Elevator risk control safety management system and method based on Internet of Things

Abstract

The invention discloses an elevator risk control safety management system and method based on the Internet of Things, and relates to the technical field of elevator safety risk control, and the elevator risk control safety management system comprises the steps of S1, data acquisition, S2, data transmission, S3, data processing, S4, model establishment, S5, foreign matter judgment, S6, countermeasures, S7, automatic measures, S8, self-calibration and S9, cyclic monitoring. The elevator shaft is scanned through the multiple three-dimensional laser scanning sensors to collect information, a model can be reconstructed in real time and compared with a standard model, foreign matter in the shaft can be found in time, potential risks such as collision and damage to safety parts caused by the foreign matter to elevator operation are effectively avoided, and elevator operation safety is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of elevator safety risk control, and in particular relates to an elevator risk control safety management system and method based on the Internet of Things. Background Art

[0002] With the increasing popularity of high-rise buildings, elevators have become an indispensable part of people's daily lives and work. Elevator safety directly impacts the safety of users' lives and property. However, traditional elevator management methods rely on manual inspections and regular maintenance, which can be time-consuming and inadequate to detect potential elevator failures and risks. Especially in high-rise buildings, elevator failures can have serious consequences.

[0003] When a foreign object falls into the elevator shaft, if it blocks a key position of the elevator shaft, the elevator may stop at the fault location, causing passengers to be trapped in the car, which will have adverse effects on the passengers' psychology and physical health. To address the above problems, the following solutions are proposed. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides an elevator risk control safety management method based on the Internet of Things, comprising: Step S1, data collection: using multiple three-dimensional laser scanning sensors to scan the interior of the elevator shaft to collect point cloud data; Step S2, data transmission: transmitting the collected point cloud data to the cloud server; Step S3, data processing: the received point cloud data is preprocessed in a data preprocessing module; Step S4, establishing a model: establishing an initial standard model; Step S5, foreign body determination: comparing the real-time point cloud data with the standard model through the data comparison module to determine whether there is a foreign body; Step S6, countermeasures: once an abnormal foreign object mark is detected in the hoistway, the cloud server immediately transmits the information to the elevator control room; Step S7, automatic measures: automatically taking corresponding safety measures according to the location of the foreign object; Step S8, self-calibration: regularly perform a comprehensive scan of the shaft during the elevator stop period and automatically update the reference model in the system; Step S9, cyclic monitoring: After completing the above-mentioned countermeasures, automatic measures and self-calibration operations, the process returns to the data collection stage to implement cyclic monitoring.

[0005] Preferably, the specific process of step S3, data processing is as follows: S31. Data cleaning: First perform mean filtering, determine the neighborhood for each point, and calculate the mean coordinates of the points in the neighborhood , where n is the number of points, If the distance between the point and the mean point exceeds the threshold, it is a noise point and is removed. Then, median filtering is performed to determine the neighborhood and sort the coordinate values ​​in the neighborhood. The odd-numbered points take the middle value, and the even-numbered points take the middle value. If the distance between the new point and the original point exceeds the threshold, the original point is a noise point and is removed. Then, according to the known elevator shaft If the point coordinates are out of range, they will be removed; S32, data sampling: first perform uniform sampling to determine Direction sampling interval ,according to ,in, is an integer, for Starting coordinates, conditional point selection, and then random sampling: set the sampling ratio ,generate Random number ,like Then keep the point.

[0006] S33, coordinate transformation: First, perform translation transformation: Assume that the origin of the scanning device coordinate system is in the global coordinate system coordinate ,point After transformation , then perform rotation transformation: let Axis rotation angle , there is a circle Axis rotation The rotation matrix , the point coordinates are multiplied by , followed by linear normalization: Coordinate normalization , 、 Similarly, the standard deviation is normalized at the end. coordinate , 、 Similarly; S34, area screening: First, screen based on the spatial range to determine the car operation area direction , 、 direction ,like and and The points are retained and then screened based on geometric features. The key features are identified using the geometric feature recognition algorithm, and the relevant points are retained based on the relationship between the point and the feature, such as the distance and angle; Preferably, the specific process of step S4, establishing the model is as follows: S41: Data preparation and space partitioning; S42: Normal vector calculation; S43: Equation solving and surface reconstruction.

[0007] Preferably, the S41 specifically includes: S411. Determine the spatial range of the point cloud data: Assume that the coordinates of the point cloud data points are , determine its position in three-dimensional space 、 、 Range of directions 、 、 ; S412, octree initialization and point partitioning: divide the space into sides with length Small cube, point exist Direction cube index , direction , direction , the total index can be obtained through Calculate, where and yes and The number of small cubes divided in the direction, so as to divide the points into the corresponding small cubes.

[0008] Preferably, the S42 specifically includes: S421, small cube point set selection: for each small cube with data points, select the point set therein; S422: Select three non-collinear points A in the point set ( )、B( )、C( ), calculate the vector =BA=( - , - , - ), =CA=( - , - , - ), normal vector = ,in , are unit vectors in the x, y, and z directions respectively.

[0009] Preferably, the S43 specifically includes: S431: Determine the equation based on the physical meaning of the point cloud data: ,in , is the function to be sought, Related to point cloud distribution and normal vector; S432: Divide the space into grids, grid points ( ), Direction grid spacing , for the second-order partial derivative in exist( ) , similarly for , for , bring in and From the equation: + + = , and the solution is The value of S433: According to the solution The value of the isosurface is D, which will satisfy =D's mesh nodes or boundary points are connected into triangular facets to construct the initial three-dimensional model.

[0010] Preferably, the S5 specifically includes: S51: Obtain the real-time reconstructed model using the same method, simultaneously obtain the corresponding reference points and direction axes of the reconstructed model and the standard model, and calculate the coordinate difference between the reference points of the reconstructed model and the standard model; S52: Calculate the coordinate difference between the bottom center of the reconstructed model and the bottom center of the standard model (0, 0, 0); S53: Rotate the reconstructed model using the rotation matrix to align the long axis directions, and measure the dimensions of the standard model in all directions; S54: Calculate the scale and scale the coordinates of the reconstructed model; S55: Then, the geometric features of the vertex information of the reconstructed model are extracted, and then feature matching is performed; S56: For each vertex of the reconstructed model, calculate the Euclidean distance to all vertices of the standard model: , take the vertex corresponding to the minimum distance to establish a match, for the matching point pair , calculate the Euclidean distance , calculate the mean distance ( is the number of matching points), variance and maximum value , check vertex matching; S57: If a large number of vertices cannot be matched or the matching distance is too large, it is marked as a foreign object.

[0011] Preferably, an elevator risk control and safety management system based on the Internet of Things includes an elevator shaft, a cloud server, an elevator control room, and a maintenance worker's mobile terminal; The elevator shaft includes a processor, a GPS, a three-dimensional laser scanning sensor, an elevator brake, a rotary encoder, and a first wireless communication module; The cloud server includes a second wireless communication module, a data comparison module, a database and a self-checking module; The elevator control room includes a third wireless communication module and a display screen; The processor is bidirectionally connected to the GPS, the output end of the three-dimensional laser scanning sensor is connected to the input end of the processor, the output end of the processor is electrically connected to the input end of the elevator brake, the output end of the rotary encoder is connected to the input end of the processor, the processor is bidirectionally connected to the first wireless communication module, the first wireless communication module is bidirectionally connected to the second wireless communication module, the second wireless communication module is bidirectionally connected to the data comparison module, the data preprocessing module and the self-calibration module respectively, the data preprocessing module is bidirectionally connected to the model reconstruction module, the model reconstruction module is bidirectionally connected to the database, the data comparison module is bidirectionally connected to the database, the database is bidirectionally connected to the self-calibration module, the second wireless communication module is bidirectionally connected to the mobile terminal and the third wireless communication module respectively, the output end of the third wireless communication module is unidirectionally connected to the input end of the display screen, the third wireless communication module is bidirectionally connected to the first wireless communication module, and the mobile terminal is bidirectionally connected to the first wireless communication module.

[0012] Preferably, the three-dimensional laser scanning sensor is a device that uses laser technology to measure three-dimensional spatial information of an object, and is used to scan and obtain point cloud data in the elevator shaft; The elevator brake is used to control the running speed and stop of the elevator; The rotary encoder is a sensor that converts rotational displacement into a series of digital pulse signals. The frequency of the pulse signal is proportional to the rotational speed of the rotating part. By counting and analyzing the pulse signals, the speed information of the rotating part is obtained, which is used to detect the moving speed of the elevator.

[0013] Preferably, the maintenance worker's mobile terminal is used to receive information sent by the cloud server.

[0014] The present invention has the following beneficial effects: 1. The present invention uses multiple three-dimensional laser scanning sensors to scan the elevator shaft and collect information. It can reconstruct the model in real time and compare it with the standard model, so as to promptly detect foreign objects in the shaft and effectively avoid potential dangers caused by foreign objects to elevator operation, such as collisions and damage to safety components, thereby ensuring the safety of elevator operation.

[0015] 2. Once a foreign object is detected, the system can automatically take corresponding safety measures based on the location of the foreign object. If the foreign object is near the elevator car's running path and may cause a collision, the elevator brake can be activated to immediately stop the elevator or dock it at the nearest safe floor. If there is potential damage to safety components, the elevator speed can be adjusted. At the same time, a rotary encoder is used to monitor the speed in real time to avoid further expansion of risks.

[0016] 3. The present invention takes into account the possible slight deformation of the elevator shaft during long-term use, the installation or maintenance of new equipment, etc. The system regularly scans the shaft in full when the elevator is stopped, compares it with the initial standard model, and automatically updates the reference model to ensure the long-term accuracy of foreign object monitoring.

[0017] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is the internal system framework diagram of the present invention; Figure 2 It is a step flow framework diagram of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] See also Figure 2 As shown, the present invention is an elevator risk control safety management method based on the Internet of Things, comprising: Step S1, data collection: using multiple three-dimensional laser scanning sensors to scan the interior of the elevator shaft to collect point cloud data; Step S2, data transmission: transmitting the collected point cloud data to the cloud server; Step S3, data processing: the received point cloud data is preprocessed in a data preprocessing module; Step S4, establishing a model: establishing an initial standard model; Step S5, foreign body determination: comparing the real-time point cloud data with the standard model through the data comparison module to determine whether there is a foreign body; Step S6, countermeasures: once an abnormal foreign object mark is detected in the hoistway, the cloud server immediately transmits the information to the elevator control room; Step S7, automatic measures: automatically taking corresponding safety measures according to the location of the foreign object; Step S8, self-calibration: regularly perform a comprehensive scan of the shaft during the elevator stop period and automatically update the reference model in the system; Step S9, cyclic monitoring: After completing the above-mentioned countermeasures, automatic measures and self-calibration operations, the process returns to the data collection stage to implement cyclic monitoring.

[0022] The specific process of step S3, data processing is as follows: S31. Data cleaning: First perform mean filtering, determine the neighborhood for each point, and calculate the mean coordinates of the points in the neighborhood , where n is the number of points, If the distance between the point and the mean point exceeds the threshold, it is a noise point and is removed. Then, median filtering is performed to determine the neighborhood and sort the coordinate values ​​in the neighborhood. The odd-numbered points take the middle value, and the even-numbered points take the middle value. If the distance between the new point and the original point exceeds the threshold, the original point is a noise point and is removed. Then, according to the known elevator shaft If the point coordinates are out of range, they will be removed; S32, data sampling: first perform uniform sampling to determine Direction sampling interval ,according to ( is an integer, for Starting coordinates) and other conditions to select points, and then perform random sampling: set the sampling ratio ,generate Random number ,like Then keep the point.

[0023] S33, coordinate transformation: First, perform translation transformation: Assume that the origin of the scanning device coordinate system is in the global coordinate system coordinate ,point After transformation , then perform rotation transformation: let Axis rotation angle , there is a circle Axis rotation The rotation matrix , the point coordinates are multiplied by , followed by linear normalization: Coordinate normalization , 、 Similarly, the standard deviation is normalized at the end. coordinate , 、 Similarly; S34, area screening: First, screen based on the spatial range to determine the car operation area direction , 、 direction ,like and and The points are retained and then screened based on geometric features. The key features are identified using the geometric feature recognition algorithm, and the relevant points are retained based on the relationship between the point and the feature, such as the distance and angle; Step S4: The specific process of establishing the model is as follows: S41: Data preparation and space partitioning; S42: Normal vector calculation; S43: Equation solving and surface reconstruction.

[0024] 2. An elevator risk control safety management method based on the Internet of Things according to claim 2, characterized in that S41 specifically includes: S411. Determine the spatial range of the point cloud data: Assume that the coordinates of the point cloud data points are , determine its position in three-dimensional space 、 、 Range of directions 、 、 ; S412, octree initialization and point partitioning: divide the space into sides with length Small cube, point exist Direction cube index , direction , direction , the total index can be obtained through Calculate, where and yes and The number of small cubes divided in the direction, so as to divide the points into the corresponding small cubes.

[0025] S42 specifically includes: S421, small cube point set selection: for each small cube with data points, select the point set therein; S422: Select three non-collinear points A in the point set ( )、B( )、C( ), calculate the vector =BA=( - , - , - ), =CA=( - , - , - ), normal vector = ,in , are unit vectors in the x, y, and z directions respectively.

[0026] S43 specifically includes: S431: Determine the equation based on the physical meaning of the point cloud data: ,in , is the function to be sought, Related to point cloud distribution and normal vector; S432: Divide the space into grids, grid points ( ), Direction grid spacing , for the second-order partial derivative in exist( ) , similarly for , for , bring in and From the equation: + + = , and the solution is The value of S433: According to the solution The value of the isosurface is D, which will satisfy =D's mesh nodes or boundary points are connected into triangular facets to construct the initial three-dimensional model.

[0027] S5 specifically includes: S51: Obtain the real-time reconstructed model using the same method, simultaneously obtain the corresponding reference points and direction axes of the reconstructed model and the standard model, and calculate the coordinate difference between the reference points of the reconstructed model and the standard model; S52: Calculate the coordinate difference between the bottom center of the reconstructed model and the bottom center of the standard model (0, 0, 0); S53: Rotate the reconstructed model using the rotation matrix to align the long axis directions, and measure the dimensions of the standard model in all directions; S54: Calculate the scale and scale the coordinates of the reconstructed model; S55: Then, the geometric features of the vertex information of the reconstructed model are extracted, and then feature matching is performed; S56: For each vertex of the reconstructed model, calculate the Euclidean distance to all vertices of the standard model: , take the vertex corresponding to the minimum distance to establish a match, for the matching point pair , calculate the Euclidean distance , calculate the mean distance ( is the number of matching points), variance and maximum value , check vertex matching; S57: If a large number of vertices cannot be matched or the matching distance is too large, it is marked as a foreign object.

[0028] See also Figure 1 As shown, an elevator risk control and safety management system based on the Internet of Things includes an elevator shaft, a cloud server, an elevator control room, and a maintenance worker's mobile terminal; The elevator shaft includes a processor, a GPS, a three-dimensional laser scanning sensor, an elevator brake, a rotary encoder, and a first wireless communication module; The cloud server includes a second wireless communication module, a data comparison module, a database and a self-correction module; The elevator control room includes a third wireless communication module and a display screen; The processor is bidirectionally connected to the GPS, the output end of the three-dimensional laser scanning sensor is connected to the input end of the processor, the output end of the processor is electrically connected to the input end of the elevator brake, the output end of the rotary encoder is connected to the input end of the processor, the processor is bidirectionally connected to the first wireless communication module, the first wireless communication module is bidirectionally connected to the second wireless communication module, the second wireless communication module is bidirectionally connected to the data comparison module, the data preprocessing module and the self-calibration module respectively, the data preprocessing module is bidirectionally connected to the model reconstruction module, the model reconstruction module is bidirectionally connected to the database, the data comparison module is bidirectionally connected to the database, the database is bidirectionally connected to the self-calibration module, the second wireless communication module is bidirectionally connected to the mobile terminal and the third wireless communication module respectively, the output end of the third wireless communication module is unidirectionally connected to the input end of the display screen, the third wireless communication module is bidirectionally connected to the first wireless communication module, and the mobile terminal is bidirectionally connected to the first wireless communication module.

[0029] A 3D laser scanning sensor is a device that uses laser technology to measure the three-dimensional spatial information of an object. It is used to scan and obtain point cloud data in the elevator shaft. Elevator brakes are used to control the speed and stop of the elevator; A rotary encoder is a sensor that converts rotational displacement into a series of digital pulse signals. The frequency of the pulse signal is proportional to the rotational speed of the rotating part. By counting and analyzing the pulse signals, the speed information of the rotating part is obtained, which is used to detect the moving speed of the elevator.

[0030] The maintenance worker's mobile terminal is used to receive information sent by the cloud server.

[0031] A specific application of this embodiment is: Step 1: Data collection: Multiple 3D laser scanning sensors installed at key locations at the top, middle, and bottom of the elevator shaft are activated simultaneously. The sensors continuously emit laser beams, performing real-time, high-precision scans of the entire shaft space. The laser beams form a dense scan point cloud data within the shaft. A 3D laser scanning sensor is a device that uses laser technology to measure the three-dimensional spatial information of an object. Step 2: Data transmission: Multiple 3D laser scanning sensors integrate the scanned point cloud data through a processor and transmit it to a second communication module in a cloud server via a first communication module. The collected point cloud data is then transmitted to the cloud server in real time. The first wireless communication module and subsequent communication modules can be any of a Wi-Fi module, a Bluetooth module, a ZigBee module, a 4G / 5G module, or a narrowband Internet of Things (NB-IoT) module. Step 3: Data processing: The second wireless communication module transmits the received point cloud data to the data preprocessing module for data preprocessing, specifically: S31. Data cleaning: Remove noise points and mean filter: For each point in the point cloud data, determine a spherical neighborhood with each point as the center. The radius is set according to the point cloud density and noise situation. Calculate the average value of the coordinates of all points in the neighborhood. Set the point There are n points in the neighborhood of ( ), then within the neighborhood Average value of coordinates , similarly calculated and , if point P is equal to the mean point If the distance exceeds the specified threshold, P is considered to be a noise point and is removed; Remove noise points, median filter: Similarly, determine the spherical neighborhood with each point as the center, and calculate the coordinate values ​​of all points in the neighborhood according to Sort the directions and take the middle value after sorting as the filtered coordinate value. For example, within the neighborhood The coordinate values ​​are sorted from small to large. , if n is an odd number, then the median is ; If it is an even number, the median can be taken ,for and Similarly, the filtered coordinate values ​​are combined into a new point. If the distance difference between the new point and the original point is too large and exceeds the set threshold, the original point is considered a noise point and removed. Eliminate invalid data based on coordinate range judgment: Determine a reasonable coordinate range based on the known size of the elevator shaft and the scanning area. For example, if the elevator shaft is known to be The range of directions is , The direction is , The direction is , for each point in the point cloud data ,if , , , then the point is considered to be invalid data and will be removed; Eliminate invalid data based on the rationality of data values: According to the characteristics of the scanning equipment and past experience, set a reasonable range of coordinate values, for example, the measurement accuracy of the scanning equipment is within Direction , the normal measurement value should be within a certain range. If the coordinate value of a point exceeds this reasonable range, such as a point If the coordinate value is much larger than the maximum value that the scanning device can measure, the point will be judged as invalid data and discarded; S32, data sampling: Uniform sampling: Determine the sampling interval based on the spatial distribution of point cloud data. For example, Set the sampling interval for each direction , starting from the starting position of the point cloud data, select points according to the sampling interval. For example, direction, point of view Coordinates satisfy ( is an integer, for Starting coordinates), the point is selected. and The same applies to direction; Random sampling: set the sampling ratio ( ), represents the ratio of points to be retained, for each point in the point cloud data, generate A random number between ,if , then keep the point; otherwise, remove the point; S33. Coordinate transformation: Coordinate system translation transformation: Determine the translation relationship between the scanning device coordinate system and the target global coordinate system (such as the coordinate system with the center of the bottom of the elevator shaft as the origin). Assume that the coordinates of the origin of the scanning device coordinate system in the global coordinate system are , for each point in the point cloud data , the coordinates in the global coordinate system Calculated by the following formula: ; Coordinate system rotation transformation: Determine the rotation angle of the scanning device coordinate system relative to the global coordinate system by measuring or knowing the device installation angle and other information. For example, The rotation angles in the axial direction are , use the rotation matrix to transform the coordinates. Axis rotation Angle rotation matrix for: , similarly we can get around Axis rotation Angle rotation matrix He Rao Axis rotation Angle rotation matrix , first multiply the point coordinates by , and then multiply by , and finally multiply by , get the coordinates after rotation in the global coordinate system; Data normalization, linear normalization: for point cloud data Coordinates, find the minimum and maximum value , for each point Coordinates , through the formula Normalize it, Coordinates are mapped to interval, for and The coordinates are normalized in the same way; Data normalization, standard deviation normalization: calculating point cloud data Mean of coordinates and standard deviation , for each point Coordinates , through the formula Normalize so that after normalization The coordinates of have the characteristics of mean 0 and standard deviation 1. and The coordinates are calculated in the same way; S34, Regional screening: Extract key areas - Screening based on spatial scope: Determine the spatial scope of key areas based on the geometric structure of the elevator shaft and the focus of attention. For example, for the car operation area, determine the Direction from the lowest operating position of the car To the highest operating position ,exist and The direction is based on the cross-sectional dimensions of the car and appropriately extended to a certain range , for each point in the point cloud data ,if and and , then the point is retained, otherwise, the point is removed; Extract key areas based on geometric feature screening: Using some geometric feature recognition algorithms, surface features of objects related to elevator operation safety are identified. For example, the position of the guide rail is determined by identifying straight line features, and the position of the shaft wall is determined by identifying plane features. Point cloud data points related to these key geometric features are retained, while other irrelevant points (such as points on decorative structures on the shaft top) are eliminated. The distance and angle between the point and the geometric features are calculated to determine whether the point belongs to the key area. Step 4: Build a model and establish an initial standard model; For example, a simple geometric object, such as a partially concave cube, is 3D reconstructed by obtaining point cloud data of the object's surface using a 3D laser scanning sensor. After the collected point cloud data is preliminarily processed, only the points in some key areas of the object surface are retained, a total of 1000 points. The coordinates of these points are given in the form of three-dimensional Cartesian coordinates. For example, the coordinates of point A are ( ) = (1.2, 2.5, 3.1), the coordinates of point B are ( ) = (1.5, 2.2, 2.8), the coordinates of point C are ( ) = (1.4, 2.3, 2.9), and so on for all 1000 points; To build an octree structure, we first determine the range of the entire three-dimensional space. Assume that the approximate spatial range of the point cloud data is determined as follows: the x direction is from 0 to 5, the y direction is from 0 to 5, and the y direction is from 0 to 5. This spatial range is divided into small cubes of different levels (octree structure). For example, the initial division is into small cubes with a side length of 1. Taking point A (1.2, 2.5, 3.1) as an example, the index of the small cube where it is located is calculated as follows: In the x direction, the point is located in the first small cube (because ), indexed as =1; In the y direction, the second small cube (because ), indexed as =2; In the z direction, the third small cube (because ), indexed as =3; Therefore, the overall index of the small cube where point A is located can be represented by some encoding method, such as simply combining the indexes of the three directions to obtain a unique index value. Here, we assume that the combination method is = +10× +100× =1+10×2+100×3=321. In this way, all 1000 points are divided into corresponding small cubes to complete the preliminary construction of the octree structure. For each small cube, find the point cloud data points contained in it (search based on the small cube index). Take the small cube containing point A (index 321) as an example. Assume that the cube also contains another 4 points, which are recorded as B, C, D, and E respectively. To calculate the normal vector of the plane composed of these points, a common method is to use vector cross product to calculate. First, calculate two vectors: vector =BA=( - , - , - )=(1.5-1.2,2.2-2.5,2.8-3.1)=(0.3,-0.3,-0.3); vector =CA=( - , - , - )=(1.4-1.2,2.3-2.5,2.9-3.1)=(0.2,-0.2,-0.2); Then calculate the normal vector by vector cross product : = × , according to the vector cross product formula: = ,in , are the unit vectors in the x, y, and z directions respectively, Expanding the determinant yields: = ((-0.3)×(-0.2)-(-0.3)×(-0.2))- (0.3×(-0.2)-0.2×(-0.3))+ (0.3×(-0.2)-0.2×(-0.3))= ; The normal vector calculated here is zero because the selected points are special. In actual calculations, a more appropriate combination of points is usually selected to calculate a non-zero normal vector. According to this method, the corresponding normal vector is calculated for the point cloud data in each small cube, and these normal vectors are used as boundary conditions for subsequent equations. Then solve the equation to get a scalar function, in three-dimensional space, using the formula: ,in is the Laplace operator, in the Cartesian coordinate system, , is an unknown function to be solved (in the context of 3D reconstruction, it can be understood as a function related to the surface generated by point cloud data). is a known source function that describes the distribution of sources in space that generate "fields", such as electric fields and gravitational fields; According to the solution value, select the appropriate equivalent surface value D, and set all =D's mesh nodes or small cube boundary points are connected to form triangular facets or other suitable geometric representations, thereby constructing a reconstructed three-dimensional model. In this way, an initial standard model of an elevator shaft is first established; Step 5: Foreign body judgment: The above method can also be used to build a reconstruction model in real time. The data comparison module compares the reconstruction model of the real-time point cloud data with the standard model to determine whether there is a foreign body. Specifically: Step S441: Obtain the corresponding reference points and direction axes of the reconstructed model and the standard model, and calculate the coordinate difference between the reference points of the reconstructed model and the standard model. For the standard model, the bottom center coordinates are clearly defined as (0, 0, 0) in its design data. For the reconstructed model, the bottom center is approximately determined by calculating the center of gravity of the point cloud data in 3D space. Step S442: Calculate the coordinate difference between the bottom center of the reconstructed model and the bottom center of the standard model (0, 0, 0), translate the reconstructed model so that the reference points coincide, analyze the directional features such as the long axis of the model, and calculate the rotation angle; Step S443: Rotation adjustment is performed by analyzing the long axis direction of the model. For the standard elevator shaft model, its long axis direction is specified to be along the vertical direction of the elevator shaft. For the reconstructed model, the main direction of its point cloud data on the horizontal plane is calculated, and the rotation angle is calculated based on the difference in the long axis direction. For example, if the long axis of the standard model is at an angle of 30° to the X-axis, while the long axis of the reconstructed model is at an angle of 40° to the X-axis, the reconstructed model needs to be rotated 10° around the Z-axis to align the long axis directions. Measure the dimensions of the standard model in the X, Y, and Z directions. For example, if the standard model is 10 meters high (in the Z direction), 3 meters wide (in the X direction), and 3 meters deep (in the Y direction), calculate the actual dimensions of the reconstructed model in the corresponding directions. Assuming the reconstructed model is 9.8 meters high, 2.95 meters wide, and 2.9 meters deep, calculate the scaling ratio in each direction. The scaling ratio in the Z direction is 10 / 9.8, the scaling ratio in the X direction is 3 / 2.95, and the scaling ratio in the Y direction is 3 / 2.9. Step S444: Calculate the ratio of the corresponding directional dimensions of the reconstructed model to the standard model, and scale the coordinates of the reconstructed model; Step S445: The reconstructed model is then subjected to geometric feature extraction from vertex information, followed by feature matching. For a standard model, a vertex list and its coordinates are directly obtained based on its vertex information. For example, the elevator shaft of a standard model has defined vertices at the intersections and corners of each floor. For the reconstructed model, vertex information is extracted from the model data structure reconstructed from the point cloud data. For example, for a model represented by triangular facets, the coordinates of the three vertices of each triangular facet are extracted. Step S446: For the point pairs that have established matching relationships (reconstructed model points and standard model points), calculate the Euclidean distance between each pair of matching points, for example, the matching point pair ,distance , calculate the mean distance of all matching points , add up all the distance values ​​and divide them by the total number of matching points to get the mean. For example, if there are 1000 matching points and the total distance is 10, then Meters, then calculate the variance of the distance, first calculate the square of the difference between each distance value and the mean, add these squared values, and then divide by the total number of matching points to get the variance For example, the calculated variance is 0.0001 square meters, and then find the maximum distance , traverse all matching point distance values ​​and record the maximum value, for example, the maximum distance value is 0.05 meters; Step S47: Check the vertex matching situation. If a large number of vertices cannot be matched or the matching distance is too large, mark them as foreign objects. For example, if the matching distance of more than 10% of the vertices exceeds 2 times the average distance, there is a problem.

[0032] Step 5: Countermeasures: Once an abnormal foreign object mark is detected in the hoistway, the cloud server immediately transmits the information to the third communication module in the elevator control room. A prominent alarm message pops up on the display screen in the elevator control room, indicating the location, size, and possible impact range of the foreign object. A detailed alarm notification, such as a screenshot of the 3D hoistway model with the foreign object location highlighted, is also sent to the mobile terminal of the elevator maintenance personnel. Step 6: Automatic measures. Based on the location of the foreign object and the potential threat to the elevator's operation, appropriate safety measures are automatically taken. If the foreign object is near the elevator car's path and could cause a collision, the cloud server will activate the elevator brake through the processor, causing the elevator to stop immediately or park at the nearest safe floor. If the foreign object poses a potential risk to safety components such as guide rails and ropes, the elevator's speed will be adjusted using the brakes. The rotary encoder measures the elevator's speed in real time to avoid further risks and prioritize maintenance personnel for cleaning and repair. Step 7: Self-calibration: The elevator shaft may experience minor deformation during long-term use, installation of new equipment, or maintenance work. The system has a self-calibration function. It regularly scans the shaft during elevator downtime and compares it with the initial standard model. The reference model in the system is automatically updated to ensure the accuracy of foreign object detection. Step 8: Circular monitoring. After completing the above-mentioned countermeasures, automatic measures, and self-calibration operations, the process returns to the data collection stage and continues the next round of real-time monitoring of foreign objects in the three-dimensional space of the elevator shaft, realizing cyclic monitoring and continuously ensuring the safe operation of the elevator.

[0033] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0034] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An elevator risk control safety management method based on the Internet of Things, characterized by: Management methods include: Step S1, data collection: using multiple three-dimensional laser scanning sensors to scan the inside of the elevator shaft to collect point cloud data; Step S2, data transmission: transmitting the collected point cloud data to the cloud server; Step S3, data processing: the received point cloud data is preprocessed in a data preprocessing module; Step S4, establishing a model: establishing an initial standard model; Step S5, foreign body determination: comparing the real-time point cloud data with the standard model through the data comparison module to determine whether there is a foreign body; Step S6, countermeasures: once an abnormal foreign body mark is detected in the hoistway, the cloud server immediately transmits the information to the elevator control room; Step S7, automatic measures: automatically taking corresponding safety measures according to the location of the foreign object; Step S8, self-calibration: regularly perform a comprehensive scan of the shaft during the period when the elevator stops running, and automatically update the reference model in the system; Step S9, cyclic monitoring: After completing the above-mentioned countermeasures, automatic measures and self-calibration operations, the process returns to the data collection stage to implement cyclic monitoring.

2. The elevator risk control safety management method based on the Internet of Things according to claim 1 is characterized in that: The specific process of step S3, data processing is as follows: S31, data cleaning: first perform mean filtering, determine the neighborhood for each point, and calculate the mean coordinates of the points in the neighborhood , where n is the number of points, If the distance between the point and the mean point exceeds the threshold, it is a noise point and is removed. Median filtering is performed to determine the neighborhood. The coordinate values ​​in the neighborhood are sorted. The odd-numbered points take the middle value, and the even-numbered points take the If the distance between the new point and the original point exceeds the threshold, the original point is a noise point and is removed. If the point coordinates are out of range, they will be removed; S32, data sampling: uniform sampling, determine Direction sampling interval ,according to , is an integer, for Select the starting coordinates, perform random sampling, and set the sampling ratio ,generate Random Number ,like , then keep the point; S33, coordinate conversion: perform translation transformation: assume that the origin of the scanning device coordinate system is in the global coordinate system ,point After transformation , and then perform rotation transformation, let Axis rotation angle , there is a winding Axis rotation The rotation matrix , the point coordinates are multiplied by , followed by linear normalization; in, The coordinate normalization formula is: , , Similarly, the standard deviation is normalized. The coordinates are , , Similarly; S34, area screening: screening based on the spatial range to determine the car operation area direction , , direction ,like and and Then retain the points, and then screen them based on geometric features, use the geometric feature recognition algorithm to identify key features, and retain related points based on the distance and angle relationship between the points and the features.

3. The elevator risk control safety management method based on the Internet of Things according to claim 1 is characterized in that: The specific process of step S4, establishing the model is as follows: S41: Data preparation and space partitioning; S42: Normal vector calculation; S43: Equation solving and surface reconstruction.

4. The elevator risk control safety management method based on the Internet of Things according to claim 2 is characterized in that: The S41 specifically includes: S411. Determine the spatial range of the point cloud data: Assume that the coordinates of the point cloud data points are , determine its , , Range of directions , , ; S412, octree initialization and point division: divide the space into two parts with side lengths of Small cube, point exist Direction cube index , direction , direction , the total index can be obtained through Calculate, where and yes and The number of small cubes divided in the direction, so as to divide the points into corresponding small cubes.

5. The elevator risk control safety management method based on the Internet of Things according to claim 2 is characterized in that: The S42 specifically includes: S421, small cube point set selection: for each small cube with data points, select the point set therein; S422: Select three non-collinear points A from the point set ( )、B( )、C( ), calculate the vector =BA=( - , - , - ), =CA=( - , - , - ), the normal vector = ,in , are unit vectors in the x, y, and z directions respectively.

6. The elevator risk control safety management method based on the Internet of Things according to claim 2 is characterized in that: The S43 specifically includes: S431: Determine the equation based on the physical meaning of the point cloud data: ,in , is the function to be sought, Related to point cloud distribution and normal vectors; S432: Divide the space into grids, grid points ( ), Direction grid spacing , for the second-order partial derivative In exist( ) , similarly for , for , bring in and From the equation: + + = , and the solution is The value of S433: According to the solution The value of the isosurface D will satisfy =D's mesh nodes or boundary points are connected into triangular facets to construct an initial three-dimensional model.

7. The elevator risk control safety management method based on the Internet of Things according to claim 2 is characterized in that: The S5 specifically includes: S51: Obtain the real-time reconstruction model in the same way, and simultaneously obtain the corresponding reference points and direction axes of the reconstruction model and the standard model, and calculate the coordinate difference between the reference points of the reconstruction model and the standard model; S52: Calculate the coordinate difference between the bottom center of the reconstructed model and the bottom center of the standard model (0,0,0); S53: Rotate the reconstructed model through the rotation matrix to make the long axis direction consistent, and measure the size of the standard model in each direction; S54: Calculate the scale and scale the coordinates of the reconstructed model; S55: then extracting geometric features of vertex information from the reconstructed model, and then performing feature matching; S56: For each vertex of the reconstructed model, calculate the Euclidean distance to all vertices of the standard model: , take the vertex corresponding to the minimum distance to establish a match, for the matching point pair , calculate the Euclidean distance , calculate the mean distance ,in is the number of matching points and variance and maximum value , check vertex matching; S57: If a large number of vertices cannot be matched or the matching distance exceeds a threshold, it is marked that a foreign object exists.

8. An elevator risk control and safety management system based on the Internet of Things, comprising an elevator shaft, a cloud server, an elevator control room and a maintenance worker's mobile terminal, characterized in that: The elevator shaft includes a processor, a GPS, a three-dimensional laser scanning sensor, an elevator brake, a rotary encoder and a first wireless communication module; The cloud server includes a second wireless communication module, a data comparison module, a database and a self-checking module; The elevator control room includes a third wireless communication module and a display screen; The processor is bidirectionally connected to the GPS, the output end of the three-dimensional laser scanning sensor is connected to the input end of the processor, the output end of the processor is electrically connected to the input end of the elevator brake, the output end of the rotary encoder is connected to the input end of the processor, the processor is bidirectionally connected to the first wireless communication module, the first wireless communication module is bidirectionally connected to the second wireless communication module, the second wireless communication module is bidirectionally connected to the data comparison module, the data preprocessing module and the self-correction module respectively, the data preprocessing module is bidirectionally connected to the model reconstruction module, the model reconstruction module is bidirectionally connected to the database, the data comparison module is bidirectionally connected to the database, the database is bidirectionally connected to the self-correction module, the second wireless communication module is bidirectionally connected to the mobile terminal and the third wireless communication module respectively, the output end of the third wireless communication module is unidirectionally connected to the input end of the display screen, the third wireless communication module is bidirectionally connected to the first wireless communication module, and the mobile terminal is bidirectionally connected to the first wireless communication module.

9. The elevator risk control safety management system based on the Internet of Things according to claim 7 is characterized in that: The three-dimensional laser scanning sensor is a device that uses laser technology to measure the three-dimensional spatial information of an object, and is used to scan and obtain point cloud data in the elevator shaft; The elevator brake is used to control the running speed and stop of the elevator; The rotary encoder is a sensor that converts rotational displacement into a series of digital pulse signals. The frequency of the pulse signal is proportional to the rotation speed of the rotating part. By counting and analyzing the pulse signal, the speed information of the rotating part is obtained, which is used to detect the moving speed of the elevator.

10. The elevator risk control safety management system based on the Internet of Things according to claim 7, characterized in that: The maintenance worker's mobile terminal is used to receive information sent by the cloud server.

Images