A building corridor personnel abnormal contact detection method, device and medium

Abstract

The application relates to a building corridor personnel abnormal contact detection method, equipment and medium, the method comprising: acquiring the pixel point number of a personnel contour, personnel identification frame information, personnel boundary coordinates and personnel key point information in a corridor personnel; determining whether there is a coordinate missing situation in the personnel key point information; if there is a coordinate missing situation, different corrections are performed according to different parts of the coordinate missing situation to obtain the pixel point number of the corrected personnel contour; determining a depth estimation value according to a pre-trained standard personnel depth detection model, the boundary coordinates and the pixel point number of the corrected personnel contour; determining a personnel relative distance according to the depth estimation value and the identification frame information; determining a personnel instantaneous contact score according to the personnel relative distance; determining a personnel abnormal contact score according to the personnel instantaneous contact score; judging the size of the personnel abnormal contact score and a preset threshold value to determine whether the personnel has a personnel abnormal contact behavior and further determine whether to alarm.

Description

Technical Field

[0001] This application relates to the field of prevention and control technology, and in particular to a method, equipment and medium for detecting abnormal contact of people in building corridors. Background Technology

[0002] With limited manpower for epidemic prevention and control, relying solely on manual supervision makes it difficult to achieve effective control without any oversights. Furthermore, existing monitoring and analysis technologies have significant errors in estimating the distance between people, which cannot guarantee the accuracy of monitoring results. Summary of the Invention

[0003] To address the aforementioned issues, this application provides a method, electronic device, and storage medium for detecting abnormal contact between people in building corridors.

[0004] In a first aspect, embodiments of this application provide a method for detecting abnormal contact among people in a building corridor, including:

[0005] The system obtains the number of pixels in the outline of the person to be detected in the building corridor, the bounding box information of the person to be detected, the boundary coordinates of the person to be detected, and the key point information of the person to be detected.

[0006] Determine if there are any missing coordinates in the key information of the person to be tested;

[0007] If it is determined that there are missing coordinates in the key point information of the person to be inspected, different corrections are made to the body shape of the person to be inspected according to the different parts of the missing coordinates to obtain the number of pixels of the corrected person outline.

[0008] The depth estimate of the person to be detected is determined based on a pre-trained standard person depth detection model, the boundary coordinates of the person to be detected, and the number of pixels in the corrected person contour.

[0009] The relative distance between the persons is determined based on the depth estimate of the persons to be detected and the identification box information of the persons to be detected;

[0010] The instantaneous contact score of the person being tested is determined based on the relative distance between the persons being tested;

[0011] The abnormal contact score of the person being tested is determined based on the instantaneous contact score of the person being tested.

[0012] The system compares the score of abnormal contact with a person with a preset threshold. If the score is greater than the preset threshold, the system determines that an abnormal contact has occurred and sends an alarm message.

[0013] The key information for personnel includes: the coordinates of the personnel's shoulders, neck, hips, and knees.

[0014] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the number of pixels in the corrected person's outline is obtained by applying different corrections to the body shape of the person to be detected based on different locations where coordinates are missing, including:

[0015] When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person being inspected are not completely missing, the contour of the person being inspected is divided into two parts using the dividing line y = b1, and the number of pixels n2 and n3 in the two parts are determined. Here, n2 is the number of pixels in the region below the line with a ordinate less than b1, and n1 = n2 + n3.

[0016] The number of pixels in the corrected person's outline is m1 = m2 + m3, obtained by making different corrections to the person's body shape.

[0017] When the coordinates of the shoulders (sx1, sy1), (sx2, sy2) and the coordinates of the knees (kx1, ky1), (kx2, ky2) of the person being tested are not missing, m2 and m3 are determined by the following formula:

[0018]

[0019]

[0020] in,

[0021]

[0022]

[0023]

[0024]

[0025] h represents the height of a standard person, s represents the shoulder width of a standard person, e represents the waist width of a standard person, k represents the knee distance of a standard person, (nx, ny0) represents the neck coordinates of the person to be tested, c1 is the first correction constant obtained from training with historical data, c2 is the second correction constant obtained from training with historical data, and c3 is the third correction constant obtained from training with historical data.

[0026] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the method further includes: adjusting the number of pixels in the corrected person's outline based on different locations with missing coordinates.

[0027] When the coordinates (sx1, sy10, (sx2, sy2) of the shoulders of the person being inspected are not missing, but the coordinates (kx1, ky1), (kx2, ky2) of the knees of the person being inspected are missing, the number of pixels of the corrected person outline is m1 = m2 + m3 after different corrections are made to the body shape of the person being inspected. m2 and m3 are determined by the following formula:

[0028]

[0029] in,

[0030]

[0031]

[0032]

[0033] c4 is the fourth correction constant obtained from training with historical data, and c6 is the sixth correction constant obtained from training with historical data.

[0034] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the method further includes: adjusting the number of pixels in the corrected person's outline based on different locations with missing coordinates.

[0035] When the knee coordinates (kx1, ky1) and (kx2, ky2) of the person being inspected are not missing, but the shoulder coordinates (sx1, sy1) and (sx2, sy2) are missing, the number of pixels in the corrected person's outline is m1 = m2 + m3 after different corrections are made to the person's body shape. m2 and m3 are determined by the following formula:

[0036]

[0037]

[0038] in,

[0039]

[0040]

[0041]

[0042] c5 is the fifth correction constant obtained from training on historical data, and c7 is the seventh correction constant obtained from training on historical data.

[0043] Further, in the above-mentioned method for detecting abnormal contact among people in building corridors, based on different locations of missing coordinates, the number of pixels in the corrected person's outline is obtained by modifying the person's body shape in different ways.

[0044] When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person to be inspected are completely missing, the number of pixels m1 of the corrected person outline obtained by correcting the person's body shape is determined by the following formula:

[0045] m1 = c8n1, where c8 is the eighth correction constant obtained from training with historical data.

[0046] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the relative distance D between people is determined based on the depth estimate of the person to be detected and the identification box information of the person to be detected, using the following formula:

[0047]

[0048]

[0049] Where d1 and d2 represent the depth estimates of the first person and the second person, respectively; c9 represents the ninth correction constant obtained from training with historical data; (x1,y1,w1,h1) and (x2,y2,w2,h2) represent the personnel bounding box coordinates of the first person and the second person, respectively; L represents the length of the corridor; W1 represents the width of the corridor at one end of the corridor camera; W2 represents the width of the corridor at the other end of the corridor camera; W2 and W3 represent the actual width of the corridor.

[0050] Furthermore, in the aforementioned method for detecting abnormal contact among people in building corridors, the instantaneous contact score gd is obtained based on the relative distance D between people. i It is determined by the following formula:

[0051]

[0052] Where i represents the number of frames in the camera's video feed, ts1 is the first threshold, and ts2 is the second threshold.

[0053] Furthermore, in the aforementioned method for detecting abnormal contact among people in building corridors, the instantaneous contact score (gd) is used to determine the contact level. i The abnormal contact score G for personnel is determined, including:

[0054] Obtain the instantaneous contact score gd of people in N consecutive video frames. i ;

[0055] Based on the instantaneous contact score of the personnel, gdi The abnormal contact score G is calculated using the following formula:

[0056]

[0057] In a second aspect, embodiments of the present invention also provide an electronic device, including: a processor and a memory;

[0058] The processor executes the abnormal contact detection method for people in building corridors as described above by calling the program or instructions stored in the memory.

[0059] Thirdly, embodiments of the present invention also provide a computer-readable storage medium storing a program or instructions that cause a computer to execute the abnormal contact detection method for people in a building corridor as described in any of the above claims.

[0060] The advantages of this application's embodiments are as follows: This application obtains the number of pixels in the silhouette of people in the corridor, the information of the personnel identification box, the personnel boundary coordinates, and the information of the personnel key points; determines whether there are missing coordinates in the personnel key point information; if there are missing coordinates, different corrections are made according to different parts of the missing coordinates to obtain the number of pixels in the corrected personnel silhouette; determines the depth estimate based on the pre-trained standard personnel depth detection model, boundary coordinates, and the number of pixels in the corrected personnel silhouette; determines the relative distance of people based on the depth estimate and the identification box information; determines the instantaneous contact score of people based on the relative distance of people; determines the abnormal contact score of people based on the instantaneous contact score of people; judges the size of the abnormal contact score of people and the preset threshold to determine whether people have engaged in abnormal contact behavior; if abnormal contact behavior of people occurs, an alarm information is sent, and management personnel can promptly control people based on the alarm information. Moreover, management personnel do not need to go to the site to check or monitor in real time, resulting in a high user experience. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of a method for detecting abnormal contact of people in a building corridor, provided in an embodiment of this application.

[0063] Figure 2This is a schematic block diagram of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0064] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0066] Figure 1 This is a schematic diagram of a method for detecting abnormal contact of people in a building corridor, provided in an embodiment of this application.

[0067] In a first aspect, embodiments of this application provide a method for detecting abnormal contact among people in a building corridor, comprising eight steps from S101 to S108:

[0068] S101: Obtain the number of pixels of the outline of the person to be detected in the building corridor, the bounding box information of the person to be detected, the boundary coordinates of the person to be detected, and the key point information of the person to be detected.

[0069] Specifically, in this embodiment of the application, the number of pixels n1 of the outline of the person to be detected in the building corridor and the identification box information (x) of the person to be detected are obtained. j y j w j h j ), where x j To identify the top-left x-coordinate of the box containing the person to be tested, y j To identify the top-left ordinate of the box containing the person to be tested, w j To identify the width of the box for the person to be tested, h j The height of the bounding box for the person to be detected is given by j, where j is the sequence number of the person in the image, and the boundary coordinates of the person to be detected are (x, y, j). j +0.5w j y j +h jThe key information of the personnel includes: the coordinates of the personnel's shoulders (sx1, sy1), (sx2, sy2), the coordinates of the personnel's neck (nx, ny), the coordinates of the personnel's hip bones (gx1, gy1), (gx2, gy2), and the coordinates of the personnel's knees (kx1, ky1), (kx2, ky2).

[0070] S102: Determine whether there are any missing coordinates in the key point information of the person to be inspected.

[0071] Specifically, in this embodiment of the application, the key point information of the person obtained includes: the coordinates of the person's shoulders (sx1, sy1), (sx2, sy2), the coordinates of the person's neck (nx, ny), the coordinates of the person's hip (gx1, gy1), (gx2, gy2), and the coordinates of the person's knees (kx1, ky1), (kx2, ky2), and further determines whether there are any missing coordinates in the key point information of the person to be detected, such as missing knee coordinates or missing neck coordinates.

[0072] S103: If it is determined that there are missing coordinates in the key point information of the person to be inspected, different corrections are made to the body shape of the person to be inspected according to the different parts of the missing coordinates to obtain the number of pixels of the corrected person outline.

[0073] Specifically, in the embodiments of this application, the following describes a method for obtaining the number of pixels m1 of the corrected person's outline by making different corrections to the body shape of the person to be detected under different coordinate missing conditions.

[0074] S104: Determine the depth estimate of the person to be detected based on the pre-trained standard person depth detection model, the boundary coordinates of the person to be detected, and the number of pixels in the corrected person contour.

[0075] Specifically, in this embodiment, the pre-trained standard person depth detection model is obtained by selecting a standard person of normal build, recording the standard person's height h, shoulder width s, waist width e, and knee distance k, dividing the corridor from one end of the camera to the other into n equal intervals, where n is a set positive integer, recording the corridor length L, the corridor width W1 at one end of the corridor in the image and the corridor width W2 at the other end, and the actual width W3 of the corridor, collecting images of the standard person walking and standing in different directions in each interval, where the person's direction includes three directions: front, oblique, and side, labeling the standard person in the image, adding depth labels to the person to be detected in different intervals, where the depth label represents the distance from the person to one end of the corridor where the camera is located, recording the number of pixels in the outline of the person to be detected, and using the labeled standard person images to train the standard person depth detection model.

[0076] The number of pixels m1 in the corrected personnel outline and the personnel boundary coordinates (x) j +0.5w j ,y j + h j Substituting the values ​​into the personnel depth detection model, we obtain the estimated depth value d of the person to be detected. j .

[0077] S105: Determine the relative distance between the person and the person to be detected based on the depth estimate and the identification box information of the person to be detected.

[0078] S106: Determine the instantaneous contact score of the person being tested based on the relative distance between the persons being tested and the person being tested.

[0079] S107: Determine the abnormal contact score of the person to be tested based on the instantaneous contact score of the person to be tested.

[0080] Specifically, in the embodiments of this application, the methods for determining the relative distance between people, determining the instantaneous contact score of the person to be tested, and determining the instantaneous contact score of the person to be tested are described below with specific formulas.

[0081] S108: Determine the difference between the abnormal contact score of the person to be tested and the preset threshold. If the result is that the abnormal contact score of the person to be tested is greater than the preset threshold, it is determined that there is abnormal contact behavior of the person to be tested, and an alarm message is sent.

[0082] Specifically, in this embodiment, the abnormal contact score of the person to be tested is compared with a preset threshold to determine whether the person to be tested has engaged in abnormal contact behavior. If abnormal contact behavior occurs, an alarm message is sent, thereby achieving effective control of personnel during the prevention and control work.

[0083] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the number of pixels in the corrected person's outline is obtained by applying different corrections to the body shape of the person to be detected based on different locations where coordinates are missing, including:

[0084] When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person being inspected are not completely missing, the contour of the person being inspected is divided into two parts using the dividing line y = b1, and the number of pixels n2 and n3 in the two parts are determined. Here, n2 is the number of pixels in the region below the line with a ordinate less than b1, and n1 = n2 + n3.

[0085] The number of pixels in the corrected person's outline is m1 = m2 + m3, obtained by making different corrections to the person's body shape.

[0086] When the coordinates of the shoulders (sx1, sy1), (sx2, sy2) and the coordinates of the knees (kx1, ky1), (kx2, ky2) of the person being tested are not missing, m2 and m3 are determined by the following formula:

[0087]

[0088]

[0089] in,

[0090]

[0091]

[0092]

[0093]

[0094] h represents the height of a standard person, s represents the shoulder width of a standard person, e represents the waist width of a standard person, k represents the knee distance of a standard person, (nx, ny) represents the neck coordinates of the person to be tested, c1 is the first correction constant obtained from training with historical data, c2 is the second correction constant obtained from training with historical data, and c3 is the third correction constant obtained from training with historical data.

[0095] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the method further includes: adjusting the number of pixels in the corrected person's outline based on different locations with missing coordinates.

[0096] When the coordinates (sx1, sy1) and (sx2, sy2) of the shoulders of the person being inspected are not missing, but the coordinates (kx1, ky1) and (kx2, ky2) of the knees of the person being inspected are missing, the number of pixels of the corrected person outline is m1 = m2 + m3 after different corrections are made to the body shape of the person being inspected. m2 and m3 are determined by the following formula:

[0097]

[0098] in,

[0099]

[0100]

[0101]

[0102] c4 is the fourth correction constant obtained from training with historical data, and c6 is the sixth correction constant obtained from training with historical data.

[0103] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the method further includes: adjusting the number of pixels in the corrected person's outline based on different locations with missing coordinates.

[0104] When the knee coordinates (kx1, ky1) and (kx2, ky2) of the person being inspected are not missing, but the shoulder coordinates (sx1, sy1) and (sx2, sy2) are missing, the number of pixels in the corrected person's outline is m1 = m2 + m3 after different corrections are made to the person's body shape. m2 and m3 are determined by the following formula:

[0105]

[0106]

[0107] in,

[0108]

[0109]

[0110]

[0111] c5 is the fifth correction constant obtained from training on historical data, and c7 is the seventh correction constant obtained from training on historical data.

[0112] Furthermore, in the above-mentioned method for detecting abnormal contact among people in a building corridor, the method further includes adjusting the size of the person being detected based on the different locations of missing coordinates to obtain the corrected number of pixels in the person's outline.

[0113] When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person to be inspected are completely missing, the number of pixels m1 of the corrected person outline obtained by correcting the person's body shape is determined by the following formula:

[0114] m1 = c8n1, where c8 is the eighth correction constant obtained from training with historical data.

[0115] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the relative distance D of the person to be detected is determined based on the depth estimate and the identification box information of the person to be detected, using the following formula:

[0116]

[0117]

[0118] Where d1 and d2 represent the depth estimates of the first person and the second person, respectively; c9 represents the ninth correction constant obtained from training with historical data; (x1,y1,w1,h1) and (x2,y2,w2,h2) represent the personnel bounding box coordinates of the first person and the second person, respectively; L represents the length of the corridor; W1 represents the width of the corridor at one end of the corridor camera; W2 represents the width of the corridor at the other end of the corridor camera; W2 and W3 represent the actual width of the corridor.

[0119] Furthermore, in the above-mentioned method for detecting abnormal contact among people in building corridors, the instantaneous contact score gd of the person to be detected is obtained based on the relative distance D between the person and the person being detected. i It is determined by the following formula:

[0120]

[0121] Where i represents the number of frames in the camera's video feed, ts1 is the first threshold, and ts2 is the second threshold.

[0122] Specifically, in the embodiments of this application, ts1 and ts2 can be flexibly determined according to the actual situation.

[0123] Furthermore, in the aforementioned method for detecting abnormal contact among people in building corridors, the instantaneous contact score (gd) of the person to be detected is used as the basis for detection. i The abnormal exposure score G of the person to be tested is determined, including:

[0124] Obtain the instantaneous contact score (gd) of the person to be detected in N consecutive video frames. i ;

[0125] Based on the instantaneous contact score of the person to be tested, gd i The abnormal exposure score G of the person being tested is determined by the following formula:

[0126]

[0127] Specifically, in this embodiment of the application, the instantaneous contact score of the person to be detected in N consecutive frames of video images is gd. i The abnormal contact score G of the person to be tested is obtained by summing the results.

[0128] In a second aspect, embodiments of the present invention also provide an electronic device, including: a processor and a memory;

[0129] The processor executes the abnormal contact detection method for people in building corridors as described above by calling the program or instructions stored in the memory.

[0130] Thirdly, embodiments of the present invention also provide a computer-readable storage medium storing a program or instructions that cause a computer to execute the abnormal contact detection method for people in a building corridor as described in any of the above claims. Figure 2 This is a schematic block diagram of an electronic device provided in an embodiment of this disclosure.

[0131] like Figure 2 As shown, the electronic device includes at least one processor 201, at least one memory 202, and at least one communication interface 203. The various components in the electronic device are coupled together via a bus system 204. The communication interface 203 is used for information transmission with external devices. It is understood that the bus system 204 is used to implement communication between these components. In addition to a data bus, the bus system 204 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 204.

[0132] It is understood that the memory 202 in this embodiment can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0133] In some implementations, memory 202 stores elements such as executable units or data structures, or subsets thereof, or extended sets thereof: operating systems and applications.

[0134] The operating system includes various system programs, such as the framework layer, core library layer, and driver layer, used to implement various basic business functions and handle hardware-based tasks. The application programs include various applications, such as media players and browsers, used to implement various application functions. The program implementing any method in the abnormal contact detection method for people in a building corridor provided in this application embodiment can be included in the application programs.

[0135] In this embodiment of the application, the processor 201 executes the steps of various embodiments of the abnormal contact detection method for people in building corridors provided in this application by calling the program or instructions stored in the memory 202, specifically, the program or instructions stored in the application.

[0136] The system obtains the number of pixels in the outline of the person to be detected in the building corridor, the bounding box information of the person to be detected, the boundary coordinates of the person to be detected, and the key point information of the person to be detected.

[0137] Determine if there are any missing coordinates in the key information of the person to be tested;

[0138] If it is determined that there are missing coordinates in the key point information of the person to be inspected, different corrections are made to the body shape of the person to be inspected according to the different parts of the missing coordinates to obtain the number of pixels of the corrected person outline.

[0139] The depth estimate of the person to be detected is determined based on a pre-trained standard person depth detection model, the boundary coordinates of the person to be detected, and the number of pixels in the corrected person contour.

[0140] The relative distance between the persons is determined based on the depth estimate of the persons to be detected and the identification box information of the persons to be detected;

[0141] The instantaneous contact score of personnel is determined based on their relative distance.

[0142] An abnormal contact score is determined based on the instantaneous contact score.

[0143] The system compares the score of abnormal contact with a person with a preset threshold. If the score is greater than the preset threshold, the system determines that an abnormal contact has occurred and sends an alarm message.

[0144] The key information for personnel includes: the coordinates of the personnel's shoulders, neck, hips, and knees.

[0145] Any method in the abnormal contact detection method for personnel in a building corridor provided in this application embodiment can be applied to, or implemented by, the processor 201. The processor 201 can be an integrated circuit chip with signal capabilities. During implementation, each step of the above method can be completed by the integrated logic circuits in the hardware of the processor 201 or by instructions in software form. The processor 201 can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional device.

[0146] The steps of any method in the abnormal contact detection method for people in a building corridor provided in this application embodiment can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software units in the decoding processor. The software units can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 202. The processor 201 reads the information in memory 202 and, in conjunction with its hardware, completes the steps of the abnormal contact detection method for people in a building corridor.

[0147] Those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments.

[0148] Those skilled in the art will understand that the descriptions of the various embodiments have different focuses, and for parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0149] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for detecting abnormal contact among people in a building corridor, characterized in that, include: The system obtains the number of pixels in the outline of the person to be detected in the building corridor, the bounding box information of the person to be detected, the boundary coordinates of the person to be detected, and the key point information of the person to be detected. Determine whether there are any missing coordinates in the key point information of the person to be tested; If it is determined that there are missing coordinates in the key point information of the person to be detected, the number of pixels of the corrected person outline is obtained by making different corrections according to the different parts of the missing coordinates. The depth estimate of the person to be detected is determined based on a pre-trained standard person depth detection model, the boundary coordinates of the person to be detected, and the number of pixels in the corrected person contour. The relative distance between the persons is determined based on the depth estimate of the persons to be detected and the identification box information of the persons to be detected; The instantaneous contact score gd is determined based on the relative distance D between the individuals to be tested. i ; Based on the instantaneous contact score of the person to be tested, gd i Determine the score G for abnormal contact with personnel; The abnormal contact score of the person to be tested is compared with a preset threshold. If the result is that the abnormal contact score of the person to be tested is greater than the preset threshold, it is determined that the person to be tested has engaged in abnormal contact behavior, and an alarm message is sent. The key information of the person to be tested includes: the coordinates of the person's shoulders, neck, hips, and knees; The determination of the relative distance D between the person and the person being detected, based on the estimated depth value and the identification box information of the person being detected, is made using the following formula: Where d1 and d2 represent the depth estimates of the first person and the second person, respectively; c9 represents the ninth correction constant obtained from training with historical data; (x1,y1,w1,h1) and (x2,y2,w2,h2) represent the personnel bounding box coordinates of the first person and the second person, respectively; L represents the length of the corridor; W1 represents the width of the corridor at one end of the corridor camera; W2 represents the width of the corridor at the other end of the corridor camera; and W3 represents the actual width of the corridor.

2. The method for detecting abnormal contact among people in a building corridor according to claim 1, characterized in that, The process of applying different corrections to the body shape of the person being inspected based on the different locations of missing coordinates to obtain the number of pixels in the corrected person's outline includes: When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person being inspected are not completely missing, the dividing line y = b1 is used to divide the outline of the person being inspected into two parts, and the number of pixels n2 and n3 in the two parts are determined. Here, n2 is the number of pixels in the region below the line with a ordinate less than b1, and n1 = n2 + n3. The number of pixels in the corrected person's outline is m1 = m2 + m3, obtained by making different corrections to the person's body shape. When the coordinates of the shoulders (sx1, sy1), (sx2, sy2) and the coordinates of the knees (kx1, ky1), (kx2, ky2) of the person being tested are not missing, m2 and m3 are determined by the following formula: in, h represents the height of a standard person, s represents the shoulder width of a standard person, e represents the waist width of a standard person, k represents the knee distance of a standard person, (nx, ny) represents the neck coordinates of the person to be tested, c1 is the first correction constant obtained from training with historical data, c2 is the second correction constant obtained from training with historical data, and c3 is the third correction constant obtained from training with historical data.

3. The method for detecting abnormal contact among people in a building corridor according to claim 2, characterized in that, The method of obtaining the number of pixels in the corrected human silhouette by applying different corrections to the body shape of the person to be detected based on different locations with missing coordinates also includes: When the coordinates (sx1, sy1) and (sx2, sy2) of the shoulders of the person being inspected are not missing, but the coordinates (kx1, ky1) and (kx2, ky2) of the knees are missing, the number of pixels in the corrected person's outline is m1 = m2 + m3 after different corrections are made to the person's body shape. m2 and m3 are determined by the following formula: in, c4 is the fourth correction constant obtained from training with historical data, and c6 is the sixth correction constant obtained from training with historical data.

4. The method for detecting abnormal contact among people in a building corridor according to claim 2, characterized in that, The method of obtaining the number of pixels in the corrected human silhouette by applying different corrections to the body shape of the person to be detected based on different locations with missing coordinates also includes: When the knee coordinates (kx1, ky1) and (kx2, ky2) of the person being inspected are not missing, but the shoulder coordinates (sx1, sy1) and (sx2, sy2) are missing, the number of pixels in the corrected person's outline is m1 = m2 + m3 after different corrections are made to the person's body shape. m2 and m3 are determined by the following formula: in, c5 is the fifth correction constant obtained from training on historical data, and c7 is the seventh correction constant obtained from training on historical data.

5. The method for detecting abnormal contact among people in a building corridor according to claim 2, characterized in that, The method of obtaining the number of pixels in the corrected human silhouette by applying different corrections to the body shape of the person to be detected based on different locations with missing coordinates also includes: When the hip coordinates (gx1, gy1) and (gx2, gy2) of the person to be inspected are completely missing, the number of pixels m1 of the corrected person outline obtained by correcting the person's body shape is determined by the following formula: m1 = c8n1, where c8 is the eighth correction constant obtained from training with historical data.

6. The method for detecting abnormal contact of people in a building corridor according to claim 1, characterized in that, The instantaneous contact score gd of the person being tested is obtained based on the relative distance D between the persons being tested. i It is determined by the following formula: Where i represents the number of frames in the camera's video feed, ts1 is the first threshold, and ts2 is the second threshold.

7. The method for detecting abnormal contact of people in a building corridor according to claim 1, characterized in that, Based on the instantaneous contact score of the person to be tested, gd i The abnormal contact score G for personnel is determined, including: Obtain the instantaneous contact score gd of people in N consecutive video frames. i ; Based on the instantaneous contact score of the personnel, gd i The abnormal contact score G is calculated using the following formula:

8. An electronic device, characterized in that, include: Processor and memory; The processor executes the abnormal contact detection method for people in building corridors as described in any one of claims 1 to 7 by calling the program or instructions stored in the memory.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instructions that cause a computer to perform a method for detecting abnormal contact of people in a building corridor as described in any one of claims 1 to 7.

Images