Construction site safety management and control method based on regional positioning visualization and related equipment

Abstract

The application discloses a construction site safety management and control method based on regional positioning visualization and related equipment, and comprises the following steps: performing feature point extraction operation on a first local construction drawing and a first standard GIS base map, then performing weighted feature point matching operation on the first local construction drawing and the first standard GIS base map to obtain a fine matching point pair; mapping the first local construction drawing to the first standard GIS base map based on the fine matching point pair to obtain a second standard GIS base map; performing safety region boundary extraction on the second standard GIS base map to generate an initial topological geometric body; assigning a space-time region code to the initial topological geometric body to obtain a target topological geometric body; and performing space collision detection on the target topological geometric body and a dynamic moving object, and when collision is detected and a current collision timestamp matches a high-risk valid time period, triggering cross-module early warning. The application can improve the accuracy of safety management and control, and can be widely applied to the technical field of human-computer interaction and safety management and control.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction and safety management technology, and in particular to a construction site safety management method and related equipment based on regional positioning visualization. Background Technology

[0002] The construction of smart construction sites is accelerating, and the application of digital technologies in the field of construction site safety management is gradually becoming more widespread. HSE systems, video surveillance, smart safety helmets, and other equipment have become important supports for safety management. However, safety management in complex construction environments still faces key technical bottlenecks: On the one hand, construction site scenarios are dynamic and ever-changing, with flexible safety zone boundaries, highly personalized safety levels and management needs, and complex situations such as severe obstruction, the addition of temporary work areas, and cross-operation in multiple areas of critical and major projects, requiring extremely high accuracy and flexibility in safety labeling. On the other hand, the industry's requirements for integrated, real-time, and traceable safety management are continuously increasing, while existing technologies have significant shortcomings: automatic identification technology has poor adaptability in complex scenarios and struggles to accurately capture temporarily changed safety areas; safety configurations are mostly fixed templates, lacking deep correlation with safety scores (generated based on probability, exposure, and severity), and cannot flexibly adapt to the management needs of different types of safety areas; management rules lack linkage, and safety area management is disconnected from modules such as critical and major projects, hazard investigation, and on-site monitoring records; visualization and interactive functions are weak, base map adaptability is insufficient, manual labeling tools are limited, and multi-terminal collaborative response capabilities are lacking. Furthermore, existing technologies mostly focus on optimizing single functional modules, lacking a systematic design that covers the entire process from area labeling, attribute configuration, rule linkage to multi-module collaboration, making it impossible to efficiently and accurately manage the safety of complex construction sites. Summary of the Invention

[0003] In view of this, the main objective of the embodiments of the present invention is to provide a method and related equipment for construction site safety management based on regional positioning visualization, in order to solve at least one of the problems of the prior art. The present invention can improve the accuracy of safety management.

[0004] To achieve the above objectives, one aspect of the present invention provides a construction site safety management method based on regional positioning visualization, the method comprising: Feature point extraction is performed on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; Based on the first original descriptor and the second original descriptor, a weighted feature point matching operation is performed on the first partial construction drawing and the first standard GIS base map to obtain a finely matched point pair; Based on the precisely matched point pairs, the first partial construction drawing is mapped to the first standard GIS base map to obtain the second standard GIS base map; The security zone boundaries are extracted from the second standard GIS base map to generate the initial topological geometry; Spatiotemporal region encoding is assigned to the initial topological geometry to obtain the target topological geometry; Spatial collision detection is performed on the target topological geometry and the dynamically moving object. When a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region encoding, a cross-module warning is triggered.

[0005] In some embodiments, the step of extracting feature points from the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor includes the following steps: The ORB algorithm is used to detect key points and generate descriptors for the first partial construction drawing, so as to obtain the first key point and the corresponding first original descriptor. The ORB algorithm is used to detect key points and generate descriptors on the first standard GIS base map to obtain the second key point and the corresponding second original descriptor.

[0006] In some embodiments, the step of performing a weighted feature point matching operation on the first partial construction drawing and the first standard GIS base map based on the first original descriptor and the second original descriptor to obtain a finely matched point pair includes the following steps: Using a lightweight semantic segmentation model, key structural category identification operations are performed on the first partial construction drawing and the first standard GIS base map respectively to obtain the first identification result and the second identification result. Based on the first identification result, a first weight is assigned to the first feature point in the first original descriptor that belongs to the preset high-value region, and a second weight is assigned to the second feature point in the first original descriptor that belongs to the background region, to obtain a first weighted descriptor; Based on the second identification result, a first weight is assigned to the third feature point in the second original descriptor that belongs to the preset high-value region, and a second weight is assigned to the fourth feature point in the second original descriptor that belongs to the background region, to obtain the second weighted descriptor; Obtain the Hamming distance between each of the first weighted descriptors and each of the second weighted descriptors; Matching priority is adjusted based on weight magnitude; wherein, the first weight is greater than the second weight; Based on the Hamming distance and the matching priority, coarse matching point pairs between the first weighted descriptor and the second weighted descriptor are obtained; The coarse-matched point pairs are processed by the multi-scale RANSAC algorithm to remove mismatched points, thereby obtaining fine-matched point pairs.

[0007] In some embodiments, mapping the first partial construction drawing to the first standard GIS base map based on the precisely matched point pairs to obtain the second standard GIS base map includes the following steps: Based on the precisely matched point pairs, obtain the affine transformation matrix between the first partial construction drawing and the first standard GIS base map; Based on the affine transformation matrix, the pixel coordinates of the first local construction drawing are mapped to the geospatial coordinates of the first standard GIS base map to obtain the second standard GIS base map.

[0008] In some embodiments, the step of extracting the security area boundary of the second standard GIS base map to generate an initial topological geometry includes the following steps: In response to the first operation command on the security source in the second standard GIS base map, a preset semantic segmentation model is invoked to generate the initial polygon boundary; The initial polygon boundary is rendered to generate an editable graphic with control nodes; In response to a second operation command on a control node in the editable graphic, the initial polygon boundary is corrected to obtain a corrected polygon boundary. Based on the modified polygon boundary and the influence range parameters of the security source, the initial topological geometry representing the boundary of the security area is generated by a geometric algorithm.

[0009] In some embodiments, the process of assigning spatiotemporal region encoding to the initial topological geometry to obtain the target topological geometry includes the following steps: Add a time validity attribute to the initial topological geometry; the time validity attribute includes at least the high-risk valid time period; Based on the initial topological geometry and the time validity attribute, a spatiotemporal region code is generated, and the spatiotemporal region code is embedded in the initial topological geometry to obtain the target topological geometry.

[0010] In some embodiments, the spatial collision detection of the target topological geometry and the dynamically moving object, when a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches a high-risk valid time period in the spatiotemporal region encoding, triggers a cross-module warning, including the following steps: Acquire the position stream data of the dynamically moving objects at the construction site; Based on the location flow data and the static reference boundary of the target topological geometry, a spatial intersection operation is performed between the dynamically moving object and the target topological geometry using the ray method or the separating axis theorem to obtain a collision detection result; the collision detection result includes the collision state and the current collision timestamp; Obtain the time validity attribute from the spatiotemporal region code of the target topological geometry, and extract the high-risk valid time period from the time validity attribute; When the collision state is that the dynamically moving object collides with the target topological geometry, and the current collision timestamp is within the high-risk valid time period, it is determined to be a valid risk collision; In response to the effective risk collision, a preset instruction template is invoked to conduct hazard investigation and on-site audible and visual alarms.

[0011] To achieve the above objectives, another aspect of this invention proposes a construction site safety management and control device based on regional positioning visualization, the device comprising: The feature point extraction module is used to perform feature point extraction operations on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; The feature point matching module is used to perform a weighted feature point matching operation on the first partial construction drawing and the first standard GIS base map based on the first original descriptor and the second original descriptor to obtain finely matched point pairs; The spatial coordinate system mapping module is used to map the first local construction drawing to the first standard GIS base map based on the precise matching point pair to obtain the second standard GIS base map; The first dynamic topology generation module is used to extract the safe area boundary of the second standard GIS base map and generate the initial topological geometry. The second dynamic topology generation module is used to assign spatiotemporal region encoding to the initial topology geometry to obtain the target topology geometry. The collision detection and early warning module is used to perform spatial collision detection between the target topological geometry and the dynamically moving object. When a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region encoding, a cross-module early warning is triggered.

[0012] To achieve the above objectives, another aspect of the present invention provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described above.

[0013] To achieve the above objectives, another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0014] To achieve the above objectives, another aspect of the present invention provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0015] The embodiments of the present invention include at least the following beneficial effects: The present invention provides a construction site safety management method and related equipment based on regional positioning visualization. This scheme extracts feature points from a first partial construction drawing and a first standard GIS base map to obtain a first original descriptor and a second original descriptor. This can transform the geometric texture in the drawing into a machine-recognizable high-dimensional vector, effectively filtering out noise interference caused by messy lines. Based on the first original descriptor and the second original descriptor, a weighted feature point matching operation is performed on the first partial construction drawing and the first standard GIS base map to obtain finely matched point pairs. By introducing a weight mechanism, the alignment accuracy between heterogeneous data is improved, and the mismatch rate caused by scale differences is reduced. Based on the finely matched point pairs, The first partial construction drawing is mapped to the first standard GIS base map to obtain the second standard GIS base map, eliminating coordinate deviations from multi-source data. Safe zone boundaries are extracted from the second standard GIS base map to generate an initial topological geometry. Spatiotemporal region codes are assigned to the initial topological geometry to obtain the target topological geometry, providing a data foundation for refined management of dynamic risks. Spatial collision detection is performed between the target topological geometry and dynamically moving objects. When a collision is detected between a dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region code, a cross-module warning is triggered. Through a spatiotemporal dual verification mechanism, false alarms from single spatial detection are effectively avoided, ensuring the accuracy and timeliness of the warning trigger. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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 effort.

[0017] Figure 1This is a flowchart of a construction site safety management method based on regional positioning visualization provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the region boundary extraction operation provided in an embodiment of the present invention; Figure 3 This is a visual schematic diagram of the region boundary provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this invention; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this invention as detailed in the appended claims.

[0019] It should be noted that although functional modules are divided in the system diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first / S100" and "second / S200" in the specification, claims, and the foregoing drawings may be used herein to describe various concepts, but unless specifically stated otherwise, these concepts are not limited by these terms. These terms are used only to distinguish one concept from another. For example, first information may also be referred to as second information without departing from the scope of the embodiments of the invention, and similarly, second information may also be referred to as first information. Depending on the context, the words "if" or "when" as used herein may be interpreted as "when," "in response to a determination," or "in the event of a determination."

[0020] The terms “at least one,” “multiple,” “each,” “any,” etc., used in this invention, “at least one” includes one, two, or more than two; “multiple” includes two or more than two; “each” refers to each of the corresponding multiple; and “any” refers to any one of the multiple.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0022] Before providing a detailed description of the embodiments of the present invention, some of the nouns and terms involved in the embodiments of the present invention will be explained first. The nouns and terms involved in the embodiments of the present invention are subject to the following interpretations.

[0023] Random Sample Consensus (RANSAC) is an efficient method for estimating mathematical model parameters from a sample set containing outliers. The algorithm assumes that the data contains inliers that can be described by the model and outliers that deviate from the model. It initializes the model by randomly sampling a minimum sample set and iteratively filters the inlier set to optimize the parameters based on a set threshold.

[0024] ORB (Oriented FAST and Rotated BRIEF) is an image feature detection and description algorithm that combines speed and robustness. This algorithm significantly improves the stability of feature matching by improving the orientation sensitivity of the FAST keypoint detector and the rotation invariance of the BRIEF descriptor.

[0025] A Geographic Information System (GIS) is a specific type of spatial information system. It is a technological system that, with the support of computer hardware and software, collects, stores, manages, processes, analyzes, displays, and describes geographic distribution data across the entire or part of the Earth's surface (including the atmosphere).

[0026] Seed points are key points used in image processing and brain function analysis to start regions or nodes for region growth or functional connectivity calculations.

[0027] In related technologies, there are methods that use ORB feature matching and affine transformation to achieve the positioning and alignment of construction drawings and GIS base maps. However, construction drawings contain a large number of repetitive textures, symmetrical structures, and linear components (such as walls, steel bars, and scaffolding). Conventional ORB algorithms will generate a large number of mismatched feature points, which cannot be completely solved by RANSAC algorithm alone. The final positioning and alignment error is generally greater than 0.5 meters. Moreover, the alignment result is a one-time output. The coordinate deviation found in subsequent field applications cannot be corrected by reversing the alignment parameters. The error will be passed on to the boundary configuration and collision verification stages, eventually leading to control failure. The area configuration of existing technologies is only a simple mode of "purely manual hand drawing" or "AI initial extraction + manual node drag and drop fine adjustment". In addition, the control level and control rules are only statically set during the initial area configuration and will not be dynamically adjusted according to the cross-operation status, personnel / equipment intrusion frequency, operation time period changes, and construction progress nodes. Furthermore, the control attributes are not deeply coupled with the spatial topology of area positioning, and are only simple label binding, which cannot adapt to the highly dynamic safety control requirements of construction sites. Furthermore, existing technologies generally employ ray casting and the SAT split axis theorem for spatial collision verification. To ensure control effectiveness, fixed high-frequency full calculations are used for all configured areas. When hundreds of control areas and thousands of positioning terminals operate concurrently on a large-scale construction site, serious waste of computing power, system lag, and response delays occur. Moreover, it can only achieve simple "entry area triggering early warning" and cannot achieve deep coupling verification of spatial topology and time dimension. Conventional technical means cannot balance the core contradiction between "control accuracy" and "computing power consumption".

[0028] In view of this, this invention provides a construction site safety management method and related equipment based on regional positioning visualization. This solution solves the problem of accurate alignment of heterogeneous data by extracting and weighted matching feature points between local construction drawings and standard GIS base maps. Subsequently, the matching results are used to map construction details onto the geographic base map, and the topological structure of the safety area is extracted, achieving accurate conversion from physical space to digital space. Based on this, this solution introduces spatiotemporal regional coding, combining static safety boundaries with dynamic temporal risks. Finally, through a collision detection mechanism under spatiotemporal constraints, the system can accurately identify intrusion behaviors during high-risk periods, thereby triggering effective early warnings.

[0029] The construction site safety management method based on regional positioning visualization provided in this invention relates to the field of human-computer interaction and safety management technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle-mounted terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the construction site safety management method based on regional positioning visualization, but is not limited to the above forms.

[0030] Figure 1 This is an optional flowchart of a construction site safety management method based on regional positioning visualization provided in an embodiment of the present invention. Figure 1 The method may include, but is not limited to, steps S100 to S600: Step S100: Perform feature point extraction on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; Step S200: Based on the first original descriptor and the second original descriptor, perform a weighted feature point matching operation on the first partial construction drawing and the first standard GIS base map to obtain a finely matched point pair; Step S300: Based on the precise matching point pairs, the first partial construction drawing is mapped to the first standard GIS base map to obtain the second standard GIS base map; Step S400: Extract the security area boundary from the second standard GIS base map to generate the initial topological geometry; Step S500: Assign spatiotemporal region encoding to the initial topological geometry to obtain the target topological geometry; Step S600: Perform spatial collision detection on the target topological geometry and the dynamically moving object. When a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region coding, trigger a cross-module warning.

[0031] In step S100 of some embodiments, the system automatically extracts ORB feature points from the first partial construction drawing and the first standard GIS base map, thereby achieving matching between the two. Optionally, the first partial construction drawing can be a raster map exported from CAD or UAV imagery; extracting ORB feature points from the two images respectively includes keypoint detection and descriptor generation.

[0032] In some embodiments, step S100 may include, but is not limited to, steps S110 to S120: Step S110: Use the ORB algorithm to detect key points and generate descriptors for the first partial construction drawing to obtain the first key point and the corresponding first original descriptor. Step S120: Use the ORB algorithm to detect key points and generate descriptors on the first standard GIS base map to obtain the second key points and the corresponding second original descriptors.

[0033] In step S110 of some embodiments, the system employs the ORB algorithm specifically for feature parsing of the first partial construction drawing. Since the construction drawing mainly consists of numerous line segments, corner points, and specific legends, traditional corner point detection is easily affected by line interference. Therefore, the FAST (Features from Accelerated Segment Test) algorithm is used to quickly locate the first key point at high-density line intersections and primitive endpoints. Subsequently, the main direction of the feature point is determined by calculating the gray-level centroid of the key point's neighborhood, eliminating possible rotational deviations that may occur during the scanning or import of the construction drawing, and generating a first original descriptor with rotation invariance. This process transforms complex CAD vector information or raster images into binary feature vectors, preserving the structural details of the drawing while significantly reducing the computational complexity of subsequent matching.

[0034] In step S120 of some embodiments, feature extraction is performed on a first standard GIS base map. Given that GIS base maps contain complex geographic features (such as road networks, building outlines, and water system boundaries) and typically have multi-scale characteristics, the ORB algorithm is invoked to detect second keypoints on the constructed multi-scale image pyramid to ensure that geographic features at different scaling levels can be captured. Simultaneously, a second raw descriptor is generated using the BRIEF descriptor. Since GIS base maps typically have standard coordinate systems and projection information, the feature points extracted here focus more on reflecting the real geographic environment's topology, thus providing an accurate feature benchmark for subsequent map-geographic cross-modal matching and solving the problem of unified representation of heterogeneous data in terms of feature dimensions.

[0035] In step S200 of some embodiments, based on the original descriptors of the first partial construction drawing and the first standard GIS base map, a semantic priority weighting mechanism for key structures at the construction site (such as building outlines, equipment edges, road boundaries, etc.) is introduced to improve matching accuracy. For example, the semantic priority weighting mechanism assigns different weights to ORB feature points in different areas according to the semantic category of the key structures at the construction site. During the feature matching stage, when calculating the matching distance, the higher-weighted feature points have a higher matching priority, thereby improving the overall matching accuracy.

[0036] In some embodiments, step S200 may include, but is not limited to, steps S210 to S270: Step S210: Using a lightweight semantic segmentation model, key structural category identification operations are performed on the first partial construction drawing and the first standard GIS base map respectively to obtain the first identification result and the second identification result. Step S220: Based on the first recognition result, assign a first weight to the first feature points in the first original descriptor that belong to the preset high-value region, and assign a second weight to the second feature points in the first original descriptor that belong to the background region, to obtain the first weighted descriptor; Step S230: Based on the second recognition result, assign a first weight to the third feature point in the second original descriptor that belongs to the preset high-value region, and assign a second weight to the fourth feature point in the second original descriptor that belongs to the background region, to obtain the second weighted descriptor; Step S240: Obtain the Hamming distance between each first weighted descriptor and each second weighted descriptor; Step S250: Adjust the matching priority based on the weight; wherein the first weight is greater than the second weight; Step S260: Based on Hamming distance and matching priority, obtain coarse matching point pairs between the first weighted descriptor and the second weighted descriptor; Step S270: Using the multi-scale RANSAC algorithm, perform a mismatch point removal operation on the coarse matching point pairs to obtain fine matching point pairs.

[0037] In step S210 of some embodiments, the system pre-trains a lightweight semantic segmentation model. Using this model, key structural categories are identified in both the first partial construction drawing and the first standard GIS base map, yielding corresponding identification results. The identification results record whether the feature points in the original descriptor belong to a high-value area or a background area. Optionally, high-value areas may include, but are not limited to, building outlines, equipment edges, road boundaries, etc.

[0038] In step S220 of some embodiments, based on the first identification result, the first feature points in the first original descriptor that belong to high-value areas such as building outlines, equipment edges, and road boundaries are given a higher weight (i.e., the first weight, such as 1.5-2.0 times), while the second feature points in the first original descriptor that belong to the background area are given a lower weight (i.e., the second weight, such as 0.5-0.8 times).

[0039] In step S230 of some embodiments, based on the second recognition result, the third feature points in the second original descriptor that belong to high-value areas such as building outlines, equipment edges, and road boundaries are given a higher weight (i.e., the first weight, such as 1.5-2.0 times), while the fourth feature points in the second original descriptor that belong to the background area are given a lower weight (i.e., the second weight, such as 0.5-0.8 times).

[0040] In step S240 of some embodiments, a similarity metric is calculated on the feature vectors after weight enhancement. For example, each first weighted descriptor extracted from the construction drawing and each second weighted descriptor extracted from the GIS base map are traversed, and the Hamming distance between each pair is calculated using bitwise operations. The Hamming distance reflects the degree of difference between two binary descriptors at the bit positions; the smaller the value, the more similar the features. Since the descriptors at key points have already undergone weight gain, the Hamming distance calculation process essentially measures the similarity between the construction drawing pixel features with importance labels and the geographic information features in a high-dimensional feature space, providing accurate values ​​for subsequent filtering and ranking.

[0041] In step S250 of some embodiments, a weight-guided priority mechanism is introduced. For example, a matching rule is set: when multiple descriptor pairs have similar Hamming distances, the pair consisting of feature points with higher weights (e.g., located in core components or high-risk areas) is prioritized. This mechanism effectively combats interference features caused by the complexity of the GIS base map background, ensuring that high-value key structural information is locked in the early stages of matching, thus improving the robustness of the matching process.

[0042] In step S260 of some embodiments, a coarse matching operation is performed based on Hamming distance and matching priority. For example, a composite filtering condition is constructed: a pair of descriptors is only considered a valid coarse matching pair if the Hamming distance between them is below a preset threshold and they conform to a set weight priority logic (i.e., higher-weight feature points have higher matching priority). This process generates a preliminary set of correspondences, reducing a large number of feature combinations to a very small candidate range and ensuring high recall of high-weight features, thus laying a data foundation for subsequent precise matching.

[0043] In step S270 of some embodiments, a multi-scale RANSAC (random sampling consensus) algorithm is used to perform geometric constraint verification, thereby eliminating the mismatched points remaining in the coarse matching, and finally retaining only the fine matching point pairs with the smallest error and the strongest geometric consistency.

[0044] In step S300 of some embodiments, after removing mismatched points using the RANSAC algorithm, the affine transformation matrix between images is calculated. Through this transformation matrix, the pixel coordinates of the local construction drawing are automatically mapped to the geospatial coordinates of the standard GIS, achieving accurate overlay.

[0045] In some embodiments, step S300 may include, but is not limited to, steps S310 to S320: Step S310: Based on the precise matching point pairs, obtain the affine transformation matrix between the first partial construction drawing and the first standard GIS base map; Step S320: Based on the affine transformation matrix, map the pixel coordinates of the first local construction drawing to the geospatial coordinates of the first standard GIS base map to obtain the second standard GIS base map.

[0046] In step S310 of some embodiments, the high-precision correspondence of the precisely matched point pairs is used to solve for the spatial mapping parameters, constructing a mathematical model regarding spatial position offset, rotation, scaling, and shear deformation. Optionally, the least squares method is used to fit and calculate multiple sets of corresponding two-dimensional pixel coordinates to obtain a 3×3 affine transformation matrix. This matrix not only encapsulates the scale difference between the first local construction drawing and the first standard GIS base map, but also corrects the geometric distortion caused by different shooting angles or projection methods, thereby transforming discrete matching points into a continuous and unified mathematical mapping relationship, laying the geometric foundation for spatial alignment across data sources.

[0047] In step S320 of some embodiments, after solving the parameters of the geometric model, specific coordinate mapping and image fusion operations are performed. Specifically, based on the affine transformation matrix, coordinate transformation operations are performed on each pixel coordinate point in the first local construction drawing, stretching and aligning it from the independent drawing pixel coordinate system to the geospatial coordinate system of the first standard GIS base map. During this process, bilinear interpolation or nearest neighbor interpolation can be used to process pixel resampling to ensure image quality. The resulting second standard GIS base map accurately overlays the construction details onto the geographic base map, providing a precise spatial carrier for subsequent safety zone delineation.

[0048] For example, the coordinate mapping formula is: ; In the formula, Represents pixel coordinates; Represents geographic spatial coordinates; , , , Represents scaling and rotation parameters; , This represents the translation parameter.

[0049] In some embodiments, after mapping the pixel coordinates of the first partial construction drawing to the geospatial coordinates of the first standard GIS base map, fine-tuning can be performed using manually selected control points to make the coordinate calibration error less than or equal to 0.1 meters, thereby improving the adaptability of the heterogeneous base map in dynamic construction environments.

[0050] In step S400 of some embodiments, a human-machine collaborative boundary extraction model is proposed and deployed on an edge GIS computing node. First, the security sources on the second standard GIS base map are coarsely segmented, then fine-tuned through human-machine interaction. The user-tuned boundary node data is fed back to the edge AI model in real time for local parameter fine-tuning (using a few-sample incremental learning mechanism). This allows the model to quickly adapt to the specific occlusion patterns and lighting conditions of the current construction site, improving the accuracy of automatic extraction in similar subsequent scenarios. Finally, using the fine-tuned vector boundary nodes, combined with the influence range parameters of the security sources, the initial topological geometry of the dynamic security influence area is automatically generated. In some embodiments, step S400 may include, but is not limited to, steps S410 to S440: Step S410: In response to the first operation command on the security source in the second standard GIS base map, a preset semantic segmentation model is invoked to generate the initial polygon boundary. Step S420: Render the initial polygon boundary to generate an editable graphic with control nodes; Step S430: In response to a second operation command on a control node in the editable graphic, the initial polygon boundary is corrected to obtain the corrected polygon boundary; Step S440: Based on the parameters of the modified polygon boundary and the influence range of the security source, an initial topological geometry representing the boundary of the security area is generated using a geometric algorithm.

[0051] In step S410 of some embodiments, the user first provides seed points for the safety source by dot-matrix on the second standard GIS base map, or by dragging a rough rectangle around the safety source area on the second standard GIS base map. In response to the selected seed point or the dragged rectangle, the system invokes a preset image semantic segmentation model, using the seed point or rectangle as prior input, to automatically identify the outline of the target object (i.e., the safety source) and generate an initial polygon boundary. Optionally, the safety source may include, but is not limited to, tower cranes, deep foundation pits, material storage areas, etc.; the preset image semantic segmentation model may be a deep learning network based on an attention mechanism.

[0052] In step S420 of some embodiments, the automatically generated initial polygon boundary is vectorized and rendered on the front-end page as a Bézier curve or polygon with control nodes, resulting in... Figure 2 The editable graphic shown.

[0053] In step S430 of some embodiments, such as Figure 2 As shown, users can drag and drop the control nodes of the editable graphic to make local corrections to the boundaries of parts of the site that are obscured by scaffolding or in blind spots, thus obtaining the corrected polygonal boundaries.

[0054] In some embodiments, the obtained node data of the corrected polygon boundary is fed back to the image semantic segmentation model in real time, and a few-sample incremental learning mechanism is used to fine-tune the local parameters of the semantic segmentation model, thereby adapting to the specific occlusion mode and lighting conditions of the current construction site and improving the accuracy of automatic recognition and extraction of target object contours in subsequent similar scenes.

[0055] In step S440 of some embodiments, based on the corrected polygon boundary and combined with the influence range parameters of the safety source (such as equipment parameters like tower crane boom length), an initial topological geometry of the dynamic safety influence area is automatically generated, i.e., the geometric boundary representation of the safety area itself. This initial topological geometry is a polygonal boundary region generated through geometric calculations based on the location and influence range parameters of the safety source.

[0056] In step S500 of some embodiments, after the vectorization of the security area boundary is completed, spatiotemporal attribute binding and encoding encapsulation operations are performed on the initial topological geometry. This transforms the static geometric space boundary into a dynamic control unit with a time dimension. By assigning a unique spatiotemporal region code to the initial topological geometry, the geometry can carry and associate risk status information within a specific time period, ultimately outputting a target topological geometry with a unique identifier and time attributes.

[0057] In some embodiments, step S500 may include, but is not limited to, steps S510 to S520: Step S510: Add a time validity attribute to the initial topological geometry; the time validity attribute shall include at least a high-risk valid time period; Step S520 generates a spatiotemporal region code based on the initial topological geometry and the time validity attribute, and embeds the spatiotemporal region code into the initial topological geometry to obtain the target topological geometry.

[0058] In step S510 of some embodiments, in response to the user's configuration of safety risk rules, a time validity attribute is assigned to the initial topological geometry. For example, the risk level of the safety area within a specific time interval is set as the time validity attribute according to the construction plan. For instance, the operating radius of a tower crane is defined as a high-risk valid period from 9:00 AM to 11:00 AM daily, and the remaining time is low-risk or invalid. Associating this time validity attribute with the initial topological geometry clarifies when the geometry possesses high-risk early warning effectiveness, providing a data foundation for subsequent precise control.

[0059] In step S520 of some embodiments, a spatiotemporal region code that uniquely identifies the geometry and its time validity attribute is generated based on the spatial characteristics of the initial topological geometry and its bound time validity attribute. Subsequently, this code is formally written into the data structure of the initial topological geometry as an extended attribute field, so that the static initial topological geometry, which originally only contained point, line and surface coordinates, is encapsulated into a target topological geometry that integrates spatial location, time window and unique identifier.

[0060] In some embodiments, such as Figure 3 As shown, the generated target topological geometry can be displayed on the front-end page, and its risk level can be marked and displayed using different color blocks. For example, based on the time validity attribute (i.e., high-risk valid period) of the embedded target topological geometry, it is determined whether the area is currently in a high-risk period. If it is, according to preset color mapping rules (e.g., red, orange, and yellow corresponding to different risk levels), a specific highlight color and transparency are assigned to the filled surface of the target topological geometry, and it is then overlaid as a dynamic layer on the base map, thereby achieving an intuitive and hierarchical visualization of the physical space boundary and real-time risk status. Optionally, according to the Job Condition Hazard Assessment (LEC) method, a comprehensive risk score d is calculated by multiplying the accident probability L, the frequency of personnel exposure to hazardous environments E, and the severity of accident consequences C. (Reference) Figure 2 , Figure 3 Based on the range of 'd' values, five levels of operational risk can be identified, with a higher 'd' value indicating a higher risk. When 'd' ≥ 320, the risk level is extremely dangerous, and work cannot continue; this risk area is marked in red. When 160 ≤ 'd' < 320, the risk level is high risk, requiring immediate rectification; this risk area is marked in orange. When 70 ≤ 'd' < 160, the risk level is significant risk, requiring rectification; this risk area is marked in yellow. When 20 ≤ 'd' < 70, the risk level is moderate risk, requiring attention; this risk area is marked in blue. When 'd' < 20, the risk level is slightly risky but acceptable; this risk area is marked in green. This hierarchical visual display clearly defines the control requirements for corresponding areas and allows for timely response to rectification requests.

[0061] In step S600 of some embodiments, the location stream data of the dynamically moving object transmitted back by the IoT device at the construction site is accessed in real time. The spatial intersection operation between the static reference boundary of the target topological geometry and the real-time position of the dynamically moving object is periodically performed, and the collision detection result including the collision status flag and the current collision timestamp is output. If the boundary interference (collision) is detected, the system immediately extracts the time validity attribute bound in the spatiotemporal region code of the target topological geometry, reads the high-risk valid time period, and verifies whether the current collision timestamp falls within the time period. When the two match completely, it is determined to be a valid risk collision. Then, the preset instruction template is called, the hidden danger investigation module is linked to automatically generate a hidden danger work order, and the on-site sound and light alarm is triggered simultaneously for immediate warning.

[0062] In some embodiments, step S600 may include, but is not limited to, steps S610 to S650: Step S610: Obtain the position flow data of dynamically moving objects in the construction site; Step S620: Based on the position flow data and the static reference boundary of the target topological geometry, perform spatial intersection calculation between the dynamically moving object and the target topological geometry using the ray method or the separating axis theorem to obtain the collision detection result; the collision detection result includes the collision state and the current collision timestamp. Step S630: Obtain the time validity attribute from the spatiotemporal region encoding of the target topological geometry, and extract the high-risk valid time period from the time validity attribute; Step S640: When the collision state is that the dynamically moving object collides with the target topological geometry, and the current collision timestamp is within the high-risk valid time period, it is determined to be a valid risk collision. Step S650: In response to a valid risk collision, a preset instruction template is invoked to conduct hazard investigation and on-site audible and visual alarms.

[0063] In step S610 of some embodiments, location stream data of dynamically moving objects in the construction site are collected in real time. Optionally, the dynamically moving objects may include, but are not limited to, mobile machinery, personnel positioning tags and other on-site Internet of Things devices.

[0064] In step S620 of some embodiments, using the Ray Casting Algorithm or the Separating Axis Theorem (SAT), at a frequency of 5 frames per second, spatial intersection operations are performed on the dynamically moving object and the target topological geometry based on the position stream data and the static reference boundary. This yields collision detection results, including the collision state reflecting whether interference (collision) has occurred and the current collision timestamp. The safe zone polygon boundary is a static or semi-static spatial reference region (i.e., the static reference boundary of the target topological geometry), determined by the location and influence range of the safety source; the dynamically moving object is a real-time moving entity. The system uses spatial collision detection to determine whether the dynamically moving object enters or crosses the safe zone boundary, thereby determining whether to trigger a warning.

[0065] In step S630 of some embodiments, the spatiotemporal region code embedded in the target topological geometry is obtained, the time validity attribute in the code is extracted, and the high-risk valid time period in the time validity attribute is used for subsequent valid risk collision determination and early warning.

[0066] In steps S640 to S650 of some embodiments, when interference (collision) is detected between the topological boundary of a dynamically moving object and the boundary of the target topological geometry region, and the current timestamp matches the high-risk valid time period of the region, a cross-module early warning based on a preset instruction template (including linkage with the hidden danger investigation module and the on-site audible and visual alarm) is immediately triggered.

[0067] This invention also provides a construction site safety management and control system based on regional positioning visualization. The system follows a full-link evolution mode encompassing multi-source base map spatial fusion, human-machine collaborative boundary extraction and topology generation, spatiotemporal multi-dimensional attribute binding, dynamic spatial collision calculation and rule distribution, and cross-terminal millisecond-level response and closed-loop operation. Exemplarily, the entire system consists of the following five modules: The multi-source heterogeneous base map intelligent mapping and edge visual interaction module has built-in ORB feature matching and affine transformation algorithms to support dynamic mapping of heterogeneous maps to the standard GIS coordinate system; it integrates an edge AI semantic segmentation model and provides a human-computer collaborative high-precision boundary extraction tool with "outline coarse extraction, vector node fine-tuning, and user feedback-driven edge AI model adaptive optimization".

[0068] The safety quantitative assessment and spatiotemporal topology attribute configuration module is used to break the limitations of static attributes, establish a mapping between safety scores and three-dimensional models, and dynamically bind the geometric topological features of safety polygons with temporal validity to generate a unique spatiotemporal area code containing spatial, temporal, responsible person, and critical engineering attributes.

[0069] The dynamic linkage rule engine module based on spatial geometric collision serves as the core computing power hub of the platform. Using spatiotemporal region encoding as an index, it employs spatial intersection algorithms such as the Separating Axis Theorem (SAT) to calculate the spatiotemporal interference (collision) state between the coordinates of dynamic IoT devices on site and the boundaries of custom safety polygons in high frequency and real time, accurately triggering linkage early warning commands across modules (such as hidden danger investigation and on-site recording).

[0070] The high-concurrency cross-terminal collaborative computing power distribution and response module is used to optimize the system's computing power allocation and support the concurrent spatiotemporal collision detection of massive coordinate points. It achieves millisecond-level data distribution of secure polygon topology nodes and rules after adjustment, and drives second-level collaborative responses of visualization screens, mobile terminals, and on-site IoT alarm devices in parallel through message queues.

[0071] The end-to-end spatiotemporal data structured storage and traceability module is used to build a structured database that adapts to high-frequency dynamic coordinate updates and polygon vector data storage. It automatically records operation logs for the entire process from base map alignment and boundary fine-tuning to early warning distribution, ensuring efficient traceability of complex multi-module penetrating queries and spatiotemporal data.

[0072] This invention also provides a construction site safety management and control device based on regional positioning visualization, which can realize the above-mentioned construction site safety management and control method based on regional positioning visualization. The device includes: The feature point extraction module is used to perform feature point extraction operations on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; The feature point matching module is used to perform weighted feature point matching operations on the first partial construction drawing and the first standard GIS base map based on the first original descriptor and the second original descriptor to obtain finely matched point pairs; The spatial coordinate system mapping module is used to map the first local construction drawing to the first standard GIS base map based on the precise matching point pairs, so as to obtain the second standard GIS base map; The first dynamic topology generation module is used to extract the safe area boundaries from the second standard GIS base map and generate the initial topological geometry. The second dynamic topology generation module is used to assign spatiotemporal region encoding to the initial topological geometry to obtain the target topological geometry. The collision detection and early warning module is used to perform spatial collision detection between the target topological geometry and dynamically moving objects. When a collision is detected between a dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region encoding, a cross-module early warning is triggered.

[0073] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0074] This invention also provides an electronic device, which includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including a tablet computer, an in-vehicle computer, or similar device.

[0075] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0076] refer to Figure 4 , Figure 4 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 701 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention. The memory 702 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 702 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 702 and is called and executed by the processor 701. The input / output interface 703 is used to implement information input and output; The communication interface 704 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 705 transmits information between various components of the device (e.g., processor 701, memory 702, input / output interface 703, and communication interface 704); The processor 701, memory 702, input / output interface 703, and communication interface 704 are connected to each other within the device via bus 705.

[0077] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0078] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0079] This invention also provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0080] In summary, the construction site safety management method and related equipment based on regional positioning visualization according to embodiments of the present invention have the following advantages: 1. This invention, through the introduction of feature matching and affine transformation algorithms, achieves intelligent spatial alignment between multi-source local drawings such as CAD and UAV orthophotos and standard GIS base maps. Simultaneously, it constructs a human-machine collaborative boundary extraction mechanism driven by AI contour initial extraction, manual vector node fine-tuning, and user feedback-driven edge AI model adaptive optimization. This mechanism overcomes the limitations of purely automatic recognition technology in environments with severe occlusion and dynamic changes, replacing inefficient manual drawing and enabling rapid and accurate annotation of polygonal boundaries in safe areas. This enhances the ability to adapt heterogeneous base maps and achieve high-precision safe positioning in complex construction scenarios.

[0081] 2. In this embodiment of the invention, the extracted safe area polygon boundary is dynamically bound to the time validity. Using a spatial topology intersection algorithm, the spatial collision (interference) status between dynamic moving objects (personnel / mechanical equipment) and the custom safe area in real time is calculated. This constructs a highly sensitive spatiotemporal linkage early warning rule engine, which breaks through the traditional early warning threshold based on static distance or single attribute, thereby enhancing the geometric topology and spatiotemporal multi-dimensional perception capabilities of dynamic safety early warning.

[0082] 3. This invention uses spatiotemporal safety zone coding as the core association key, deeply integrating business modules such as monitoring video entry, hidden danger investigation, and critical engineering projects. When a dynamically moving object triggers spatiotemporal collision rules, it instantly generates and issues multi-dimensional control instructions containing associated monitoring images and response priorities, providing intuitive operational feedback and achieving second-level linkage response from early warning triggering to multi-terminal collaborative intervention, thereby strengthening the deep linkage and agile response loop across modules.

[0083] 4. According to the three-layer collaborative architecture of central platform scheduling, large screen monitoring and display, and mobile terminal on-site operation, the embodiments of the present invention rationally allocate complex spatial collision calculation and feature extraction, improve the system operation smoothness when there are multiple attribute changes and massive coordinate concurrent comparisons, realize efficient computing power scheduling and resource allocation in large-scale construction sites, and optimize the system computing power and terminal collaboration efficiency in highly dynamic and concurrent scenarios.

[0084] 5. The embodiments of the present invention construct a new technical closed loop from intelligent mapping of multi-source base maps, generation of human-computer interaction boundaries, early warning of spatiotemporal collision topology to real-time intervention of multiple terminals. This fundamentally solves the problems of large blind spots in complex scene recognition, rigid static configuration, and fragmented control data in the existing technology, greatly improving the level of intelligence and flexible configuration efficiency of safety management at construction sites, and realizing four-dimensional (space and time) precise safety management through human-machine collaboration.

[0085] 6. By using an optimized GIS coordinate calibration algorithm and manual drawing tools, the system achieves accurate matching between the safety zone and the actual on-site scene. Combined with a safety score quantitative assessment model (including dynamic weights), it solves the problems of ambiguous areas and subjective judgment of safety levels in traditional management and control, and effectively ensures the accuracy of safety marking.

[0086] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0087] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0088] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0089] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0090] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0091] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above 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 one or more embodiments or examples.

[0092] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0093] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A construction site safety management method based on regional positioning visualization, characterized in that, Includes the following steps: Feature point extraction is performed on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; Based on the first original descriptor and the second original descriptor, a weighted feature point matching operation is performed on the first partial construction drawing and the first standard GIS base map to obtain a finely matched point pair; Based on the precisely matched point pairs, the first partial construction drawing is mapped to the first standard GIS base map to obtain the second standard GIS base map; The security area boundaries are extracted from the second standard GIS base map to generate the initial topological geometry; Spatiotemporal region encoding is assigned to the initial topological geometry to obtain the target topological geometry; Spatial collision detection is performed between the target topological geometry and the dynamically moving object. When a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region encoding, a cross-module warning is triggered.

2. The method according to claim 1, characterized in that, The step of extracting feature points from the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor includes the following steps: The ORB algorithm is used to detect key points and generate descriptors for the first partial construction drawing, so as to obtain the first key point and the corresponding first original descriptor. The ORB algorithm is used to detect key points and generate descriptors on the first standard GIS base map to obtain the second key point and the corresponding second original descriptor.

3. The method according to claim 1, characterized in that, The step of performing weighted feature point matching on the first partial construction drawing and the first standard GIS base map based on the first original descriptor and the second original descriptor to obtain finely matched point pairs includes the following steps: Using a lightweight semantic segmentation model, key structural category identification operations are performed on the first partial construction drawing and the first standard GIS base map respectively to obtain the first identification result and the second identification result. Based on the first identification result, a first weight is assigned to the first feature point in the first original descriptor that belongs to the preset high-value region, and a second weight is assigned to the second feature point in the first original descriptor that belongs to the background region, to obtain a first weighted descriptor; Based on the second identification result, the third feature point belonging to the preset high-value region in the second original descriptor is assigned a first weight, and the fourth feature point belonging to the background region in the second original descriptor is assigned a second weight to obtain the second weighted descriptor; Obtain the Hamming distance between each of the first weighted descriptors and each of the second weighted descriptors; Matching priority is adjusted based on weight magnitude; wherein, the first weight is greater than the second weight; Based on the Hamming distance and the matching priority, coarse matching point pairs between the first weighted descriptor and the second weighted descriptor are obtained; By using the multi-scale RANSAC algorithm, mismatched points are removed from the coarse matching point pairs to obtain fine matching point pairs.

4. The method according to claim 1, characterized in that, The process of mapping the first partial construction drawing to the first standard GIS base map based on the precisely matched point pairs to obtain the second standard GIS base map includes the following steps: Based on the precisely matched point pairs, obtain the affine transformation matrix between the first partial construction drawing and the first standard GIS base map; Based on the affine transformation matrix, the pixel coordinates of the first local construction drawing are mapped to the geospatial coordinates of the first standard GIS base map to obtain the second standard GIS base map.

5. The method according to claim 1, characterized in that, The step of extracting the security area boundary from the second standard GIS base map and generating the initial topological geometry includes the following steps: In response to the first operation command on the security source in the second standard GIS base map, a preset semantic segmentation model is invoked to generate the initial polygon boundary; The initial polygon boundary is rendered to generate an editable graphic with control nodes; In response to a second operation command on a control node in the editable graphic, the initial polygon boundary is corrected to obtain a corrected polygon boundary. Based on the modified polygon boundary and the influence range parameters of the security source, the initial topological geometry representing the boundary of the security area is generated by a geometric algorithm.

6. The method according to claim 1, characterized in that, The process of assigning spatiotemporal region encoding to the initial topological geometry to obtain the target topological geometry includes the following steps: Add a time validity attribute to the initial topological geometry; the time validity attribute includes at least the high-risk valid time period; Based on the initial topological geometry and the time validity attribute, a spatiotemporal region code is generated, and the spatiotemporal region code is embedded in the initial topological geometry to obtain the target topological geometry.

7. The method according to claim 1, characterized in that, The step of performing spatial collision detection on the target topological geometry and the dynamically moving object, and triggering a cross-module warning when a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches a high-risk valid time period in the spatiotemporal region encoding, includes the following steps: Acquire the position stream data of the dynamically moving objects at the construction site; Based on the location flow data and the static reference boundary of the target topological geometry, a spatial intersection operation is performed between the dynamically moving object and the target topological geometry using the ray method or the separating axis theorem to obtain a collision detection result; the collision detection result includes the collision state and the current collision timestamp; Obtain the time validity attribute from the spatiotemporal region code of the target topological geometry, and extract the high-risk valid time period from the time validity attribute; When the collision state is that the dynamically moving object collides with the target topological geometry, and the current collision timestamp is within the high-risk valid time period, it is determined to be a valid risk collision; In response to the effective risk collision, a preset instruction template is invoked to conduct hazard investigation and on-site audible and visual alarms.

8. A construction site safety management and control device based on regional positioning visualization, characterized in that, include: The feature point extraction module is used to perform feature point extraction operations on the first partial construction drawing and the first standard GIS base map to obtain the first original descriptor and the second original descriptor; The feature point matching module is used to perform a weighted feature point matching operation on the first partial construction drawing and the first standard GIS base map based on the first original descriptor and the second original descriptor to obtain finely matched point pairs; The spatial coordinate system mapping module is used to map the first local construction drawing to the first standard GIS base map based on the precise matching point pair to obtain the second standard GIS base map; The first dynamic topology generation module is used to extract the safe area boundary of the second standard GIS base map and generate the initial topological geometry. The second dynamic topology generation module is used to assign spatiotemporal region encoding to the initial topology geometry to obtain the target topology geometry. The collision detection and early warning module is used to perform spatial collision detection between the target topological geometry and the dynamically moving object. When a collision is detected between the dynamically moving object and the target topological geometry, and the current collision timestamp matches the high-risk valid time period in the spatiotemporal region encoding, a cross-module early warning is triggered.

9. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images