Monitoring system for preventing elevator from carrying electric vehicle and elevator

Abstract

The invention relates to the technical field of elevator safety monitoring, in particular to a monitoring device for preventing an elevator from carrying an electric vehicle and the elevator, and the system comprises a target detection module, a contour matching module, a track capturing module, an early warning triggering module and an elevator control module. According to the method, by fusing the image edge gradient, the morphological characteristics and the pixel contact characteristics, the target contour is accurately extracted, environmental interference is reduced, structural matching analysis is utilized, human body natural characteristic edges are removed, the electric vehicle recognition accuracy is improved, track capture is combined with human body key point displacement and the track change rate, and a path continuity threshold value is set; the method comprises the following steps: stabilizing a target form, avoiding the influence of shielding or angle change on recognition, triggering an alarm only when an illegal behavior is clear based on contour proportion and staying duration data, reducing false alarm interference, linking an elevator control system with an alarm signal, dynamically setting locking time to execute a running stopping instruction, uploading an intrusion record, forming traceable management, and improving the safety of an elevator. And the safety supervision capability is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator safety monitoring, and in particular to a monitoring system and an elevator for preventing an elevator from carrying an electric vehicle. Background Art

[0002] The field of elevator safety monitoring technology includes a comprehensive management system for elevator operating status, fault detection, abnormal behavior monitoring and safety protection. The core content of this technical field involves real-time monitoring of elevators, data collection and analysis, illegal behavior identification, automatic alarm and safety warning. The system is usually composed of sensor equipment, video monitoring, data transmission, intelligent analysis devices, etc., aiming to improve the safety of elevators and prevent illegal use, elevator accidents and abnormal situations. In recent years, with the increase in the use rate of electric vehicles, the number of accidents such as fires and explosions caused by electric vehicles illegally entering elevators has increased. Elevator safety monitoring technology has further expanded to the prevention and management of electric vehicle transportation behavior to ensure the safety of elevator passengers and building environment.

[0003] Among them, the monitoring system to prevent elevators from carrying electric vehicles refers to a security monitoring system specifically used to identify, detect and intervene in the entry of electric vehicles into the elevator. The device usually includes detection methods such as image recognition, weight detection, and infrared sensing to determine whether there are electric vehicles in the elevator. When an electric vehicle is detected entering the elevator, the system will remind or prevent it through voice alarms, warning lights, remote control signals, etc. Some devices are also combined with the elevator access control system to prevent electric vehicles from entering the elevator or making the elevator unable to start. In addition, the device can analyze the characteristics of electric vehicles based on the image recognition technology of the camera inside the elevator using a deep learning model to achieve automatic identification. By combining a variety of monitoring methods, the device can effectively improve the safety of elevator transportation and avoid the safety hazards caused by electric vehicles illegally entering the elevator.

[0004] In the existing elevator management process, when detecting electric vehicles, it is limited to a single detection method and is easily disturbed by environmental factors, resulting in insufficient recognition accuracy. For example, weight detection is affected by multiple passengers in the elevator, making it difficult to accurately distinguish electric vehicles from other large luggage; infrared sensing is easily affected by external temperature changes, resulting in false alarms or missed alarms. In traditional image recognition solutions, simple feature matching methods are often used, which have weak adaptability to the structure of electric vehicles. In the case of complex light, partial occlusion or diverse electric vehicle shapes, recognition failure or misjudgment is prone to occur. Although some solutions are combined with the elevator access control system for control, they are only based on static detection and only trigger alarms after the electric vehicle enters the elevator. They lack the ability to predict violations, resulting in delayed response and inability to effectively prevent electric vehicles from entering. In addition, the alarm triggering logic of the existing system is relatively simple, and it does not fully combine the intrusion time, contour ratio and residence time of the electric vehicle for multi-dimensional comprehensive judgment, which is easy to cause frequent alarms or omission of violations, making it difficult to achieve the ideal level of security prevention effect. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a monitoring system and an elevator for preventing an elevator from carrying an electric vehicle.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A monitoring system for preventing elevators from carrying electric vehicles comprises: The target detection module obtains the image edge gradient, morphological features and pixel contact features, calculates the gradient changes of adjacent pixels and establishes a global gradient map, analyzes the edge areas with strong continuity as candidate contours, calculates the curvature and screens the electric vehicle contour area to obtain the target contour candidate area; The contour matching module selects the rigid structure area according to the target contour candidate area, removes the edge area that conforms to the natural undulation characteristics of the human body according to the comparison between the electric vehicle parts and the human clothing characteristics, and obtains the structural matching contour area; The trajectory capture module sets a path continuity threshold based on the structure matching contour area, combines the displacement of key points of the human body and the trajectory change rate, and screens the stable area of ​​morphological change to obtain an independent contour area of ​​the electric vehicle; The early warning trigger module determines whether the electric vehicle enters the elevator door according to the independent contour area, contour proportion and residence time data of the electric vehicle, sets the alarm condition according to the intrusion time, calculates the priority and triggers the sound and light alarm to generate an intrusion alarm signal; The elevator control module executes elevator control, sets the elevator locking time, executes the stop operation instruction and uploads the intrusion record based on the intrusion alarm signal, and obtains the elevator blocking instruction.

[0007] As a further solution of the present invention, the target contour candidate area includes a strong continuity edge area, a global gradient mapping, and a candidate contour curvature calculation result; the structure matching contour area includes a rigid structure area, an electric vehicle component feature area, and a human body edge area after elimination; the electric vehicle independent contour area includes a morphological change stable area, a trajectory change rate screening area, and a path continuity threshold area; the intrusion alarm signal includes contour proportion data, stay time data, intrusion time threshold, and alarm priority; the elevator blocking instruction includes elevator locking time, stop operation instruction, and uploading intrusion record data.

[0008] As a further solution of the present invention, the target detection module includes: The edge gradient calculation submodule obtains the grayscale value of image pixels, calculates the grayscale change rate of adjacent pixels, determines the gradient direction and amplitude, selects the gradient amplitude pixels that meet the threshold, extracts the gradient mutation points, calculates the curvature, contrast change rate and edge direction consistency, and generates gradient mutation feature parameters; The global gradient mapping submodule calculates the distribution of pixel gradient directions based on the gradient mutation characteristic parameters, constructs a gradient direction histogram, calculates the mean gradient change of adjacent regions, generates a global gradient change matrix, sets dynamic thresholds for the pixels of the global gradient change matrix, screens the regions with prominent gradient changes, and generates a global gradient mapping matrix; The target contour screening submodule extracts the contour of the gradient change area based on the global gradient mapping matrix, calculates the boundary curvature and contrast change rate, screens the contours with stable curvature and prominent contrast changes, calculates edge continuity and pixel contact features, screens the contour areas that meet the morphological features, and obtains the target contour candidate area.

[0009] As a further solution of the present invention, the contour matching module includes: The target contour extraction submodule obtains the target contour candidate area, calculates the gradient change of the pixel area edge, selects the boundary pixels whose change meets the gradient threshold, connects the continuous boundary pixels to form a preliminary contour, calculates the curvature change rate of the preliminary contour, selects the contour segment whose curvature change rate is greater than the curvature threshold, and obtains the high curvature target contour area; The symmetry screening submodule calculates the symmetry offset of the pixel distribution inside the contour area based on the high curvature target contour area, screens the area where the offset is less than the set threshold, removes the non-rigid structure area where the offset exceeds the set threshold, calculates the offset mean of the morphological symmetry axis, and screens the area where the offset mean is less than the average threshold to obtain the rigid structure contour area; The structural matching optimization submodule detects the edge curvature change trend based on the rigid structure contour area, calculates the curvature change of the local area, screens the edge areas whose change is greater than the standard threshold, calculates the matching coefficient, screens the areas whose matching coefficient is greater than the reference threshold, and obtains the structural matching contour area.

[0010] As a further solution of the present invention, the calculation formula for screening the area where the offset is less than the set threshold is specifically: ; Calculate the mean offset of the morphological symmetry axis and filter the area where the mean offset is less than the average threshold; in, represents the mean value of the deviation of the symmetry axis of the pattern, Represents the number of pixels in the contour area. and Respectively represent Pixels in and The position on the axis, and Respectively represent pixels in the reference area and Position on the axis.

[0011] As a further solution of the present invention, the trajectory capture module includes: The trajectory rate analysis submodule obtains the displacement information of key points of the human body, calculates the displacement distance in adjacent time frames, screens the high change rate area according to the motion path continuity threshold, calls the trajectory change rate to calculate the time series distribution, and performs sequence analysis in combination with the time frame to obtain the trajectory change rate distribution data; The target area screening submodule detects the area with drastic rate changes in a short period of time based on the trajectory change rate distribution data, calculates its proportion in the global trajectory, analyzes the continuity of the time series, screens the area with stable motion trend, and obtains the stable motion area distribution data; The independent contour extraction submodule performs morphological analysis based on the stable motion area distribution data and the rigid structure data of the electric vehicle, extracts consistent structural feature areas, calculates regional boundary changes, screens stable morphological change areas, and establishes independent contour areas of the electric vehicle.

[0012] As a further solution of the present invention, the early warning triggering module includes: The contour detection submodule obtains the electric vehicle contour area data, detects the pixel distribution of the areas inside and outside the elevator door, extracts the contour features of the target area, calculates the contour area ratio, calls the elevator door status information, compares the contour ratio threshold, determines whether the electric vehicle is close to or partially enters the elevator door area, and generates the contour ratio parameter; The trajectory analysis submodule extracts the electric vehicle trajectory data based on the contour ratio parameter, calculates the dwell time near the elevator door, analyzes the trajectory deviation in combination with the motion trajectory, compares the dwell time threshold, determines whether there is a continuous intrusion trend, and obtains the trajectory dwell time parameter; The alarm trigger submodule analyzes whether the intrusion time exceeds the reference threshold based on the trajectory dwell time parameter and the elevator door status data, and determines whether the alarm trigger condition is met. If the alarm trigger condition is met, the sound and light alarm is activated to generate an intrusion alarm signal.

[0013] As a further solution of the present invention, the calculation formula for extracting the contour features of the target area is specifically: ; Calculate the contour ratio, Represents the contour ratio, Representative The contour pixel value of the pixel point, represent The absolute value of Representative The area of ​​the region corresponding to the pixel point is Represents the total number of pixels in the contour area. Represents the total area of ​​the elevator door area.

[0014] As a further solution of the present invention, the elevator control module includes: The alarm signal processing submodule obtains the intrusion alarm signal, detects the signal source and level, sets the locking parameters, calls the elevator control data, matches the elevator floor and operation status, calculates the deviation value between the elevator operation parameters and the locking parameters, screens the executable locking duration, and obtains the elevator locking duration data; The elevator operation control submodule compares the current operation status of the elevator based on the elevator locking time data, determines whether it is in the adjustable range, calculates the stop trigger condition, compares the current position of the elevator, the target floor and the dispatching information, selects the stop mode, and generates the elevator dispatch stop instruction; The prohibition instruction execution submodule detects the elevator response status, confirms the shutdown, collects the intrusion alarm signal and the shutdown time data, integrates the log record shutdown parameters, uploads them to the intrusion record database, and generates the elevator blocking instruction based on the elevator dispatch shutdown instruction.

[0015] Based on the same inventive concept, an elevator is also proposed to prevent the elevator from carrying electric vehicles, and the elevator is equipped with the above-mentioned monitoring system.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by fusing the edge gradient, morphological features and pixel contact features of the image, the target contour is accurately extracted to reduce environmental interference. The natural feature edges of the human body are eliminated by structural matching analysis to improve the accuracy of electric vehicle recognition. The trajectory capture combines the displacement of key points of the human body with the trajectory change rate, sets the path continuity threshold, stabilizes the target morphology, and avoids the impact of occlusion or angle changes on recognition. Based on the contour ratio and dwell time data, the alarm is triggered only when the violation is clear, reducing false alarm interference. The elevator control system links the alarm signal, dynamically sets the lock time to execute the stop operation command, and uploads the intrusion record to form traceable management and improve safety supervision capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a system flow chart of the present invention.

[0018] Figure 2 It is a flow chart of the submodules of the present invention.

[0019] Figure 3 Flowchart of the target detection module of the present invention.

[0020] Figure 4 Flowchart of the contour matching module of the present invention.

[0021] Figure 5 Flow chart of the trajectory capture module of the present invention.

[0022] Figure 6 This is a flow chart of the early warning trigger module of the present invention.

[0023] Figure 7 It is a flow chart of the elevator control module of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0025] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0026] See also Figure 1 A monitoring system for preventing elevators from carrying electric vehicles comprises: The target detection module obtains the image edge gradient, morphological features and pixel contact features, calculates the gradient changes of adjacent pixels, establishes a global gradient mapping, sets a dynamic change threshold, performs vector analysis on the edge area, screens the candidate areas of the electric vehicle contour, and combines the curvature and contrast change rate to obtain the candidate area of ​​the target contour.

[0027] The contour matching module analyzes the morphological symmetry of the pixel area according to the target contour candidate area, screens the rigid structure area, eliminates the edge area that conforms to the natural undulating characteristics of human clothing, and obtains the structural matching contour area.

[0028] The trajectory capture module combines the displacement information of key points of the human body, analyzes the trajectory change rate, sets the motion path continuity threshold, screens the target areas that change dramatically in a short period of time, eliminates the target areas that are not in accordance with the human motion rules, and combines the rigid structure data of the electric vehicle to select the stable morphological change area to obtain the independent contour area of ​​the electric vehicle.

[0029] The early warning trigger module is based on the independent contour area of ​​the electric vehicle and combines the alarm rules to calculate the contour ratio, trajectory dwell time and elevator door status data, determine whether the electric vehicle has entered the elevator door, set the alarm activation conditions according to the intrusion time, trigger the sound and light alarm, and obtain the intrusion alarm signal.

[0030] According to the intrusion alarm signal, the elevator control module calls the elevator control system data, sets the elevator lock time, executes the stop operation command, uploads the intrusion record, and obtains the elevator blocking command.

[0031] The target contour candidate area includes the strong continuity edge area, the global gradient mapping, and the candidate contour curvature calculation results; the structure matching contour area includes the rigid structure area, the electric vehicle component feature area, and the human body edge area after elimination; the electric vehicle independent contour area includes the morphological change stable area, the trajectory change rate screening area, and the path continuity threshold area; the intrusion alarm signal includes the contour proportion data, the dwell time data, the intrusion time threshold, and the alarm priority; the elevator blocking command includes the elevator locking time, the stop operation command, and the uploading of the intrusion record data.

[0032] See also Figure 3 and Figure 2 , the target detection module includes: The edge gradient calculation submodule obtains the grayscale value of image pixels, calculates the grayscale change rate of adjacent pixels, determines the gradient direction and amplitude, screens the gradient amplitude pixels that meet the threshold, extracts the gradient mutation points, calculates the curvature, contrast change rate and edge direction consistency, and generates gradient mutation feature parameters.

[0033] The step of obtaining the grayscale value of image pixels first requires extracting image data through corresponding image processing software or tools. Each pixel has a grayscale value, usually ranging from 0 to 255, where 0 represents black and 255 represents white. Next, the grayscale change rate of adjacent pixels is calculated, which can be achieved by a simple differential calculation method, such as calculating the grayscale difference of adjacent pixels in the horizontal, vertical or diagonal direction, expressed as: ΔI=I(x,y)-I(x+1,y), where I(x,y) represents the grayscale value at position (x,y) in the image. Using these differences, the magnitude of the grayscale change of adjacent pixels can be determined, and then according to a predetermined threshold (for example, set to 15, that is, when the grayscale difference of adjacent pixels exceeds 15, the area is considered to be an edge), the pixels with larger changes are screened out, and these pixels with larger changes represent the edge area in the image. Then, the mutation points of these edges are extracted, and the edge detection algorithm such as the Sobel operator can be used to further calculate the gradient amplitude and select the most prominent edge part. Next, we use a mathematical formula to evaluate the curvature of the edge and the prominence of the brightness change by combining the curvature and contrast change rate, and filter out qualified areas by setting a threshold (for example, curvature greater than 0.5, contrast change rate greater than 0.3). The curvature is calculated based on the curvature of the edge in the image. Usually, if the curvature value is greater than 0.5, it means that the edge change is more significant, so it is set as a reasonable threshold.

[0034] The calculation method of contrast change rate is: contrast change rate = (I max -I min ) / I avg , where I max and I min are the maximum and minimum grayscale values ​​within the contour area, I avg is the average grayscale value of the area. If the contrast change rate is greater than 0.3, the area will be considered to have a prominent contrast change. Finally, based on these features, the gradient mutation feature parameters are calculated and generated. These parameters include information such as the direction and intensity of the edge mutation, providing basic data for subsequent global gradient mapping.

[0035] The global gradient mapping submodule calculates the pixel gradient direction distribution based on the gradient mutation feature parameters, constructs the gradient direction histogram, calculates the mean gradient change of adjacent regions, generates the global gradient change matrix, sets dynamic thresholds for the pixels of the global gradient change matrix, screens the regions with prominent gradient changes, and generates the global gradient mapping matrix.

[0036] Based on the gradient mutation feature parameters, the pixel gradient direction distribution is constructed. This process aggregates the gradient direction data of each pixel into a direction histogram, which is similar to dividing the gradient direction of each pixel (such as the gradient direction calculated by the Sobel operator) into multiple intervals, each of which represents a certain angle range. For example, it is divided into 0°-45°, 45°-90°, etc., so that a histogram representing the gradient direction distribution of the image can be obtained. Next, the mean gradient change of adjacent regions can be calculated using the sliding window method. Assuming that the window size is 3x3 pixels, for each pixel in the window, its average gradient value is calculated and compared with the surrounding area. If the change exceeds the set dynamic threshold (such as set to 0.2), it is considered that the gradient change in this area is large. In this way, a global gradient change matrix is ​​obtained, which represents the gradient change of different regions in the entire image. Next, for each pixel in the global gradient change matrix, a dynamic threshold is set to filter out the area with prominent gradient changes. For example, a dynamic threshold can be set by calculating the change value of each position in the matrix and the standard deviation of the neighborhood. If the gradient change of a pixel exceeds the threshold (for example, set to 2 times the standard deviation), the pixel is marked as a prominent area. Finally, a global gradient mapping matrix is ​​generated, which shows all areas with significant gradient changes in the image, providing data support for subsequent target contour screening.

[0037] The target contour screening submodule extracts the contour of the gradient change area based on the global gradient mapping matrix, calculates the boundary curvature and contrast change rate, screens the contours with stable curvature and prominent contrast changes, calculates edge continuity and pixel contact features, screens the contour areas that meet the morphological characteristics, and obtains the target contour candidate area.

[0038] Based on the global gradient mapping matrix, the target contour screening submodule first extracts the contours of the area where the gradient changes are more obvious. This step can be completed by an image segmentation algorithm, such as Canny edge detection, through which the coordinates of all edge points in the image are extracted and connected into a contour. Next, the boundary curvature of the contour is calculated. The curvature calculation formula is: k=(d²y / dx²) / (1+(dy / dx)²)^(3 / 2), where dy / dx is the tangent slope of the contour, and d²y / dx² is the second-order derivative, which represents the curvature of the curve. If the calculated curvature value is greater than 0.5, it is considered that the contour has a large curvature. If the threshold is set to 0.5, it means that those areas with more significant edge mutations are selected. For example, if the curvature value in a certain contour area is greater than 0.5, the edge change in this area is obvious and meets the target screening conditions. Next, the contrast change rate of the contour is calculated, which is defined as contrast change rate=(I max -I min ) / I avg , where I maxand I min are the maximum and minimum grayscale values ​​within the contour area, I avg is the average gray value of the area. If the contrast change rate is greater than a preset threshold (for example, set to 0.3), the contour has a significant change in contrast. Next, those contours with smooth curvature and prominent contrast changes are screened out. These contours are more likely to be the edges of the target. Calculate the continuity of the edge, that is, by checking whether the contour is continuous, the distance threshold method is used to judge. If the distance between two adjacent edge points is less than 2 pixels, the edge is considered to be continuous. This setting is based on the edge characteristics of typical target images. When the gray values ​​of adjacent pixels change greatly, the pixel distance between them is usually small. Therefore, 2 pixels are set as a reasonable threshold, which means that these edges should belong to the same continuous part of the target. Finally, combined with the pixel contact feature, the contour area that meets the morphological characteristics is screened out. For example, through morphological operations such as corrosion and expansion, the target contour is further refined, and finally the target contour candidate area is obtained. These areas may contain target objects and serve as the basic data for subsequent recognition and tracking.

[0039] See also Figure 4 and Figure 2 , the contour matching module includes: The target contour extraction submodule obtains the target contour candidate area, calculates the gradient change of the pixel area edge, screens the boundary pixels whose change meets the gradient threshold, connects the continuous boundary pixels to form a preliminary contour, calculates the preliminary contour curvature change rate, screens the contour segments whose curvature change rate is greater than the curvature threshold, removes the low curvature boundary, and obtains the high curvature target contour area.

[0040] The target contour extraction submodule obtains the target contour candidate area, calculates the edge gradient change of the pixel area, screens the boundary pixels whose change meets the gradient threshold, connects the continuous boundary pixels to form a preliminary contour, calculates the preliminary contour curvature change rate, and screens the contour segments whose curvature change rate is greater than the curvature threshold to obtain the high curvature target contour area. First, the target contour extraction submodule performs edge detection on the input image and analyzes the edge features of the image by calculating the edge gradient change of the pixel area. In practical applications, considering the grayscale change of the image, the intensity of the image edge can be reflected by calculating the gradient value of each pixel point, for example, using the Sobel operator to perform edge detection on the image. The gradient change calculation formula is: ; in, is the gradient value, is the image gray value, and The images are and By setting a gradient threshold, we can filter out the boundary pixels whose change is greater than the threshold, and then obtain the boundary contour in the image. Next, we connect the continuous boundary pixels to form a preliminary contour, and further calculate the curvature change rate of the preliminary contour. The curvature calculation formula is: ; in, is the curvature, is the change in tangent angle, represent The absolute value of is the change in arc length. By setting the curvature threshold, the contour segments with a curvature change rate greater than the threshold are screened out, and finally a high curvature target contour area is formed. For example, assuming that the edge gradient change in the image is 10 and the curvature change rate is 0.5, by setting the gradient threshold to 8 and the curvature threshold to 0.3, the target contour area that meets the conditions is screened out. At this time, the output high curvature target contour area is an area with significant curvature change, which is convenient for subsequent target recognition and processing.

[0041] The symmetry screening submodule calculates the symmetry offset of the pixel distribution inside the contour area based on the high curvature target contour area, screens the areas where the offset is less than the standard threshold, removes the non-rigid structure areas where the offset exceeds the threshold, calculates the offset mean of the morphological symmetry axis, and screens the areas where the offset mean is less than the reference threshold to obtain the rigid structure contour area.

[0042] Filter the area where the offset sets the standard threshold, using the formula: ; Calculate the mean deviation of the morphological symmetry axis and select the area with the average reference threshold of the mean deviation; in, represents the mean value of the deviation of the symmetry axis of the pattern, Represents the number of pixels in the contour area. and Respectively represent Pixels in and The position on the axis, and Respectively represent pixels in the reference area and Position on the axis; The parameters in the formula are defined as follows: : The mean value of the offset of the morphological symmetry axis, which indicates the average degree of symmetry offset between the contour area and the reference area. It is obtained by statistically analyzing the offset of each pixel position.

[0043] : The number of pixels in the contour area. The contour area is segmented by image processing methods, and the number of pixels in the area is calculated.

[0044] and : No. The position coordinates of the pixel points on the x and y axes are obtained by extracting the coordinates of the pixels in the contour area.

[0045] and : No. The position coordinates of the pixel points on the x and y axes in the reference area. The reference area is the reference image area selected by comparison with the contour area, and its coordinates are obtained by the image registration method.

[0046] Quantification process: (Number of pixels): Extract the contour area through image processing algorithm and count the number of pixels in the area. Assuming there are 1000 pixels in the contour area, then =1000.

[0047] and (Coordinates of each pixel in the contour area): For the first pixel in the contour area, its position is determined by the image coordinate system. Assume that the coordinates of the first pixel are and .

[0048] and (Coordinates of each pixel in the reference area): The position of the pixel in the reference area is obtained by image registration of the reference area. Assume that the coordinates of the pixel in the reference area are and .

[0049] For the first pixel, calculate the offset of its position: ; ; Sum these two offsets and take the square root to get the offset of the pixel: ; Similarly, calculate the offsets of all 1000 pixels and find the average value. Assume that the sum of the offsets of all 1000 pixels is .

[0050] Divide the sum of all offsets by , we can get the mean value of the offset of the symmetry axis of the shape : ; Calculated =7, indicating that the average overall offset between the contour area and the reference area is 7 unit lengths. This value reflects the deviation of the symmetry of the contour area from the reference area. The larger the value, the more serious the offset, and the smaller the value, the better the symmetry of the contour area.

[0051] The structural matching optimization submodule detects the edge curvature change trend based on the rigid structure contour area, calculates the curvature change of the local area, screens the edge areas with a change greater than the set threshold, removes the flexible deformation edges with high change, calculates the matching coefficient, screens the areas with a matching coefficient greater than the average threshold, and obtains the structural matching contour area.

[0052] The structural matching optimization submodule detects the edge curvature change trend based on the rigid structure contour area, calculates the local area curvature change, screens the edge area with a change greater than the set threshold, calculates the matching coefficient, and screens the area with a matching coefficient greater than the average threshold to obtain the structural matching contour area. First, the structural matching optimization submodule analyzes the edge of the rigid structure contour area and detects the curvature change trend. In order to capture the curvature change of the local area, the calculation method of the local curvature change is adopted. The local curvature change can be obtained by calculating the curvature change between different points in the area. The calculation formula is: ; in, is the local curvature variation, and is the curvature value of two adjacent points. By setting a curvature change threshold, the edge area with a change greater than the threshold is screened out for further analysis and processing. Next, the matching degree of the area is determined by calculating the matching coefficient. The calculation formula of the matching coefficient is as follows: ; in, and For the two regions The curvature value at each position, is the number of pixels in the area. By calculating the matching coefficient and comparing it with the set average threshold, the area with a matching coefficient greater than the threshold is screened out, and finally the structure matching contour area is obtained. Assuming that the threshold of the matching coefficient is set to 0.8, when the matching coefficient of a certain area is 0.9, the area is the structure matching contour area. These areas show a strong consistency in practical applications and are suitable for further processing.

[0053] See also Figure 5 and Figure 2, the trajectory capture module includes: The trajectory rate analysis submodule obtains the displacement information of key points of the human body, calculates the displacement distance in adjacent time frames, screens the high change rate area according to the motion path continuity threshold, calls the trajectory change rate to calculate its time series distribution, combines multiple time frames for sequence analysis, and obtains the trajectory change rate distribution data; The trajectory rate analysis submodule calculates the displacement distance in adjacent time frames by obtaining the displacement information of key points of the human body, and screens out the high change rate area according to the continuity threshold of the motion path. First, the sensor equipment is used to capture the key point positions of the human body movement to form continuous time series data, and then the displacement of the key points in each time frame is calculated. The displacement formula is: ; in, and Time frames The horizontal and vertical coordinates of the key points are calculated by this formula. The displacement of the key points of the human body at each time point is calculated. Then, the displacement change rate between adjacent time frames is calculated: ; For example, assuming that between the first frame and the second frame, the displacement of the key point is 5cm and the time interval is 0.1 seconds, the change rate is 50cm / s. According to the preset motion path continuity threshold (such as 0.3m / s), the areas where the rate exceeds this threshold are screened out. These areas are the parts that may have sharp changes. By combining with the time series, the specific trajectory change rate distribution data is further determined.

[0054] The target area screening submodule detects the area with drastic rate changes in a short period of time based on the trajectory change rate distribution data, calculates its proportion in the global trajectory, analyzes the continuity of the time series, eliminates the area with non-human motion rules, screens the area with stable motion trend, and obtains the stable motion area distribution data.

[0055] The target area screening submodule detects areas with drastic rate changes in a short period of time based on the trajectory change rate distribution data, and calculates the proportion of these areas in the global trajectory. By comparing the change rates in adjacent time frames, if the rate change exceeds the set threshold (such as 10cm / s) in a short period of time, the area is determined to be a drastic rate change area. Further analyze the proportion of these areas in the global trajectory. For example, if the total length of the trajectory is 100m, and the rate of change in a certain area is drastic and its length is 5m, then the area accounts for 5%. Combined with the time series continuity analysis, it can be found that the movement trend of some areas is relatively stable, so they are screened as stable movement areas. The screening process of these stable areas determines that their rate changes are stable (the fluctuation of the rate of change is less than the set value, such as 1cm / s), and then obtains the distribution data of the stable movement area. For example, if the rate change of a certain area does not exceed the fluctuation range of 0.5cm / s within 1 second, the area can be considered stable.

[0056] The independent contour extraction submodule performs morphological analysis based on the stable motion area distribution data and the rigid structure data of the electric vehicle, extracts the consistent structural feature area, calculates the regional boundary changes, screens the stable morphological change area, establishes the independent contour morphology, and obtains the independent contour area of ​​the electric vehicle.

[0057] The independent contour extraction submodule performs morphological analysis based on the stable motion area distribution data and the rigid structure data of the electric vehicle, extracts the consistent structural feature area, calculates the boundary changes of the area, and screens out the area with stable morphological changes. Through in-depth analysis of the stable motion area data and combined with the rigid structure model of the electric vehicle (such as data based on geometric modeling), the degree of morphological change in the area is calculated. For example, if the boundary change of the area does not exceed the set value (such as 0.5mm) for a period of time, it means that the morphology of the area is stable. The overall morphology of the electric vehicle is fitted using a mathematical model to obtain the structural feature area. By comparing with the actual data, those areas with smaller boundary changes are screened out, which have higher stability. Finally, based on this analysis, the independent contour area of ​​the electric vehicle is constructed, and the independent contour area that meets the standards is identified through system identification and screening to ensure the stability and consistency of the electric vehicle structure.

[0058] See also Figure 6 and Figure 2 , the early warning trigger module includes: The contour detection submodule obtains the electric vehicle contour area data, detects the pixel distribution of the areas inside and outside the elevator door, extracts the contour features of the target area, calculates the contour area ratio, calls the elevator door status information, compares the contour ratio threshold, determines whether the electric vehicle is close to or partially enters the elevator door area, and generates the contour ratio parameters.

[0059] Extract the contour features of the target area using the formula: ; Calculate the contour ratio, Represents the contour ratio, Representative The contour pixel value of the pixel point, represent The absolute value of Representative The area of ​​the region corresponding to the pixel point is Represents the total number of pixels in the contour area. Represents the total area of ​​the elevator door area.

[0060] This value is obtained through image processing algorithms. Each pixel of the image is marked as a contour pixel (value 1) or a non-contour pixel (value 0) after thresholding. In this example, it is assumed that there are a total of 2000 pixels detected in the elevator door area, of which 1000 pixels belong to the electric car contour area. The value is 1. The remaining 1000 pixels are the background or the elevator door area, corresponding to The value is 0.

[0061] Each pixel of the electric vehicle outline occupies a certain area, which is usually converted using the resolution of the actual image. Assume that the resolution of the image is 1mm 2 / pixel, that is, the area of ​​each pixel is 1mm 2 In this case, all Each pixel with a value of 1 (i.e., 1000 contour pixels) has an area of ​​1 mm 2 ,therefore =1mm 2 .

[0062] In the above example, the number of pixels in the electric car contour area is 1000, so =1000.

[0063] The total area of ​​the elevator door can be calculated by actual measurement or the number of pixels in the image. Assume that the total area of ​​the elevator door is 2000mm 2 (corresponding to 2000 pixels), therefore, =2000mm 2 .

[0064] Summation part: For each pixel, and The product of gets the area contributed by the pixel. Since it is within the electric vehicle contour area, The value is 1, Value: 1mm 2 , so each pixel contributes 1mm 2For 1000 contour pixels, the sum is: ; Calculate the contour ratio: divide the sum of the areas occupied by all contour pixels by the total area of ​​the elevator door area , get the contour ratio : ; The result shows that the contour area of ​​the electric vehicle occupies half of the elevator door area, that is, 50%.

[0065] The trajectory analysis submodule extracts the electric vehicle trajectory data based on the contour ratio parameter, calculates the residence time near the elevator door, analyzes the trajectory deviation in combination with the motion trajectory, compares the residence time threshold, determines whether there is a continuous intrusion trend, and obtains the trajectory residence time parameter.

[0066] The execution process of the trajectory analysis submodule begins with the acquisition of the aforementioned contour ratio parameters. Through these parameters, the system can extract the trajectory data of the electric vehicle. The trajectory of the electric vehicle is obtained by tracking the moving path of the electric vehicle contour in the elevator door area. Specifically, the system needs to calculate the movement trajectory of the electric vehicle based on the position change of the electric vehicle in the video stream or the continuously shot images. In order to ensure the accuracy of the trajectory, the system tracks the movement path of the electric vehicle by comparing consecutive frames and applying the optical flow method (OpticalFlow) or the target detection algorithm. After the trajectory data is processed, the system will further calculate the residence time of the electric vehicle near the elevator door. If the electric vehicle stays in the elevator door area for more than the preset time threshold, the system will consider that it may have a continuous intrusion trend. For further judgment, the system analyzes the offset of the electric vehicle in combination with its movement trajectory, and calculates the trajectory offset by comparing the changes in the current trajectory with the previous trajectory. When the trajectory offset is greater than the set threshold, and the time the electric vehicle stays in the elevator door area exceeds the set residence time threshold, the system determines that the electric vehicle may have a trend of continuing to invade, obtains the trajectory residence time parameter, and prepares to further trigger the alarm.

[0067] The alarm trigger submodule analyzes whether the intrusion time exceeds the reference threshold based on the trajectory dwell time parameter and the elevator door status data, and determines whether the alarm trigger conditions are met. If the alarm trigger conditions are met, the sound and light alarm is activated to generate an intrusion alarm signal.

[0068] The execution process of the alarm trigger submodule starts with obtaining the trajectory dwell time parameter. The system first analyzes whether the intrusion time exceeds the reference threshold value in combination with the elevator door status data. The elevator door status is obtained through the connection with the elevator control system. When the elevator door is in the open state and the electric vehicle stays in the elevator door area for more than the set reference threshold value, the system will determine that the intrusion event meets the alarm trigger condition. The reference threshold value is set based on historical data and scenario requirements. Usually, this value is dynamically adjusted according to the elevator door opening time and the electric vehicle behavior pattern. For example, within 10 seconds of the elevator door opening, if the electric vehicle stays for more than 5 seconds, the system will determine that there may be danger and trigger an alarm. If this condition is met, the system will activate the sound and light alarm device and send out an intrusion alarm signal. This signal not only notifies the elevator control system to take action (such as automatically closing the elevator door), but also notifies the surrounding personnel through external alarm equipment, thereby achieving the purpose of preventing potential accidents.

[0069] See also Figure 7 and Figure 2 , the elevator control module includes: The alarm signal processing submodule obtains the intrusion alarm signal, detects the signal source and level, sets the locking parameters, calls the elevator control system data, matches the elevator floor and operating status, calculates the deviation between the elevator operating parameters and the locking parameters, screens the executable locking time, and obtains the elevator locking time data.

[0070] The alarm signal processing submodule obtains the intrusion alarm signal. First, the system receives the intrusion alarm signal from the sensor or monitoring device. The signal usually contains the time, location, signal strength and other information of the intrusion. Then, the system analyzes the source of the signal to determine whether it is an intrusion from the external environment or a false alarm from the internal device. The analysis process may involve the analysis of the signal frequency and duration. By matching with the preset intrusion mode data, it can be determined whether it meets the normal intrusion signal characteristics. Next, the system screens by judging the alarm level. Lower-level alarms may not trigger subsequent elevator control measures, while high-level alarms will cause the elevator's safety measures to start. After the level is confirmed, the system sets the locking parameters. These parameters involve the elevator's working mode, floor range and predetermined elevator operation status. The setting of the locking parameters needs to be adjusted according to the specific alarm level. For example, a high-level alarm may cause the system to enable the full elevator floor locking mode to prevent the elevator from continuing to run or staying on certain floors when encountering a threat. In this process, the system will combine the current operating status of the elevator, the preset safety rules and the actual floor information for calculation. The system compares the current position of the elevator with the target floor, calculates the current operating parameters of the elevator, and compares them with the set locking parameters to determine whether there is a deviation. The greater the deviation between the elevator operation parameters and the locking parameters, the more necessary the locking measures are, and the system further screens the executable locking duration. Assuming that the locking duration can be set between 10 and 30 minutes, the system determines the most appropriate duration based on the elevator's current position, floor, operating status, and alarm level, and finally calculates the elevator locking duration data. This data is used in the subsequent elevator operation control submodule to ensure that the elevator executes the necessary shutdown measures according to the set safety time to avoid being affected by external intrusion threats.

[0071] The elevator operation control submodule compares the elevator's current operating status based on the elevator lock duration data, determines whether it is in the adjustable range, calculates the stop trigger condition, compares the elevator's current position, target floor and scheduling information, screens the stop method, and generates the elevator scheduling stop instruction.

[0072] The elevator operation control submodule is based on the elevator lock duration data. First, the elevator control module analyzes the current operation status of the elevator to check whether the elevator is in an adjustable state range. Assuming that the elevator is currently running on a high floor, whether its current operation status is in a condition where it can stop, it is necessary to calculate the stop trigger condition. This condition may include the speed, acceleration, and distance between the operating floor and the target floor of the elevator. Specifically, the system calculates the relationship between the distance between the elevator and the target floor and the current speed and acceleration of the elevator through an algorithm, using the formula: ; in, is the distance the elevator needs to stop. is the initial velocity of the elevator, is the acceleration of the elevator, is the elevator travel time. When the speed and acceleration of the elevator meet the set standards, the elevator is considered to be in a stoppable state. On this basis, the information of the elevator's target floor and current position is input into the system, and the elevator's dispatch information is passed to the operation control submodule. The submodule selects the most appropriate stopping method based on these data. For example, if the elevator is far away from the target floor, it may adopt a deceleration stop strategy, while if it is at a closer floor, it may directly trigger an emergency stop. The screening of these stopping methods combines the elevator's running trajectory and the required time to determine the most appropriate execution time. In this process, the elevator control submodule also needs to analyze the elevator dispatch information to ensure that the appropriate stop mode is triggered according to the current operation status of the elevator, and finally generate the elevator dispatch stop instruction to ensure that the elevator can stop quickly and safely after the alarm is triggered.

[0073] The prohibition instruction execution submodule is based on the elevator dispatch stop instruction, detects the elevator response status, confirms the stop, collects the intrusion alarm signal and the stop time data, integrates the log record stop parameters, uploads them to the intrusion record database, and generates the elevator blocking instruction.

[0074] The prohibition instruction execution submodule is based on the elevator dispatching stop instruction. First, the system needs to detect the response status of the elevator and check whether the elevator is successfully stopped. This process may include confirmation through the communication signal between the control system and the elevator. If the elevator does not stop as expected, the system will repeatedly check its status and output an alarm. At this time, the elevator may execute backup stop measures until it is confirmed that the elevator is completely stopped. After the stop is successful, the system will collect and record data such as intrusion alarm signals and stop time to ensure that the stop process is fully logged, and these data will be organized into stop parameters. The stop data may include the start time of the stop, the duration of the stop, the strength of the alarm signal, etc., for subsequent analysis and tracing. After logging, the system uploads it to the intrusion record database to ensure the safe storage and audit tracking of the data. In addition, the prohibition instruction execution submodule will also generate an elevator blocking instruction to ensure that the elevator will not restart or continue to run when it is seriously threatened. Once the prohibition instruction is issued, the elevator will enter the prohibition mode according to the set conditions until it is manually released.

[0075] Based on the same inventive concept, the present invention also proposes an elevator that prevents the elevator from carrying electric vehicles, and the elevator is equipped with the above-mentioned monitoring system.

[0076] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A monitoring system for preventing elevators from carrying electric vehicles, characterized in that: The system includes: The target detection module obtains the image edge gradient, morphological features and pixel contact features, calculates the gradient changes of adjacent pixels and establishes a global gradient map, analyzes the edge areas with strong continuity as candidate contours, calculates the curvature and screens the electric vehicle contour area to obtain the target contour candidate area; The contour matching module selects the rigid structure area according to the target contour candidate area, removes the edge area that conforms to the natural undulation characteristics of the human body according to the comparison between the electric vehicle parts and the human clothing characteristics, and obtains the structural matching contour area; The trajectory capture module sets a path continuity threshold based on the structure matching contour area, combines the displacement of key points of the human body and the trajectory change rate, and screens the stable area of ​​morphological change to obtain an independent contour area of ​​the electric vehicle; The early warning trigger module determines whether the electric vehicle enters the elevator door according to the independent contour area, contour proportion and residence time data of the electric vehicle, sets the alarm condition according to the intrusion time, calculates the priority and triggers the sound and light alarm to generate an intrusion alarm signal; The elevator control module executes elevator control, sets the elevator locking time, executes the stop operation instruction and uploads the intrusion record based on the intrusion alarm signal, and obtains the elevator blocking instruction.

2. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The target contour candidate area includes a strong continuity edge area, a global gradient mapping, and a candidate contour curvature calculation result; the structure matching contour area includes a rigid structure area, an electric vehicle component feature area, and a human body edge area after elimination; the electric vehicle independent contour area includes a morphological change stability area, a trajectory change rate screening area, and a path continuity threshold area; the intrusion alarm signal includes contour proportion data, dwell time data, intrusion time threshold, and alarm priority; the elevator blocking instruction includes elevator locking time, stop operation instruction, and uploading of intrusion record data.

3. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The target detection module comprises: The edge gradient calculation submodule obtains the grayscale value of image pixels, calculates the grayscale change rate of adjacent pixels, determines the gradient direction and amplitude, selects the gradient amplitude pixels that meet the threshold, extracts the gradient mutation points, calculates the curvature, contrast change rate and edge direction consistency, and generates gradient mutation feature parameters; The global gradient mapping submodule calculates the distribution of pixel gradient directions based on the gradient mutation characteristic parameters, constructs a gradient direction histogram, calculates the mean gradient change of adjacent regions, generates a global gradient change matrix, sets dynamic thresholds for the pixels of the global gradient change matrix, screens the regions with prominent gradient changes, and generates a global gradient mapping matrix; The target contour screening submodule extracts the contour of the gradient change area based on the global gradient mapping matrix, calculates the boundary curvature and contrast change rate, screens the contours with stable curvature and prominent contrast changes, calculates edge continuity and pixel contact features, screens the contour areas that meet the morphological features, and obtains the target contour candidate area.

4. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The contour matching module comprises: The target contour extraction submodule obtains the target contour candidate area, calculates the gradient change of the pixel area edge, selects the boundary pixels whose change meets the gradient threshold, connects the continuous boundary pixels to form a preliminary contour, calculates the curvature change rate of the preliminary contour, selects the contour segment whose curvature change rate is greater than the curvature threshold, and obtains the high curvature target contour area; The symmetry screening submodule calculates the symmetry offset of the pixel distribution inside the contour area based on the high curvature target contour area, screens the area where the offset is less than the set threshold, removes the non-rigid structure area where the offset exceeds the set threshold, calculates the offset mean of the morphological symmetry axis, and screens the area where the offset mean is less than the average threshold to obtain the rigid structure contour area; The structural matching optimization submodule detects the edge curvature change trend based on the rigid structure contour area, calculates the curvature change of the local area, screens the edge areas whose change is greater than the standard threshold, calculates the matching coefficient, screens the areas whose matching coefficient is greater than the reference threshold, and obtains the structural matching contour area.

5. The monitoring system for preventing elevators from carrying electric vehicles according to claim 4, characterized in that: The calculation formula for filtering the area where the offset is less than the set threshold is specifically: ； Calculate the mean offset of the morphological symmetry axis and filter the area where the mean offset is less than the average threshold; in, represents the mean value of the deviation of the symmetry axis of the pattern, Represents the number of pixels in the contour area. and Respectively represent Pixels in and The position on the axis, and Respectively represent pixels in the reference area and Position on the axis.

6. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The trajectory capture module comprises: The trajectory rate analysis submodule obtains the displacement information of key points of the human body, calculates the displacement distance in adjacent time frames, screens the high change rate area according to the motion path continuity threshold, calls the trajectory change rate to calculate the time series distribution, and performs sequence analysis in combination with the time frame to obtain the trajectory change rate distribution data; The target area screening submodule detects the area with drastic rate changes in a short period of time based on the trajectory change rate distribution data, calculates its proportion in the global trajectory, analyzes the continuity of the time series, screens the area with stable motion trend, and obtains the stable motion area distribution data; The independent contour extraction submodule performs morphological analysis based on the stable motion area distribution data and the rigid structure data of the electric vehicle, extracts consistent structural feature areas, calculates regional boundary changes, screens stable morphological change areas, and establishes independent contour areas of the electric vehicle.

7. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The early warning trigger module includes: The contour detection submodule obtains the electric vehicle contour area data, detects the pixel distribution of the areas inside and outside the elevator door, extracts the contour features of the target area, calculates the contour area ratio, calls the elevator door status information, compares the contour ratio threshold, determines whether the electric vehicle is close to or partially enters the elevator door area, and generates the contour ratio parameter; The trajectory analysis submodule extracts the electric vehicle trajectory data based on the contour ratio parameter, calculates the dwell time near the elevator door, analyzes the trajectory deviation in combination with the motion trajectory, compares the dwell time threshold, determines whether there is a continuous intrusion trend, and obtains the trajectory dwell time parameter; The alarm trigger submodule analyzes whether the intrusion time exceeds the reference threshold based on the trajectory dwell time parameter and the elevator door status data, and determines whether the alarm trigger condition is met. If the alarm trigger condition is met, the sound and light alarm is activated to generate an intrusion alarm signal.

8. The monitoring system for preventing elevators from carrying electric vehicles according to claim 7, characterized in that: The calculation formula for extracting the contour features of the target area is specifically: ； Calculate the contour ratio, where Represents the contour ratio, Representative The contour pixel value of the pixel point, represent The absolute value of Representative The area of ​​the region corresponding to the pixel point is Represents the total number of pixels in the contour area. Represents the total area of ​​the elevator door area.

9. The monitoring system for preventing elevators from carrying electric vehicles according to claim 1, characterized in that: The elevator control module comprises: The alarm signal processing submodule obtains the intrusion alarm signal, detects the signal source and level, sets the locking parameters, calls the elevator control data, matches the elevator floor and operation status, calculates the deviation value between the elevator operation parameters and the locking parameters, screens the executable locking duration, and obtains the elevator locking duration data; The elevator operation control submodule compares the current operation status of the elevator based on the elevator locking time data, determines whether it is in the adjustable range, calculates the stop trigger condition, compares the current position of the elevator, the target floor and the dispatching information, selects the stop mode, and generates the elevator dispatch stop instruction; The prohibition instruction execution submodule detects the elevator response status, confirms the shutdown, collects the intrusion alarm signal and the shutdown time data, integrates the log record shutdown parameters, uploads them to the intrusion record database, and generates the elevator blocking instruction based on the elevator dispatch shutdown instruction.

10. An elevator that prevents the elevator from carrying electric vehicles, characterized in that: The elevator is equipped with the monitoring system according to any one of claims 1 to 9.

Images