Underground garage operation and maintenance task management system and method based on Internet of Things

Abstract

The invention provides an underground garage operation and maintenance task management system and method based on the Internet of Things, and relates to the technical field of garage management, and the system comprises a blind area recognition unit which is used for constructing a garage initial digital twin model based on the structure data of an underground garage, collecting the state data of the underground garage, and storing the state data of the underground garage; recognizing a blind area of the underground garage in combination with a grid comparison algorithm; the path planning unit is used for generating an optimal routing inspection path of the management personnel by utilizing a path planning algorithm according to the identified blind area, based on a routing inspection initial point of the operation and maintenance management personnel and a receiver preset in the underground garage, and determining a receiver to be connected by combining the optimal routing inspection path; and the inspection state monitoring unit is used for executing an inspection operation and maintenance task based on the optimal inspection path, and determining and recording the inspection state of the management personnel according to the receiver to be connected so as to realize monitoring of the operation and maintenance task of the underground garage. The operation and maintenance efficiency, the data accuracy and the management transparency are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of garage management, and in particular to an underground garage operation and maintenance task management system and method based on the Internet of Things. Background Art

[0002] Underground garages are crucial underground spaces for safely parking and managing vehicles. Their operation and maintenance directly impact vehicle safety, facility reliability, and emergency response efficiency. Comprehensive and systematic operation and maintenance management is essential to safeguard vehicle and personnel safety, ensure the stable operation of fire protection, drainage, ventilation, lighting, and other facilities, and enable rapid response to emergencies.

[0003] Currently, the operation and maintenance of underground garages is primarily based on manual inspections, static monitoring, or periodic maintenance. These issues present significant challenges, including incomplete coverage, low inspection efficiency, and untraceable data. This is particularly true in large or complex underground garages, where building obstructions, restricted viewing angles, and other factors can easily create blind spots that are difficult to monitor, preventing them from being discovered or addressed in a timely manner. Traditional monitoring equipment deployments lack dynamic updating capabilities, and inspection routes for operations and maintenance personnel are often determined based on experience. This lack of system optimization can easily lead to missed critical areas and waste of human resources. Furthermore, the lack of standardized recording methods for inspection results makes it difficult to understand inspection status in real time, and to implement automated monitoring and historical traceability.

[0004] Currently, no effective solutions have been proposed for the problems in related technologies. Summary of the Invention

[0005] In view of this, the present invention provides an underground garage operation and maintenance task management system and method based on the Internet of Things to solve the above-mentioned problems.

[0006] In order to solve the above problems, the specific technical solutions adopted by the present invention are as follows:

[0007] According to one aspect of the present invention, an underground garage operation and maintenance task management system based on the Internet of Things is provided, the system comprising:

[0008] The blind spot identification unit is used to build an initial digital twin model of the underground garage based on its structural data, collect its status data, and identify its blind spots using a grid comparison algorithm.

[0009] The path planning unit is used to generate the optimal inspection path for the operator based on the identified blind spot area, the operator's patrol starting point, and the preset receivers in the underground garage using a path planning algorithm, and determine the receiver to be connected based on the optimal inspection path;

[0010] The inspection status monitoring unit is used to perform inspection and maintenance tasks based on the optimal inspection path. According to the receiver to be connected, it determines the inspection status of the management personnel and records it to realize the monitoring of the underground garage operation and maintenance tasks.

[0011] Preferably, the blind spot identification unit includes:

[0012] The initial model building module is used to use LiDAR to obtain and preprocess the point cloud data of the underground garage, and then build the initial digital twin model of the garage based on the preprocessed point cloud data;

[0013] A real-time model building module is used to obtain the status data of the underground garage based on the camera array preset in the underground garage. Based on the status data, a stereo vision matching algorithm with parallax depth is used to build a real-time digital twin model of the garage.

[0014] The grid comparison module is used to grid the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; and combined with the grid comparison method, it identifies the blind spots of the underground garage.

[0015] Preferably, the method of acquiring status data of the underground garage based on a camera array preset in the underground garage and constructing a real-time digital twin model of the garage using a stereoscopic vision matching algorithm of parallax depth according to the status data includes:

[0016] The garage image captured by the camera array is converted into a color space, and the structured light stripe area is extracted based on the brightness channel threshold.

[0017] In the extracted structured light stripe area, the structured light is extracted based on the multi-level threshold segmentation of the hue channel, and the center line of each structured light is extracted;

[0018] Based on the center lines of each structured light, stereo matching is performed in a preset hierarchical order. The three-dimensional coordinates of the matching points are determined by calculating the parallax of the matching points. Point cloud data is generated according to the three-dimensional coordinates of the matching points, and a real-time digital twin model of the garage is constructed.

[0019] Preferably, extracting structured light within the extracted structured light stripe region based on multi-level threshold segmentation of the hue channel, and extracting the center line of each structured light comprises:

[0020] Based on the light intensity in the underground garage, the hue channel threshold of each structured light is set;

[0021] According to the hue channel threshold of each structured light, a multi-level threshold segmentation method is used to separate the structured light stripe masks of each color in the hue channel;

[0022] Morphological processing is performed on each type of structured light stripe mask to extract the initial geometric center line of each structured light. The initial geometric center line is then fitted using line segment fitting to obtain the final center line of each structured light.

[0023] Preferably, the gridding of the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; and identifying the blind area of ​​the underground garage in combination with the grid comparison method includes:

[0024] Align the coordinates of the initial digital twin model of the garage and the real-time digital twin model of the garage, and perform unified voxel grid division on the initial digital twin model of the garage and the real-time digital twin model of the garage respectively;

[0025] Mapping the real-time digital twin model of the garage after voxel grid partitioning to the initial digital twin model of the garage after voxel grid partitioning;

[0026] Based on the mapped initial digital twin model of the garage, the overlap of each grid is calculated and grids with overlap below a preset threshold are selected.

[0027] The filtered grids are marked, and the blind area of ​​the underground garage is determined based on the marked grids.

[0028] Preferably, the path planning unit includes:

[0029] The coverage relationship establishment module is used to determine the center points of all blind areas based on the identified blind areas using a spatial clustering algorithm, and to establish the coverage relationship between the receiver and the blind area center point;

[0030] The key node screening module is used to screen out the minimum set of receivers that can cover the center points of all blind areas according to the coverage relationship and use them as key nodes;

[0031] The inspection path generation module is used to combine the patrol initial point and key nodes and use the improved ant colony algorithm to generate the weighted optimal inspection path;

[0032] The receiver determination module is used to determine the set of receivers to be connected based on the key node sequence passed by the optimal inspection path.

[0033] Preferably, the method of determining the center point of all blind areas based on the identified blind area using a weighted centroid method and establishing a coverage relationship between the receiver and the blind area center point includes:

[0034] According to the point cloud data of the blind area, the weighted centroid method is used to calculate the coordinates of the center point of each blind area;

[0035] According to the receiver's location and signal characteristics, a receiver coverage model is established, and the query is accelerated through spatial indexing to establish the coverage relationship between the receiver and the center point of the blind spot.

[0036] Preferably, establishing a receiver coverage model based on the position and signal characteristics of the receiver, accelerating the query through a spatial index, and establishing a coverage relationship between the receiver and the center point of the blind spot includes:

[0037] Obtain the three-dimensional coordinates and signal coverage of all receivers and construct a spatial representation of the receiver coverage model;

[0038] Load the 3D coordinates of all blind spot center points into the preset spatial index structure and build a data index for fast proximity query;

[0039] The spatial index is used to perform spatial query on the coverage model of each receiver, filter out the blind spot center points that intersect with the receiver coverage range, and generate the coverage relationship between the receiver and the blind spot center points based on the query results.

[0040] Preferably, the inspection status monitoring unit includes:

[0041] The signal connection module is used to establish a signal connection between the transmitter preset on the administrator and the receiver to be connected when the administrator performs the inspection and maintenance task based on the optimal inspection route and inspects the receiver to be connected.

[0042] Inspection record module, used to record the connection time and receiver number of the transmitter and the receiver to be connected, and generate corresponding inspection record entries;

[0043] The inspection status judgment module is used to compare the inspection record entries with the set of receivers to be connected, determine whether the manager's current inspection has inspected all blind areas, and obtain the inspection results.

[0044] According to another aspect of the present invention, a method for managing underground garage operation and maintenance tasks based on the Internet of Things is provided, comprising the following steps:

[0045] S1. Build an initial digital twin model of the garage based on the structural data of the underground garage, collect the status data of the underground garage, and use the grid comparison algorithm to identify the blind spots of the underground garage.

[0046] S2. Based on the identified blind spots, the operator's patrol starting point, and the preset receivers in the underground garage, a path planning algorithm is used to generate the operator's optimal patrol route. The optimal patrol route is then used to determine the receiver to be connected.

[0047] S3. Execute inspection and maintenance tasks based on the optimal inspection path. Determine the inspection status of the management personnel and record it according to the receiver to be connected to realize the monitoring of underground garage operation and maintenance tasks.

[0048] The beneficial effects of the present invention are:

[0049] 1. The present invention realizes a fully automated process from blind spot detection to closed-loop management of personnel inspection tasks. It can build a digital twin model based on structural data and status information to accurately identify coverage blind spots; then reasonably arrange the inspection route through the optimal path planning algorithm, dynamically determine the task nodes, and finally, in the inspection execution stage, combine signal connection and status comparison to record the management personnel's work status in real time, ensure full coverage and traceability of tasks, and greatly improve operation and maintenance efficiency, data accuracy and management transparency.

[0050] 2. The present invention constructs a high-precision initial digital twin model through lidar point cloud, which can comprehensively reflect the spatial structure of the garage. It combines stereo vision matching with multi-level threshold extraction methods to dynamically generate a real-time digital twin model based on parallax depth. Subsequently, through voxel grid division and alignment comparison, the two models are spatially matched with high precision to accurately identify blind spots caused by occlusion, structural changes or lighting differences. This not only achieves high-resolution recognition of blind spots, but also maintains continuous monitoring of invisible areas in complex and dynamic environments.

[0051] 3. The present invention combines the patrol starting point of the management personnel and adopts an improved ant colony algorithm to generate a weighted optimal inspection path that takes into account distance, energy consumption and priority, and determines the set of receivers to be connected based on the key nodes passed in the path. This not only greatly reduces the inspection path length and time cost, but also ensures the integrity of blind spot coverage and the targeted management of receivers, providing efficient and reliable decision support for inspection scheduling and resource allocation in smart garages. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0053] Figure 1 This is a principle block diagram of an underground garage operation and maintenance task management system based on the Internet of Things according to an embodiment of the present invention;

[0054] Figure 2 The present invention is a flowchart of an underground garage operation and maintenance task management method based on the Internet of Things.

[0055] In the picture:

[0056] 1. Blind spot identification unit; 101. Initial model construction module; 102. Real-time model construction module; 103. Grid comparison module; 2. Path planning unit; 201. Coverage relationship establishment module; 202. Key node screening module; 203. Inspection path generation module; 204. Receiver determination module; 3. Inspection status monitoring unit; 301. Signal connection module; 302. Inspection recording module; 303. Inspection status judgment module. DETAILED DESCRIPTION

[0057] In order to enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0058] According to an embodiment of the present invention, an Internet of Things-based underground garage operation and maintenance task management system and method are provided.

[0059] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, an underground garage operation and maintenance task management system based on the Internet of Things is provided, the system comprising:

[0060] Blind spot identification unit 1 is used to build an initial digital twin model of the garage based on the structural data of the underground garage, collect the status data of the underground garage, and identify the blind spots of the underground garage in combination with the grid comparison algorithm;

[0061] As a preferred embodiment, the blind spot identification unit 1 includes:

[0062] An initial model building module 101 is used to acquire and pre-process point cloud data of the underground garage using a laser radar, and to build an initial digital twin model of the garage based on the pre-processed point cloud data;

[0063] It should be noted that the preprocessing of point cloud data includes: denoising and filtering; denoising is to remove random noise, error points and external interference in point cloud data; filtering is to further process abnormal points and non-target areas in point cloud data, and is often used to remove ground points and local details and enhance the performance of main target objects.

[0064] A real-time model building module 102 is configured to obtain status data of the underground garage based on a camera array preset in the underground garage, and to build a real-time digital twin model of the garage based on the status data using a stereo vision matching algorithm with parallax depth;

[0065] As a preferred embodiment, the method of acquiring the status data of the underground garage based on a camera array preset in the underground garage and constructing a real-time digital twin model of the garage using a stereoscopic vision matching algorithm with parallax depth according to the status data includes:

[0066] The garage image captured by the camera array is converted into a color space, and the structured light stripe area is extracted based on the brightness channel threshold.

[0067] It should be noted that the camera array uses binocular cameras. Specifically, a group of binocular cameras (baseline length 0.5m) can be installed every 10m on the top of the underground garage to ensure that the field of view covers all parking spaces and passages.

[0068] Color space conversion typically involves converting from RGB to HSV (hue H, saturation S, brightness V) or YUV space. When extracting structured light streaks, brightness (the V or Y channel) is the most stable reference channel, effectively separating the background from the high-contrast structured light. An adaptive or fixed threshold is then set based on the brightness channel to extract the high-brightness structured light illumination region; this region typically appears as regular, uniform, bright streaks.

[0069] In the extracted structured light stripe area, the structured light is extracted based on the multi-level threshold segmentation of the hue channel, and the center line of each structured light is extracted;

[0070] As a preferred embodiment, the extracting structured light within the extracted structured light stripe region based on multi-level threshold segmentation of the hue channel and extracting the center line of each structured light includes:

[0071] Based on the light intensity in the underground garage, the hue channel threshold of each structured light is set;

[0072] Specifically, the light intensity in underground garages (typically 50-500 lux) changes the color temperature distribution of camera-captured images. In low light, increased sensor noise causes compression of hue channel values ​​(for example, the hue of a blue stripe might shift from the standard 180° to 160°). In high light, the hue of overexposed areas diffuses (for example, a green stripe might diffuse from 120° to 100-140°).

[0073] Among them, fixed thresholds can lead to missed or false detection of stripes in areas with sudden changes in lighting (such as garage entrances and exits). Dynamic thresholds automatically adjust the hue segmentation range by sensing the ambient light intensity in real time, ensuring stable extraction under different lighting conditions. When setting dynamic thresholds, it is necessary to establish a linear / nonlinear relationship between light intensity E and hue threshold T. For example:

[0074] Blue stripes: T low =180°-0.1E, T high =180°+0.05E;

[0075] Green stripes: T low =120°-0.08E, T high =120°+0.03E;

[0076] T low Indicates the lowest value of the hue threshold, T high It represents the lowest value of the hue threshold. Finally, light sensors are evenly distributed on the roof of the garage (for example, one every 20m). The multi-point average value is used as the current light intensity input to reduce local shadow interference.

[0077] According to the hue channel threshold of each structured light, a multi-level threshold segmentation method is used to separate the structured light stripe masks of each color in the hue channel;

[0078] It should be noted that when using the multi-level threshold segmentation method to separate the color structured light stripe mask in the hue channel, it is necessary to divide the hue channel into multiple sub-intervals and set thresholds for different color structured lights (such as blue and green). For example:

[0079] Blue stripes: Main range 160°-200°, auxiliary range 150°-210° (for low-contrast areas).

[0080] Green stripes: main range 100°-140°, auxiliary range 90°-150°.

[0081] For areas with uneven lighting (such as near pillars), a sliding window (such as 11×11 pixels) is used to calculate the local hue mean and dynamically correct the threshold range to avoid over-segmentation or under-segmentation caused by the global threshold.

[0082] Morphological processing is performed on each type of structured light stripe mask to extract the initial geometric center line of each structured light. The initial geometric center line is then fitted using line segment fitting to obtain the final center line of each structured light.

[0083] It should be noted that the morphological processing performed on each type of structured light stripe mask includes thinning and breakpoint repair. Thinning compresses wider structured light stripes into one-pixel-wide lines to obtain an image representation closer to the actual beam path; while breakpoint repair is used to bridge the stripe interruptions caused by noise, occlusion, or imaging distortion, making the extracted light coherent and continuous. Through thinning algorithms (such as central axis extraction or image skeletonization), the initial geometric skeleton of the structured light stripes in the image can be obtained, providing a structural foundation for subsequent geometric fitting.

[0084] Based on the center lines of each structured light, stereo matching is performed in a preset hierarchical order. The three-dimensional coordinates of the matching points are determined by calculating the parallax of the matching points. Point cloud data is generated according to the three-dimensional coordinates of the matching points, and a real-time digital twin model of the garage is constructed.

[0085] It should be noted that the camera array uses a binocular camera layout, so each binocular camera group has a fixed left and right viewing angle hierarchy; the extracted structured light centerlines need to be matched according to the corresponding order between the left and right image pairs to ensure spatial geometric consistency;

[0086] In binocular images, the center point of the same structured light will have a pixel position offset between the left and right views. This offset is called disparity. Disparity is inversely proportional to depth: the larger the disparity, the closer the object. Therefore, when calculating the disparity of a matching point, you need to find a point on the center line of a structured light in the left image, then search for a matching point on the corresponding light in the right image (this can be based on neighborhood similarity, structured light characteristics, etc.), and then calculate the pixel offset between the two points to obtain the disparity value.

[0087] In addition, when calculating the three-dimensional coordinates based on the parallax, the calculation formula is:

[0088] Z = (f·B) / d;

[0089] Where Z represents the depth coordinate (the distance from the camera); f represents the focal length of the camera; B represents the baseline length between the binocular cameras; and d represents the disparity of the matching points.

[0090] The corresponding spatial coordinates can be further calculated using Z and the pixel coordinates in the image, that is, the two-dimensional points in the image can be converted into three-dimensional space points through stereo geometric projection.

[0091] The grid comparison module 103 is used to perform grid division on the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; and combine the grid comparison method to identify the blind spots of the underground garage.

[0092] As a preferred embodiment, the grid division of the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; and combining the grid comparison method to identify the blind area of ​​the underground garage includes:

[0093] Align the coordinates of the initial digital twin model of the garage and the real-time digital twin model of the garage, and perform unified voxel grid division on the initial digital twin model of the garage and the real-time digital twin model of the garage respectively;

[0094] It should be noted that in the two digital twin models, the initial model is the original, designed state of the garage, while the real-time model is the current actual state of the garage. In order to compare, it is necessary to first ensure that the two are spatially aligned. This is usually done through rigid transformations (for example, translation, rotation) to ensure that the two share a unified coordinate system.

[0095] Once the models are aligned, both models need to be meshed using voxels. Voxels are the basic units of three-dimensional space, similar to pixels in a two-dimensional image. Meshing divides the three-dimensional space of the garage into several small voxel units, each of which represents a certain area or volume of the garage.

[0096] Mapping the real-time digital twin model of the garage after voxel grid partitioning to the initial digital twin model of the garage after voxel grid partitioning;

[0097] Specifically, when mapping the grid of the real-time model to the grid of the initial model, it is necessary to map the voxel position of the real-time data to the corresponding voxel area of ​​the initial garage model through the transformation relationship after coordinate alignment; so that each voxel of the real-time model can be associated with the corresponding position of the initial model, thereby providing a basis for subsequent overlap calculation and blind spot identification.

[0098] Based on the mapped initial digital twin model of the garage, the overlap of each grid is calculated and grids with overlap below a preset threshold are selected.

[0099] The filtered grids are marked, and the blind area of ​​the underground garage is determined based on the marked grids.

[0100] Specifically, the overlap of each voxel grid refers to the ratio of the intersection area (or volume) of the voxel in the real-time model and the corresponding voxel area in the initial model to the total area (or volume). Then, by calculating the overlap of each grid, the degree of overlap between a certain area of ​​the real-time model and the initial model can be evaluated. By comparing with the preset threshold, the standard for determining whether the area is a blind spot is determined based on the comparison result. For example, if the overlap is less than a certain set value (such as 30%), the grid area is considered a blind spot.

[0101] Path planning unit 2 is used to generate the optimal patrol path for the operator based on the identified blind spot area, the operator's patrol starting point, and the preset receivers in the underground garage using a path planning algorithm, and determine the receiver to be connected based on the optimal patrol path;

[0102] As a preferred embodiment, the path planning unit 2 includes:

[0103] The coverage relationship establishment module 201 is used to determine the center point of all blind areas based on the identified blind areas using a spatial clustering algorithm, and establish a coverage relationship between the receiver and the blind area center point;

[0104] As a preferred embodiment, the method of determining the center point of all blind areas based on the identified blind area using the weighted centroid method and establishing the coverage relationship between the receiver and the blind area center point includes:

[0105] According to the point cloud data of the blind area, the weighted centroid method is used to calculate the coordinates of the center point of each blind area;

[0106] It should be noted that before using the weighted centroid method to calculate the coordinates of the center point of each blind spot area, it is necessary to assign different weights to different points based on the density distribution or reflection intensity distribution of the point cloud in the blind spot area. Specifically, first, it is necessary to count the local neighborhood point density of each point in the blind spot area (such as based on the number of points within a fixed radius or based on the k-nearest neighbor distance to estimate the density); then points with higher density usually indicate that the area has been observed multiple times or is a more stable part of the object surface, so it is given a higher weight; the weight can be set to the local density value of the point or its normalized value.

[0107] After weight distribution, weighted centroid calculation is required, and the calculation formula is:

[0108]

[0109] In the formula, (x i ,y i ,z i ) represents the coordinates of the i-th point; w i Indicates the weight based on density allocation; N indicates the total number of points in the blind area; (x c ,y c ,z c ) represents the coordinates of the center point.

[0110] Based on the receiver's location and signal characteristics, a receiver coverage model is established, and spatial indexing is used to accelerate queries and establish the coverage relationship between the receiver and the center point of the blind spot.

[0111] As a preferred embodiment, the receiver coverage model is established based on the position and signal characteristics of the receiver, and the query is accelerated by spatial indexing to establish the coverage relationship between the receiver and the center point of the blind spot, including:

[0112] Obtain the three-dimensional coordinates and signal coverage of all receivers and construct a spatial representation of the receiver coverage model;

[0113] It should be noted that each receiver has a fixed three-dimensional spatial position and signal coverage parameters defined by its physical characteristics, such as: coverage radius (spherical signal model, suitable for omnidirectional receivers); directivity angle and direction vector (constructing a cone or fan-shaped model, suitable for directional receivers); then, combined with the above parameters, the receiver coverage range model can be simulated in three-dimensional space using standard geometric bodies (such as spheres, cones, and polyhedrons).

[0114] Load the 3D coordinates of all blind spot center points into the preset spatial index structure and build a data index for fast proximity query;

[0115] It should be noted that when loading the 3D coordinates of all blind spot center points into the preset spatial index structure, the set of blind spot center points calculated using the weighted centroid method must first be organized into a unified data format and imported into an efficient 3D spatial index structure (such as a kd-tree, R-tree, or Octree). These structures can support fast proximity queries and range intersection operations, significantly reducing the overhead of brute-force traversal of all blind spot points in receiver coverage relationship calculations.

[0116] The spatial index is used to perform spatial query on the coverage model of each receiver (to determine which points the receiver model overlaps with), filter out the blind spot center points that intersect with the receiver coverage range, and generate the coverage relationship between the receiver and the blind spot center points based on the query results.

[0117] Specifically, a spatial index intersection query is performed on the geometric coverage model of each receiver; the query result is the set of blind spot center points that can be effectively covered by the receiver; the result is recorded as a set of "receiver-blind spot center point" mapping relationships to form a coverage matrix, that is, a Boolean structure of coverage status.

[0118] The key node screening module 202 is used to screen out the minimum receiver set that can cover the center points of all blind areas according to the coverage relationship and use it as the key node;

[0119] Specifically, it should be noted that some blind spots in the coverage relationship may be covered by multiple receivers. Therefore, a linear programming / integer linear programming model is used to construct a "minimum receiver set" coverage problem. The constraint is that the center of each blind spot is covered by at least one receiver, and the optimization goal is to minimize the number of receivers used. Finally, a heuristic algorithm (such as the greedy set cover algorithm) is used to approximate the solution.

[0120] The inspection path generation module 203 is used to generate a weighted optimal inspection path by combining the patrol initial point and key nodes using an improved ant colony algorithm;

[0121] Specifically, the improved ant colony algorithm is used to generate the weighted optimal inspection path. This means that based on the spatial distribution information of the patrol starting point and the screened minimum receiver set (i.e., key nodes), a weighted graph model including factors such as distance, energy consumption, and priority is constructed, and the improved ant colony algorithm (ACO) is used to search for an inspection path that passes through all key nodes and has the lowest total cost.

[0122] The receiver determination module 204 is configured to determine a set of receivers to be connected based on a sequence of key nodes that the optimal inspection path passes through.

[0123] It's important to note that the sequence of key nodes along the optimal inspection path directly determines the set of receivers that need to be connected and inspected, known as the set of receivers to be connected. Because these key nodes are selected through minimum coverage optimization, they collectively cover the entire blind spot area.

[0124] The inspection status monitoring unit 3 is used to perform inspection and operation tasks based on the optimal inspection path, determine the inspection status of the management personnel and record it according to the receiver to be connected, so as to realize the monitoring of the underground garage operation and maintenance tasks.

[0125] As a preferred embodiment, the inspection status monitoring unit 3 includes:

[0126] The signal connection module 301 is used to establish a signal connection between the transmitter preset on the administrator and the receiver to be connected when the administrator performs the inspection and operation task based on the optimal inspection path and inspects the receiver to be connected.

[0127] Inspection record module 302, used to record the connection time and receiver number of the transmitter and the receiver to be connected, and generate corresponding inspection record entries;

[0128] The inspection status judgment module 303 is used to compare the inspection record entries with the set of receivers to be connected, determine whether the current inspection by the manager has inspected all blind areas, and obtain the inspection result.

[0129] Specifically, in the process of executing inspection and operation tasks based on the optimal inspection path, the transmitter carried by the manager and the key receivers in the path are dynamically connected through signal connection. When the manager enters the signal coverage range of a certain receiver, communication is automatically established to ensure contactless confirmation. Subsequently, the inspection record module will record the timestamp and receiver number of each connection in real time to form a standardized inspection log entry. Finally, the inspection status judgment module compares the generated records with the preset list of key receivers to determine whether the inspection tasks for all target areas have been completed, and outputs the results of whether the inspection has met the standards. This mechanism not only realizes the closed loop of automated inspection execution and data collection, but also provides a traceable, quantifiable and evaluable execution basis for underground garage operation and maintenance tasks, significantly improving management transparency and inspection compliance.

[0130] like Figure 2 According to another embodiment of the present invention, a method for managing underground garage operation and maintenance tasks based on the Internet of Things is provided, comprising the following steps:

[0131] S1. Build an initial digital twin model of the garage based on the structural data of the underground garage, collect the status data of the underground garage, and use the grid comparison algorithm to identify the blind spots of the underground garage.

[0132] S2. Based on the identified blind spots, the operator's patrol starting point, and the preset receivers in the underground garage, a path planning algorithm is used to generate the operator's optimal patrol route. The optimal patrol route is then used to determine the receiver to be connected.

[0133] S3. Execute inspection and operation maintenance tasks based on the optimal inspection path, determine the inspection status of the management personnel and record it according to the receiver to be connected, so as to realize the monitoring of the operation and maintenance tasks of the underground garage.

[0134] In summary, with the help of the above technical solutions of the present invention, the present invention realizes a fully automated process from blind spot detection to closed-loop management of personnel inspection tasks, and can build a digital twin model based on structural data and status information to accurately identify and cover blind spots; then, through the optimal path planning algorithm, the inspection route is reasonably arranged, and the task nodes are dynamically determined. Finally, in the inspection execution stage, the signal connection and status comparison are combined to record the work status of the management personnel in real time to ensure full coverage and traceability of the task, greatly improving the operation and maintenance efficiency, data accuracy and management transparency. The present invention constructs a high-precision initial digital twin model through the lidar point cloud, which can fully reflect the garage space structure, and combines stereo vision matching with multi-level threshold extraction methods to dynamically generate a real-time digital twin model based on parallax depth; then, through voxel grid division and alignment comparison, the two models are spatially accurately matched to accurately identify blind spots caused by occlusion, structural changes or lighting differences, which not only achieves high-resolution recognition of blind spots, but also maintains continuous monitoring of invisible areas in complex and dynamic environments. The present invention combines the patrol starting point of the management personnel and adopts an improved ant colony algorithm to generate a weighted optimal inspection path that takes into account distance, energy consumption and priority, and determines the set of receivers to be connected based on the key nodes passed in the path. This not only greatly shortens the inspection path length and time cost, but also ensures the integrity of blind spot coverage and the targeted management of receivers, providing efficient and reliable decision support for inspection scheduling and resource allocation in smart garages.

[0135] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, optical storage, etc.) containing computer-usable program code.

[0136] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An underground garage operation and maintenance task management system based on the Internet of Things, characterized in that: The system includes: The blind spot identification unit is used to build an initial digital twin model of the underground garage based on its structural data, collect its status data, and identify its blind spots using a grid comparison algorithm. The path planning unit is used to generate the optimal inspection path for the operator based on the identified blind spot area, the operator's patrol starting point, and the preset receivers in the underground garage using a path planning algorithm, and determine the receiver to be connected based on the optimal inspection path; The inspection status monitoring unit is used to perform inspection and maintenance tasks based on the optimal inspection path. According to the receiver to be connected, it determines the inspection status of the management personnel and records it to realize the monitoring of the underground garage operation and maintenance tasks.

2. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 1 is characterized in that: The blind spot recognition unit includes: The initial model building module is used to use LiDAR to obtain and preprocess the point cloud data of the underground garage, and then build the initial digital twin model of the garage based on the preprocessed point cloud data; A real-time model building module is used to obtain the status data of the underground garage based on the camera array preset in the underground garage. Based on the status data, a stereo vision matching algorithm with parallax depth is used to build a real-time digital twin model of the garage. The grid comparison module is used to grid the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; and combined with the grid comparison method, it identifies the blind spots of the underground garage.

3. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 2 is characterized in that: The method of acquiring status data of the underground garage based on a camera array preset in the underground garage and constructing a real-time digital twin model of the garage using a stereoscopic vision matching algorithm with parallax depth based on the status data includes: The garage image captured by the camera array is converted into a color space, and the structured light stripe area is extracted based on the brightness channel threshold. In the extracted structured light stripe area, the structured light is extracted based on the multi-level threshold segmentation of the hue channel, and the center line of each structured light is extracted; Based on the center lines of each structured light, stereo matching is performed in a preset hierarchical order. The three-dimensional coordinates of the matching points are determined by calculating the parallax of the matching points. Point cloud data is generated according to the three-dimensional coordinates of the matching points, and a real-time digital twin model of the garage is constructed.

4. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 3 is characterized in that: The extracting of structured light within the extracted structured light stripe region by multi-level threshold segmentation based on the hue channel and extracting the center line of each structured light comprises: Based on the light intensity in the underground garage, the hue channel threshold of each structured light is set; According to the hue channel threshold of each structured light, a multi-level threshold segmentation method is used to separate the structured light stripe masks of each color in the hue channel; Morphological processing is performed on each type of structured light stripe mask to extract the initial geometric center line of each structured light. The initial geometric center line is then fitted using line segment fitting to obtain the final center line of each structured light.

5. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 2 is characterized in that: The initial digital twin model of the garage and the real-time digital twin model of the garage are meshed respectively; Combined with the grid comparison method, the blind areas of the underground garage are identified, including: Align the coordinates of the initial digital twin model of the garage and the real-time digital twin model of the garage, and perform unified voxel grid division on the initial digital twin model of the garage and the real-time digital twin model of the garage respectively; Mapping the real-time digital twin model of the garage after voxel grid partitioning to the initial digital twin model of the garage after voxel grid partitioning; Based on the mapped initial digital twin model of the garage, the overlap of each grid is calculated and grids with overlap below a preset threshold are selected. The filtered grids are marked, and the blind area of ​​the underground garage is determined based on the marked grids.

6. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 1 is characterized in that: The path planning unit includes: The coverage relationship establishment module is used to determine the center points of all blind areas based on the identified blind areas using a spatial clustering algorithm, and to establish the coverage relationship between the receiver and the blind area center point; The key node screening module is used to screen out the minimum set of receivers that can cover the center points of all blind areas according to the coverage relationship and use them as key nodes; The inspection path generation module is used to combine the patrol initial point and key nodes and use the improved ant colony algorithm to generate the weighted optimal inspection path; The receiver determination module is used to determine the set of receivers to be connected based on the key node sequence passed by the optimal inspection path.

7. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 6 is characterized in that: The method of determining the center point of all blind areas based on the identified blind area using a weighted centroid method and establishing a coverage relationship between the receiver and the blind area center point includes: According to the point cloud data of the blind area, the weighted centroid method is used to calculate the coordinates of the center point of each blind area; According to the receiver's location and signal characteristics, a receiver coverage model is established, and the query is accelerated through spatial indexing to establish the coverage relationship between the receiver and the center point of the blind spot.

8. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 7 is characterized in that: The method of establishing a receiver coverage model based on the receiver's position and signal characteristics, accelerating queries through spatial indexing, and establishing a coverage relationship between the receiver and the center point of the blind spot includes: Obtain the three-dimensional coordinates and signal coverage of all receivers and construct a spatial representation of the receiver coverage model; Load the 3D coordinates of all blind spot center points into the preset spatial index structure and build a data index for fast proximity query; The spatial index is used to perform spatial query on the coverage model of each receiver, filter out the blind spot center points that intersect with the receiver coverage range, and generate the coverage relationship between the receiver and the blind spot center points based on the query results.

9. The underground garage operation and maintenance task management system based on the Internet of Things according to claim 1 is characterized in that: The inspection status monitoring unit includes: The signal connection module is used to establish a signal connection between the transmitter preset on the administrator and the receiver to be connected when the administrator performs the inspection and maintenance task based on the optimal inspection route and inspects the receiver to be connected. Inspection record module, used to record the connection time and receiver number of the transmitter and the receiver to be connected, and generate corresponding inspection record entries; The inspection status judgment module is used to compare the inspection record entries with the set of receivers to be connected, determine whether the manager's current inspection has inspected all blind areas, and obtain the inspection results.

10. A method for managing underground garage operation and maintenance tasks based on the Internet of Things, using the underground garage operation and maintenance task management system based on the Internet of Things as claimed in any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Build an initial digital twin model of the garage based on the structural data of the underground garage, collect the status data of the underground garage, and use the grid comparison algorithm to identify the blind spots of the underground garage. S2. Based on the identified blind spots, the operator's patrol starting point, and the preset receivers in the underground garage, a path planning algorithm is used to generate the operator's optimal patrol route. The optimal patrol route is then used to determine the receiver to be connected. S3. Execute inspection and operation maintenance tasks based on the optimal inspection path, determine the inspection status of the management personnel and record it according to the receiver to be connected, so as to realize the monitoring of the operation and maintenance tasks of the underground garage.

Images