CN-121998243-A - Comprehensive intelligent management method, system and storage medium for building construction site

CN121998243ACN 121998243 ACN121998243 ACN 121998243ACN-121998243-A

Abstract

The invention discloses a comprehensive intelligent management method, a system and a storage medium for a construction site, which relate to the technical field of intelligent management of construction, and comprise the steps of generating a space distribution density field based on Gaussian kernel density estimation of UWB label positions and wireless signal inversion, generating coarse granularity density through fusion, and performing smoothing treatment to obtain a final density field; and generating an expected density field according to the working condition plan, carrying out plan matching on the entity state set, calculating the plan conformity field and the plan deviation degree, and generating a comprehensive risk field based on the plan conformity field complement and the deviation degree. According to the invention, by combining the wireless signal inversion and UWB Gaussian kernel density estimation, the spatial recognition precision of the distribution of construction site personnel and equipment is improved, and the real-time performance and scene self-adaptability of risk recognition are improved by combining the combination calculation of the plan coincidence degree field and the plan deviation degree with the static risk field superposition.

Inventors

LI JIANWEI
WU ZHUYE
CHEN XIANYING
Dun Jianbei
GAO ZHIPING
HUANG SHIYING
YANG LEI
Ba Zhanqiang
WANG LELE
LI JIANGPENG
ZHAO LIYONG

Assignees

河北诺亚能源股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260119

Claims (10)

1. The comprehensive intelligent management method for the construction site is characterized by comprising the following steps: Collecting multi-source heterogeneous data and working condition plans of a building construction site, and preprocessing, wherein the multi-source heterogeneous data comprises visual images, three-dimensional point clouds, UWB tag positions and wireless radio frequency signal data; gaussian kernel density estimation based on UWB label position and wireless signal inversion generate a space distribution density field, coarse granularity density is generated through fusion, smoothing processing is carried out, and a final density field is obtained; Performing Monte Carlo sampling based on the final density field, generating a final motion field by combining optical flow calculation and Gaussian smoothing, dividing the final density field into connected areas, calculating the mass center and average speed of each area based on the final motion field to form an entity state set, generating a desired density field according to a working condition plan, performing plan matching on the entity state set, and calculating plan conformity degree field and plan deviation degree; And generating a comprehensive risk field based on the plan coincidence degree field complement and the deviation degree, setting a risk threshold value to identify a high risk area and triggering early warning.
2. The method for comprehensive intelligent management of building construction sites according to claim 1, wherein the step of generating a spatial distribution density field by Gaussian kernel density estimation based on UWB tag positions and wireless signal inversion, generating coarse-grained density by fusion, and obtaining a final density field by smoothing comprises the steps of: Building a construction ground coordinate system based on the construction total plane graph, and dividing the construction ground coordinate system into regular two-dimensional grids; For each grid point, converting the UWB label position into continuous density distribution by adopting a Gaussian kernel density estimation method, inverting the spatial distribution of a signal source by adopting inverse distance weighted approximation of a logarithmic distance path loss model, and respectively carrying out maximum-minimum normalization processing and fusion on the density distribution and the spatial distribution to generate fused coarse-granularity density; Smoothing the fused coarse-grain density by using Gaussian filtering to obtain a final density field; setting a density threshold, screening a region with coarse grain density larger than the density threshold, and defining the region as a region of interest; setting a foreground detection threshold value, extracting operation foreground pixels of the visual image in the region of interest, and generating an operation foreground mask.
3. The method for comprehensive intelligent management of a building construction site according to claim 2, wherein the performing Monte Carlo sampling based on the final density field, generating a final motion field by combining optical flow calculation and Gaussian smoothing, dividing the final density field into connected regions, and based on the final motion field, comprises: Carrying out normalization processing on the final density field, taking the normalized final density field as a probability density function, carrying out Monte Carlo sampling in a construction plane, finding out the projection position of a sampling point in each visual image, calculating the optical flow displacement of the projection position between a current frame and a previous frame, carrying out back projection on the displaced sampling point to a ground plane, calculating the ground speed of the sampling point under the observation of a camera, calculating the arithmetic average of the ground speed, and defining the arithmetic average as the optical flow speed; The projection position is used for carrying out internal parameter calibration on each high-definition network camera through a Zhang Zhengyou calibration method, carrying out external parameter calibration on the three-dimensional point cloud through a PNP algorithm, and projecting the operation foreground mask to a ground coordinate system based on the internal parameters and the external parameters; Calculating the instantaneous speed of the UWB tag of each sampling point, combining the optical flow speed and the instantaneous speed to obtain a sparse speed field, calculating the speed vector of each regular grid point in the construction ground coordinate system by adopting an inverse distance weighted interpolation method based on the sparse speed field, and performing Gaussian smoothing to obtain a final sports field.
4. The method for comprehensive intelligent management of a building construction site according to claim 3, wherein calculating the mass center and the average speed of each region to form a solid state set, generating a desired density field according to a working condition plan, performing plan matching on the solid state set, and calculating a plan conformity field and a plan deviation degree comprises: Identifying local density maximum points on the final density field by using a peak value searching algorithm, defining the local density maximum points as density peak points, dividing the final density field into connected areas, calculating the mass center and the average speed of the connected areas, and combining the mass center and the average speed into a solid state set; calculating expected density fields of each plan by adopting a two-dimensional Gaussian model at any grid, and obtaining a total plan expected density field by linearly superposing the expected density fields of all plans; traversing the entity state set, judging whether the centroid position of each entity falls within the working range of any plan, if the centroid position meets the condition, marking the entity as an 'intra-plan entity', otherwise marking the entity as an 'out-of-plan entity'; The operation range refers to that the Euclidean distance between the centroid position and the planned operation position is smaller than or equal to the radius of the planned operation range, and the current moment is in the planned execution time window; A plan compliance field is calculated based on the marking results and the total desired density field.
5. The method for comprehensive intelligent management of a building construction site according to claim 4, wherein the generating the comprehensive risk field based on the plan compliance field complement and the deviation degree comprises: calculating the relative absolute deviation between the final density field and the total plan expected density field, generating a plan deviation degree, and calculating a complement of the plan compliance field, which is defined as a logic conflict degree; for each grid point in the construction ground coordinate system, defining a local space neighborhood by taking the local space neighborhood as the center, calculating sample covariance matrixes of speed vectors at all grid points in the local space neighborhood, carrying out eigenvalue decomposition, and defining the obtained maximum eigenvalue as the motion confusion of the grid point; Normalizing the plan deviation degree, the logic conflict degree and the motion confusion degree, and performing linear superposition to obtain a basic dynamic risk field; The method comprises the steps of (1) deriving space coordinate information of all geometric primitives marked as dangerous sources from a construction information model designed by construction organization, dividing the ground of the whole construction site into two-dimensional grids identical to a construction ground coordinate system, creating a blank binary matrix, marking all grid units covered by each dangerous source geometric primitive as1 in the binary matrix, generating a static dangerous source binary matrix aligned with the construction ground coordinate system, and defining a static dangerous field; and combining the static risk field and the basic dynamic risk field to generate a comprehensive risk field.
6. The method for comprehensive intelligent management of a building construction site according to claim 5, wherein the generating a comprehensive risk field based on the plan compliance field complement and the deviation, setting a risk threshold to identify a high risk area and triggering early warning comprises: setting a risk threshold, performing binary segmentation on the comprehensive risk field by using the risk threshold, marking the comprehensive risk field larger than the risk threshold as 1, otherwise marking the comprehensive risk field as 0, and generating an initial risk mask; and performing morphological closing operation on the initial risk mask to eliminate noise and connect adjacent areas to obtain a binary risk mask, aggregating the binary risk mask by using a connected component analysis algorithm to generate a connected domain, and sending out early warning and notifying area responsible persons.
7. The method for comprehensive intelligent management of a building construction site according to claim 1, wherein the steps of collecting multi-source heterogeneous data and working condition plans of the building construction site and preprocessing the multi-source heterogeneous data and working condition plans comprise: collecting multi-source heterogeneous data of a building construction site by using an intelligent sensor, collecting a working condition plan from a construction organization design by using an API interface, and performing time stamp synchronization, denoising and standardization treatment; The intelligent sensor comprises a high-definition network camera, a laser radar, a UWB positioning system and a Wi-Fi probe sensor; the working condition plan comprises a working center coordinate, a working range radius, a plan execution time window and an expected entity number.
8. The comprehensive intelligent management system for the construction site is used for realizing the comprehensive intelligent management method for the construction site, which is characterized by comprising the following steps: The collection processing module is used for collecting multi-source heterogeneous data and working condition plans of a building construction site and carrying out time stamp synchronization, denoising and standardization processing; The density fusion module is used for establishing a construction ground coordinate system and a regular grid, generating a space density field based on UWB Gaussian kernel density estimation and Wi-Fi path loss inversion, forming coarse granularity density distribution through weighted fusion, and performing smoothing treatment to obtain a refined final density field; the motion entity module is used for generating a sparse speed field through optical flow analysis and UWB speed calculation, interpolating to form a continuous motion field, and identifying a construction site entity and a state thereof by utilizing a watershed algorithm; The risk early warning module is used for matching the entity state with the planned expected density field, calculating the plan conformity, the deviation degree and the movement confusion degree, generating a basic dynamic risk field, combining static risk source information and dynamic risk fields derived by BIM, generating a comprehensive risk field and carrying out risk threshold detection and region early warning.
9. A computer device comprising a memory and a processor, said memory storing a computer program, characterized in that said processor, when executing said computer program, implements the steps of the method for integrated intelligent management of construction sites according to any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method for comprehensive intelligent management of a construction site according to any one of claims 1 to 7.

Description

Comprehensive intelligent management method, system and storage medium for building construction site Technical Field The invention relates to the technical field of intelligent management of building construction, in particular to a comprehensive intelligent management method, a system and a storage medium for a building construction site. Background With the continuous advancement of the digitization and the intelligent transformation of the construction industry, the management mode of the construction site gradually evolves from an artificial experience decision to a data-driven intelligent decision, and the application of technologies such as the internet of things (IoT), artificial Intelligence (AI), computer Vision (CV), wireless positioning (such as UWB and Wi-Fi probes) and the like provides new possibilities for real-time monitoring and dynamic management of the construction site. The existing construction site intelligent management technology still has obvious defects, visual detection results are easily influenced by shielding, illumination and environmental interference, and a complementary mechanism with wireless signal data is lacking, so that macroscopic coverage and microscopic precision are difficult to be simultaneously considered. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the invention provides a comprehensive intelligent management method, a system and a storage medium for a building construction site, which solve the problems that visual detection results are easily influenced by shielding, illumination and environmental interference, and a complementary mechanism with wireless signal data is lacking, so that macroscopic coverage and microscopic precision are difficult to be simultaneously considered. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, the invention provides a method for comprehensive intelligent management of a building construction site, which comprises the following steps: Collecting multi-source heterogeneous data and working condition plans of a building construction site, and preprocessing, wherein the multi-source heterogeneous data comprises visual images, three-dimensional point clouds, UWB tag positions and wireless radio frequency signal data; gaussian kernel density estimation based on UWB label position and wireless signal inversion generate a space distribution density field, coarse granularity density is generated through fusion, smoothing processing is carried out, and a final density field is obtained; Performing Monte Carlo sampling based on the final density field, generating a final motion field by combining optical flow calculation and Gaussian smoothing, dividing the final density field into connected areas, calculating the mass center and average speed of each area based on the final motion field to form an entity state set, generating a desired density field according to a working condition plan, performing plan matching on the entity state set, and calculating plan conformity degree field and plan deviation degree; And generating a comprehensive risk field based on the plan coincidence degree field complement and the deviation degree, setting a risk threshold value to identify a high risk area and triggering early warning. The invention relates to a construction site comprehensive intelligent management method, which comprises the following steps of generating a space distribution density field by Gaussian kernel density estimation based on UWB label positions and wireless signal inversion, generating coarse granularity density by fusion, and obtaining a final density field by smoothing, wherein the method comprises the following steps: Building a construction ground coordinate system based on the construction total plane graph, and dividing the construction ground coordinate system into regular two-dimensional grids; For each grid point, converting the UWB label position into continuous density distribution by adopting a Gaussian kernel density estimation method, inverting the spatial distribution of a signal source by adopting inverse distance weighted approximation of a logarithmic distance path loss model, and respectively carrying out maximum-minimum normalization processing and fusion on the density distribution and the spatial distribution to generate fused coarse-granularity density; Smoothing the fused coarse-grain density by using Gaussian filtering to obtain a final density field; setting a density threshold, screening a region with coarse grain density larger than the density threshold, and defining the region as a region of interest; setting a foreground detection threshold value, extracting operation foreground pixels of the visual image in the region of interest, and generating an operation foreground mask. The invention provides a method for comprehensively and intellig