A, D, and the preset non-user weight W 1 and the preset user weight W 2 , obtain the user security information set U={U1, U2, ..., U j , ..., U n }, where U j ={U j1 , U j2 , ..., U jr , ..., U js }, D jr Corresponding user security information U jr Obtain it by following the steps below: S210, according to A and D jr , obtain the first image area information H 1 jr =(H 1 jr1 , H 1 jr2 , H 1 jr3 ) and the second image area information H 2 jr =(H 2 jr1 , H 2 jr2 , H 2 jr3 ), where H 1 jr1 D jr The user pixel point set in the first image area corresponding to the non-user, H 1 jr2 is the set of non-user pixels in the first image area, H 1 jr3 is the inclination angle of the first target step corresponding to the first image area, H 2 jr1 D jr The user pixel point set in the second image area corresponding to the user, H 2 jr2 is the set of non-user pixels in the second image area, H 2 jr3 is the inclination angle of the second target step corresponding to the second image area; S220, according to H 1 jr2 , H 2 jr2 and W 1 , get D jr The first weight w corresponding to each non-user pixel in 1 =W 1 / L 1 1, where L 1 1 refers to H 1 jr2 The number of non-user pixels and H 2 jr2 The sum of the number of non-user pixels in ; S230, according to H 1 jr1 , H 2 jr1 and W 2 , get D jr The second weight w corresponding to each user pixel in 2 =W 2 / L 2 2, where L 2 2 refers to H 1 jr1 The number of user pixels and H 2 jr1 The sum of the number of user pixels in ; S240, according to H 1 jr1 , H 1 jr2 、w 1 and w 2 , get D jr The first centroid coordinates G of the first image area in 1 jr ; S250, according to H 2 jr1 , H 2 jr2 、w 1 and w 2 , get D jr The second centroid coordinates G of the second image area 2 jr ; S260, according to G 1 jr , G 2 jr , H 1 jr3 and H 2 jr3 , get D jr Corresponding user security information U jr ={G 1 jr , G 2 jr , H 1 jr3 , H 2 jr3 }; S300, according to U, obtain the first user security probability set F={F1, F2, ..., F j , ..., F n }, where F j ={F j1 , F j2 , ..., F jr , ..., F js }, F jr It means D jr a corresponding first user safety probability, where the first user safety probability refers to a probability that the user safety status of the corresponding user is safe; S400: According to F, determine the user safety status corresponding to each target building area.

2. The data processing system according to claim 1, characterized in that The memory also stores a sample user's action step time set t={t1, t2, ..., t g ,……,t z }, where t g It refers to the action step time of the gth sample user. The action step time refers to the time interval from the left foot being lifted to the left foot being put down, or from the right foot being lifted to the right foot being put down when the sample user walks on flat ground. g=1, 2, ..., z, where z refers to the number of sample users. The preset sampling interval T meets the following conditions: T=W0×∑ g=1 z (t g / z), W0 refers to the preset step length weight.

3. The data processing system according to claim 1, characterized in that: S210 also includes the following steps: S211, for D jr Perform semantic segmentation and transform D jr The pixels in are classified into user pixels, non-user pixels and background pixels; S212, to D jr Performing bounding box detection, and determining an image area corresponding to a first bounding box not corresponding to the user as a first image area; S213, according to D jr The user pixel point and the first image area in H 1 jr1 ; S214, according to D jr The non-user pixel points and the first image area in H 1 jr2 ; S215, obtaining D according to the first bounding box and the e(j) steps in the j-th target building area. jr The corresponding first projection overlap length set Y jr 1 ={Y jr 11 , Y jr 12 , ..., Y jr 1k , ..., Y jr 1e(j) },Y jr 1k It means that the first bounding box is in D jr The first projection in the bottom direction of D jr The overlapping length between the second projections in the bottom direction; S216, max(Y jr 1 ) is determined as the first target step corresponding to the first image area, and H is obtained according to the inclination angle of the first target step. 1 jr3 , where max() is the maximum value function; S217, according to H 1 jr1 , H 1 jr2 and H 1 jr3 , obtain the first image area information H 1 jr =(H 1 jr1 , H 1 jr2 , H 1 jr3 ).

4. The data processing system according to claim 3, characterized in that: S210 also includes the following steps: S218, to D jr Performing bounding box detection, determining the image area corresponding to the second bounding box corresponding to the user as the second image area, and acquiring the third bounding box corresponding to the user's supporting leg; S219, according to D jr The user pixel point and the second image area in H 2 jr1 ; S2110, according to D jr The non-user pixel points and the second image area in H 2 jr2 ; S2111, according to the third bounding box and the e(j) steps in the j-th target building area, obtain D jr The corresponding second projection overlap length set Y jr 2 ={Y jr 21 , Y jr 22 , ..., Y jr 2k , ..., Y jr 2e(j) },Y jr 2k It means that the third bounding box is in D jr The third projection in the bottom direction of D jr The overlapping length between the second projections in the bottom direction; S2112, max(Y jr 2 ) is determined as the second target step corresponding to the second image area, and H is obtained according to the inclination angle of the second target step. 2 jr3 ; S2113, according to H 2 jr1 , H 2 jr2 and H 2 jr3 , obtain the second image area information H 2 jr =(H 2 jr1 , H 2 jr2 , H 2 jr3 ).

5. The data processing system according to claim 4, characterized in that: G 1 jr =(x 1 jr ,y 1 jr ), x 1 jr refers to the abscissa of the first centroid of the first image region, y 1 jr is the ordinate of the first centroid of the first image region, G 1 jr Meet the following conditions: x 1 jr =(Σ α=1 γ (w 2 × 2 jr1 α )+Σ β=1 δ (w 1 × 1 jr2 β )) / (γ×w 2 +δ×w 1 ), where w 1 It means D jr The first weight corresponding to each non-user pixel in , w 2 It means D jr The second weight corresponding to each user pixel in x 2 jr1 α It means H 1 jr1 The horizontal coordinate of the αth user pixel point in x 1 jr2 β It means H 1 jr2 The horizontal coordinate of the βth non-user pixel in H, α=1, 2, ..., γ, γ refers to 1 jr1 The number of user pixels in H, β = 1, 2, ..., δ, δ refers to 1 jr2 The number of non-user pixels in the y 1 jr =(Σ α=1 γ (w 2 ×y 2 jr1 α )+Σ β=1 δ (w 1 ×y 1 jr2 β )) / (γ×w 2 +δ×w 1 ), where y 2 jr1 α It means H 1 jr1 The ordinate of the αth user pixel point in y 1 jr2 β It means H 1 jr2 The ordinate of the βth non-user pixel in .

6. The data processing system according to claim 5, characterized in that: G 2 jr =(x 2 jr ,y 2 jr ), x 2 jr refers to the abscissa of the second centroid of the second image area, y 2 jr is the ordinate of the second centroid of the second image area, G 2 jr Meet the following conditions: x 2 jr =(Σ ζ=1 λ (w 2 × 2 jr1 ζ )+Σ η=1 μ (w 1 × 2 jr2 η )) / (λ×w 2 +μ×w 1 ), where x 2 jr1 ζ It means H 2 jr1 The horizontal coordinate of the ζ-th user pixel point in 2 jr2 η It means H 2 jr2 The horizontal coordinate of the nth non-user pixel in H, ζ=1, 2, ..., λ, λ refers to 2 jr1 The number of user pixels in H, η = 1, 2, ..., μ, μ refers to 2 jr2 The number of non-user pixels in the y 2 jr =(Σ ζ=1 λ (w 1 ×y 2 jr1 ζ )+Σ η=1 μ (w 2 ×y 2 jr2 η )) / (λ×w 2 +μ×w 1 ), where y 2 jr1 ζ It means H 2 jr1 The ordinate of the ζ-th user pixel point in y 2 jr2 η It means H 2 jr2 The vertical coordinate of the nth non-user pixel in .

7. The data processing system according to claim 1, characterized in that: W 1 Obtain it by following the steps below: S01, identifying the first image region to obtain a predicted category of an object other than a user in the first image region; S02, determining the preset object weight corresponding to the predicted object category as W 1 .

8. The data processing system according to claim 1, characterized in that: W 2 Obtain it by following the steps below: S03, identifying the second image area, and obtaining a predicted height of the user in the second image area; S04: Determine the preset user weight corresponding to the predicted height as W 2 .

9. The data processing system according to claim 5, characterized in that: The inclination angle of the first target step is D jr The first line between the lower left corner target point and the edge corner point of the first target step and D jr The angle between the bottom edges of the second target step is D jr The second line between the lower left corner target point and the edge corner point of the second target step and D jr The angle between the bottom edges of , S300 specifically includes the following steps: S310, get G 1 jr and G 2 jr The third link between S320, a first intersection point between the first link and the third link, a second intersection point between the second link and the third link, G 1 jr and G 2 jr The highest point in G is determined as the first target point, and the first intersection point, the second intersection point, G 1 jr and G 2 jr The lowest point in is determined as the second target point; S330, the first target point is located at D jr The projection point on the bottom edge of is determined as the third target point; S340, acquiring a third image area according to the first target point, the second target point, and the third target point, the third image area being an area enclosed by a first target line segment between the first target point and the second target point, a second target line segment between the first target point and the third target point, and a third target line segment between the second target point and the third target point; S350, according to the category and coordinates of each pixel point in the third image area, D jr The first weight w corresponding to each non-user pixel in 1 and D jr The second weight w corresponding to each user pixel in 2 , obtain the third centroid coordinates G of the third image area 3 jr =(x 3 jr ,y 3 jr ), where x 3 jr refers to the abscissa of the third centroid of the third image area, y 3 jr refers to the ordinate of the third centroid of the third image area, G 3 jr Meet the following conditions: x 3 jr =(Σ ε=1 σ (w 2 × 3 jr1 ε )+Σ ρ=1 τ (w 1 × 3 jr2 ρ )) / (σ×w 2 +τ×w 1 ), where x 3 jr1 ε refers to the horizontal coordinate of the εth user pixel point in the third image area, x 3 jr2 ρ refers to the horizontal coordinate of the ρth non-user pixel in the third image area, ε=1, 2, ..., σ, σ refers to the number of user pixels in the third image area, ρ=1, 2, ..., τ, τ refers to the number of non-user pixels in the third image area; y 3 jr =(Σ ε=1 σ (w 2 ×y 3 jr1 ε )+Σ ρ=1 τ (w 1 ×y 3 jr2 ρ )) / (σ×w 2 +τ×w 1 ), where y 3 jr1 ε refers to the ordinate of the εth user pixel point in the third image area, y 3 jr2 ρ refers to the ordinate of the pth non-user pixel in the third image area; S360, acquiring fourth centroid coordinates G corresponding to the third image area according to the coordinates of each pixel point in the third image area 4 jr =(x 4 jr ,y 4 jr ), where x 4 jr refers to the abscissa of the fourth centroid of the third image area, y 4 jr is the ordinate of the fourth centroid of the third image area, G 4 jr Meet the following conditions: x 4 jr =(S ε=1 σ (x 3 jr1 ε )+S ρ=1 τ (x 3 jr2 ρ )) / (σ+τ); y 4 jr =(S ε=1 σ (y 3 jr1 ε )+S ρ=1 τ (y 3 jr2 ρ )) / (σ+τ); S370, according to G 3 jr and G 4 jr , get D jr The corresponding first user safety probability F jr , where D jr The corresponding first user safety probability F jr Meet the following conditions: When A j(k-1) <A jk When F jr =e^(-(x 4 jr -x 3 jr )-(y 3 jr -y 4 jr )); When A j(k-1) >A jk When F jr =e^(-(x 3 jr -x 4 jr )-(y 3 jr -y 4 jr )).

10. The data processing system according to claim 1, characterized in that: S400 specifically includes the following steps: S410, according to F, obtain the second user safety probability set E={E1, E2, . . . , E j ,……,E n }, the second user safety probability E corresponding to the jth target building area j Meet the following conditions: E j =Σ r=1 s ((K jr1 +K jr2 )×0.5×F jr ) / s, where K jr1 refers to the first preset priority corresponding to the first target step, K jr2 refers to the second preset priority corresponding to the second target step, F jr It means D jr The corresponding first user safety probability; S420, if E j ≥E0, the user safety status corresponding to the jth target building area is determined to be safe, and E0 refers to the preset safety probability threshold; S430, if E j <E0, then the user safety status corresponding to the jth target building area is determined to be unsafe.

Images