Cloud edge coordination management and control platform and method

Abstract

The invention discloses a cloud edge coordination management and control platform and method, and belongs to the technical field of unmanned aerial vehicle intelligent inspection and construction modeling. The platform comprises an unmanned aerial vehicle device carrying a millimeter wave radar and an infrared multispectral imager, an edge control layer and a cloud management and control center. The unmanned aerial vehicle collects terrain and heat source data according to a preset path, the edge side preprocesses the data by using a Gaussian splash engine, and a DSM coarse model is generated and uploaded to the cloud; and the cloud side generates a DSM fine model through error correction, further generates a construction thermodynamic diagram, and optimizes a path and an obstacle avoidance strategy template in combination with historical flight data. The scheme has a stable sensing capability in a complex environment, and improves the operation stability and modeling precision in environments such as haze and smoke dust; meanwhile, edge cloud cooperative processing is realized, the data transmission burden is effectively reduced, and the system response efficiency is improved; and in combination with DSM fine model and thermodynamic diagram output, a construction party is supported to visually manage and make decisions on earthwork volume change and operation distribution.

Description

Technical Field

[0001] The present application relates to the field of cloud-edge collaborative computing, and in particular to a cloud-edge coordination and management platform and method. Background Art

[0002] Drones are increasingly being used in construction, infrastructure maintenance, and earthwork monitoring. Traditionally, drones primarily capture imagery and conduct aerial modeling. Using onboard visual and infrared sensors, they collaborate with ground-based processing systems to generate three-dimensional models (such as 3D projection models or DSMs) for construction progress analysis, quantity calculations, and efficiency assessments.

[0003] To cope with the dynamic and complex environments of construction sites, some solutions have incorporated edge processing modules, such as GPU-embedded terminals or lightweight AI processors, to perform some image preprocessing and feature extraction, reducing the burden of data transmission. Some products also attempt to use fixed route templates combined with scheduled inspections to capture topographic changes in the construction area at different points in time, assisting in calculating earthwork volume changes and construction efficiency indicators.

[0004] However, these systems commonly suffer from the following bottlenecks: 1. Single perception methods and poor environmental adaptability: Existing drones mostly rely on visible light vision systems and conventional infrared thermal imaging equipment. When encountering low visibility conditions such as haze, smoke, and rain, sensor acquisition quality deteriorates significantly, resulting in blurred images, target recognition failures, and false alarms. 2. Inadequate obstacle avoidance capabilities, posing flight safety risks: Traditional path planning often uses static presets. During flight, drones lack the ability to perceive and avoid high-altitude obstacles (such as cranes, cables, and scaffolding), making accidental collisions particularly likely. 3. Point cloud modeling is significantly susceptible to interference, resulting in unstable DSM accuracy: Although millimeter-wave radar has penetrating capabilities, in environments with high PM2.5 concentrations and high humidity, signal scattering is severe, resulting in a low echo signal-to-noise ratio. The generated point cloud data is prone to noise and holes, affecting the accuracy of the DSM model. 4. High edge computing load and insufficient real-time performance: While unilaterally deployed lightweight AI engines can partially process image data, computing power conflicts exist between tasks such as image restoration, anomaly detection, and obstacle avoidance. This increases system response latency, making it difficult to handle the collaborative tasks of complex perception and decision-making. 5. Path rigidity and lack of learning capabilities: Existing path planning templates are mostly static configurations, lacking dynamic learning and self-update mechanisms. They cannot adaptively adjust flight paths based on changes in on-site obstacles and environmental disturbances, severely limiting the system's intelligence. Summary of the Invention

[0005] The purpose of this application is to solve the above-mentioned problems of single perception means and poor environmental adaptability; insufficient obstacle avoidance ability and flight safety risks; high edge computing load and insufficient real-time performance; strong path rigidity and lack of learning ability.

[0006] According to one aspect of the present application, a cloud-edge coordination and management platform is provided, including: The drone is equipped with a radar host and an infrared slave. The radar host is used to scan the target area and obtain scanning data; the infrared slave is used to lock onto the heat source area generated by the mechanical equipment, take detailed photos, and obtain detailed data. An edge control layer, provided on the UAV device, has a built-in lightweight Gaussian splash engine for pre-processing the scan data and detail data, marking abnormal areas where the image signal-to-noise ratio is lower than a preset threshold, and generating a DSM rough model; The cloud-based control center is used to receive the DSM rough model and optimize the DSM rough model through the deployed engineering perception data intelligent processing algorithm set to generate a DSM fine model; the cloud-based control center further calculates the earthwork volume of the target area based on the DSM fine model and generates a construction heat map.

[0007] Preferably, the radar host is a millimeter-wave radar, which is used to penetrate suspended particles in the air in haze or low-visibility environments and sense terrain contour information; the millimeter-wave radar is combined with an acoustic wave array system, which includes multiple acoustic wave transmitting and receiving units, and is used to assist in calibrating the spatial positioning results of the millimeter-wave radar; The infrared slave is a multispectral imager that operates in the 8-14 μm band and is used to obtain image information in different bands. The multispectral imager has a built-in infrared reflectivity database of metal materials to improve the multispectral imager's ability to identify the reflective characteristics of different metal surfaces. The integrated PM2.5, temperature, humidity and pressure sensor is used to detect the concentration of fine particulate matter, temperature, humidity and air pressure parameters in the environment.

[0008] The present invention also provides a cloud-edge coordination and control method, which is applied to the above-mentioned cloud-edge coordination and control platform, comprising: Generate and issue a UAV cruise path, control the UAV to cruise along the cruise path, and perform terrain scanning on the target area to collect terrain data and generate a rough DSM model; The DSM rough model is pre-processed through the edge control layer and uploaded to the cloud control center; The DSM rough model is error-corrected by the engineering perception data intelligent processing algorithm deployed by the cloud control center to generate a DSM fine model; A construction heat map is generated based on the DSM fine mold.

[0009] Preferably, the collecting of terrain data includes: obtaining terrain contour data by scanning with a radar host, and obtaining a heat source area image by scanning with an infrared slave.

[0010] Preferably, the radar host includes a millimeter wave radar and an acoustic wave array system. Accordingly, controlling the UAV to cruise on the cruise path includes: The millimeter wave radar transmits a frequency modulated continuous wave in the target area to obtain terrain contour data; Transmitting ultrasonic pulses to the target area through the acoustic wave array system and receiving echo signals after encountering obstacles; Calculate the spatial position of obstacles based on the sound wave flight time and temperature and humidity correction model; Performing spatiotemporal alignment of obstacle position data measured by the acoustic array system with millimeter-wave radar data to calculate ranging errors; Based on the ranging error, the positioning error of the millimeter wave radar is determined and corrected.

[0011] Preferably, the determining and correcting the positioning error of the millimeter wave radar includes: Adding a unified timestamp to the obstacle distance data obtained by the millimeter-wave radar and the acoustic wave array system, and converting the coordinates of the acoustic wave data into the coordinate system of the millimeter-wave radar; Calculate the deviation between the obstacle distance data obtained by the millimeter wave radar and the acoustic wave array system:

[0012] Get the PM2.5 concentration value at the corresponding time point and determine whether any of the following conditions are met: Δd exceeds the preset error threshold; The PM2.5 concentration is higher than the preset standard for dense fog; If any of the above conditions is met, the compensation mechanism will be triggered: If the haze interference is serious, the acoustic wave coordinates are directly used to replace the corresponding point cloud coordinates; If the interference level is moderate, the radar and acoustic wave coordinates are weighted and fused according to the preset weights to generate the corrected point cloud coordinates:

[0013] The corrected point cloud data is used as the coordinate data for cruising.

[0014] Preferably, the pre-processing of the DSM block mold by the edge control layer comprises: Abnormal area marking, by performing signal-to-noise ratio analysis on the image or point cloud data in the DSM rough model, identifying areas where the signal-to-noise ratio is lower than a preset threshold, and marking the abnormal area; Data enhancement processing, performing quality optimization operations on the DSM rough model data, including but not limited to noise filtering, detail sharpening and contrast adjustment; Data compression processing is to compress the pre-processed DSM rough model data.

[0015] Preferably, the engineering perception data intelligent processing algorithm is a haze scattering error compensation algorithm. Accordingly, the error correction includes: Collect real-time PM2.5 concentration and atmospheric humidity; Based on the haze forecast model, the height offset of the point cloud is predicted; The height offset of the point cloud is used to correct the original point cloud height and adjust the height data in the point cloud.

[0016] Preferably, the height offset of the predicted point cloud includes: The point cloud height is corrected using the following error correction formula:

[0017] Where, is the construction site dust coefficient, is the scattering attenuation factor, is atmospheric humidity, and the construction site dust coefficient and the scattering attenuation factor are obtained through historical data training.

[0018] Preferably, the method further comprises: Identify potential flight risk areas and obstacles within the construction area based on the latest terrain morphology and obstacle spatial distribution information contained in the DSM; Comparing and analyzing the DSM fine model with historical flight trajectories and obstacle avoidance records to form an updated path planning solution; The path planning scheme and the corresponding obstacle avoidance strategy are sent to the drone in the form of a template.

[0019] The cloud-edge coordinated management and control platform and method proposed in this application enhance environmental adaptability by building multimodal perception. This embodiment utilizes a combined perception mechanism of a radar master and infrared slaves to significantly enhance the drone's perception of complex environments (such as smoke, haze, and obstructions), improving operational stability. Modeling accuracy is improved by integrating a Gaussian splattering engine into the edge control layer, enabling image quality pre-screening and anomaly area annotation, effectively preventing low-quality data from impacting DSM construction. The cloud further optimizes rough model errors, ensuring the final DSM refinement model is closer to the actual terrain. Data is processed at the edge before being uploaded to the cloud, significantly reducing redundant transmission of raw images and sensor data, shortening return latency, and improving overall system responsiveness. Construction data visualization is supported by generating a DSM refinement model and construction heat map, enabling dynamic tracking of earthwork volume trends, providing intuitive decision support for construction teams.

[0020] Furthermore, this solution also provides accurate and efficient data collection and modeling capabilities. The drone can perform tasks without human intervention and can autonomously perform path scanning, quickly complete terrain data collection and preliminary modeling, and significantly improve inspection efficiency. The stable perception capability in complex environments such as haze, through the collaborative perception of millimeter-wave radar and acoustic wave array, and edge-end image enhancement processing, effectively alleviates the interference of low-visibility environments on modeling accuracy and improves the quality of DSM rough models. Cloud-based intelligent optimization ensures accuracy. After error correction by the cloud-based control center, the DSM rough model forms a high-precision DSM fine model, providing a solid foundation for the generation of construction heat maps and avoiding data distortion caused by noise errors in traditional modeling. The data support for construction management decisions is enhanced. As a layer that intuitively reflects the intensity and distribution of construction activities, the construction heat map provides reliable data support for construction units to optimize project scheduling, adjust resource allocation, and evaluate project quantities. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the implementation methods of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the implementation methods or the description of the prior art. Obviously, the drawings described below are only some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a logical block diagram of the cloud-edge coordination and control method described in one embodiment of the present application. DETAILED DESCRIPTION

[0023] To facilitate understanding of the present application, a more comprehensive description of the present application will be provided below with reference to the accompanying drawings. The accompanying drawings illustrate preferred embodiments of the present application. However, the present application may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the disclosure of the present application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein in the specification of this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] Please refer to Figure 1 , an embodiment of the present application provides a cloud-edge coordination and management platform, including drone equipment, an edge control layer and a cloud-based management and control center.

[0026] The drone equipment is equipped with a radar host and an infrared slave. The radar host is used to scan the terrain of the target area and obtain scanning data; the infrared slave is used to lock the heat source area generated by the mechanical equipment, take detailed photos, and obtain detailed data.

[0027] The edge control layer is set on the drone equipment and has a built-in lightweight Gaussian splash engine, which is used to pre-process the scanning data and detail data, mark abnormal areas where the image signal-to-noise ratio is lower than the preset threshold, and generate a DSM rough model.

[0028] The cloud-based control center is used to receive the DSM rough model and optimize it through the deployed engineering perception data intelligent processing algorithm set to generate a DSM fine model; the cloud-based control center further calculates the earthwork volume of the target area based on the DSM fine model and generates a construction heat map.

[0029] In this embodiment, it should be noted that the cloud-edge coordination and control platform consists of three main components, namely drone equipment, edge control layer and cloud control center, which work together to achieve high-precision perception and dynamic modeling of the construction site.

[0030] Specifically, drone equipment is equipped with two types of sensing units: The radar host is preferably a combined system integrating millimeter-wave radar and acoustic arrays, used to scan the target area during flight. The millimeter-wave radar has strong penetrating capabilities and can identify the outline of terrain objects in lightly obstructed or low-visibility environments; the acoustic array is used to assist in obstacle positioning and correction, improving ranging accuracy.

[0031] Infrared slaves are used to locate high heat source areas and are particularly suitable for identifying continuously operating construction machinery such as tower cranes, excavators, and concrete mixing plants, thereby assisting in detailed modeling or thermal map analysis.

[0032] The edge control layer serves as the local computing unit of the drone device and integrates a lightweight Gaussian splash engine. This module has edge AI image processing capabilities and can complete the following tasks before data is uploaded: Image quality assessment and abnormal area annotation, such as image frames with signal-to-noise ratios below a set threshold; Data compression and enhancement preprocessing, including preliminary feature extraction and filtering and denoising; Based on the above processing, a DSM rough model (digital surface model) is generated as the basic content for uploading structured data to the cloud.

[0033] After receiving the DSM rough model data from the edge control layer, the cloud control center further optimizes the errors and abnormal areas in the DSM model through the deployed engineering perception data intelligent processing algorithm set, and generates a DSM fine model that meets the actual terrain accuracy requirements. The cloud platform also has: The ability to perform high-precision earthwork volume calculations on DSM fine models, and quantify incoming and outgoing volumes based on historical model changes; Generate construction heat maps based on terrain data and heat source distribution to assist in construction progress management, resource scheduling, and efficiency analysis.

[0034] Implementing the technical solutions of this embodiment has the following technical benefits: Enhanced environmental adaptability: Through the combined perception mechanism of the radar master and infrared slaves, this embodiment significantly enhances the drone's perception of complex environments (such as smoke, haze, and obstructions), improving operational stability. Improved modeling accuracy: The edge control layer integrates a Gaussian splash engine to pre-screen image quality and anomaly area annotation, effectively preventing low-quality data from impacting DSM construction. The cloud further optimizes rough model errors, making the final DSM fine model more accurate to the actual terrain. Reduced transmission burden and improved real-time performance: Data is processed on the edge before being uploaded to the cloud, significantly reducing redundant transmission of raw images and sensor data, shortening return latency, and improving overall system response speed. Construction data visualization: By generating a fine DSM model and construction heat map, this platform can dynamically track earthwork volume trends, providing intuitive decision support for construction teams.

[0035] In summary, this embodiment not only improves modeling efficiency and accuracy but also enhances the system's adaptability and responsiveness to complex construction environments by establishing a tightly coordinated perception and processing link between the drone side, the edge layer, and the cloud.

[0036] Furthermore, the radar host is a millimeter-wave radar, which is used to penetrate suspended particles in the air and sense terrain contour information in haze or low-visibility environments. The millimeter-wave radar is combined with an acoustic array system, which includes multiple acoustic wave transmitting and receiving units to assist in calibrating the spatial positioning results of the millimeter-wave radar. The infrared slave is a multispectral imager that operates in the 8-14μm band and is used to obtain image information in different bands. The multispectral imager has a built-in infrared reflectivity database of metal materials to improve its ability to identify the reflective characteristics of different metal surfaces. The integrated PM2.5, temperature, humidity and pressure sensor is used to detect the concentration of fine particulate matter, temperature, humidity and air pressure parameters in the environment.

[0037] In this embodiment, it should be noted that the perception components carried by the drone equipment of the cloud-edge coordination and control platform have been further designed with multimodal integration, specifically including: millimeter-wave radar, acoustic wave array system, multispectral imager (infrared slave) and PM2.5 and temperature, humidity and pressure integrated sensor, so as to achieve stronger environmental adaptability and multi-source information fusion perception effect.

[0038] Specifically: Millimeter-wave radar, the main unit of the radar host, operates in the 30GHz to 300GHz frequency band. It has excellent penetration performance and is particularly suitable for detecting terrain contours in low-visibility or hazy environments. By transmitting frequency-modulated continuous waves (FMCW) and receiving echo signals, millimeter-wave radar constructs preliminary point cloud data of the target area in real time, which serves as the terrain foundation for generating the rough model of the DSM.

[0039] The acoustic array system, working in conjunction with the millimeter-wave radar, consists of multiple ultrasonic transmitting and receiving units positioned in different locations. The system transmits ultrasonic pulses and calculates the spatial position of obstacles based on the echo's time of flight (TOF). It further incorporates a temperature and humidity compensation model to correct for variations in sound velocity, improving positioning accuracy. When the radar echo signal-to-noise ratio degrades due to environmental interference, the acoustic array serves as an auxiliary reference to calibrate the millimeter-wave radar's point cloud results.

[0040] The infrared slave is a multispectral imager operating in the 8–14μm wavelength range, covering the long-wave infrared region. It can capture thermal images in different wavelengths in low-light or no-light conditions. This imager includes a built-in infrared reflectivity database for metal materials, comparing the reflectivity characteristics of different metals (such as steel, aluminum, and copper) in different wavelengths. This improves the accuracy of surface recognition for mechanical equipment, especially in scenes with subtle temperature differences but strong material contrast.

[0041] An integrated PM2.5, temperature, humidity, and pressure sensor supplements environmental perception. It monitors real-time environmental parameters such as fine particulate matter concentration (PM2.5), temperature, humidity, and air pressure along the flight path. This data is used to correct errors in sound velocity calculation during acoustic detection and also provides environmental support for error assessment and filtering compensation in subsequent point cloud data.

[0042] Implementing the technical solution of this embodiment has the following technical effects: The radar positioning stability and anti-interference performance are significantly improved. The millimeter-wave radar has the ability to penetrate haze and can stably output terrain contour information under low visibility conditions, overcoming the shortcoming of visual sensors that are prone to failure in bad weather.

[0043] To enhance obstacle recognition and avoidance capabilities, the platform uses an acoustic array system to perform spatiotemporal alignment and error correction on the millimeter-wave radar's scanning data. This enables multimodal complementary perception, allowing for more accurate identification of slender metal obstacles such as crane cables and scaffolding, improving dynamic obstacle avoidance capabilities.

[0044] Enhanced infrared image recognition accuracy. With the help of the infrared database matching capability of the multispectral imager, the system can still identify the material characteristics of equipment even when the heat source area is not obviously distributed or the temperature difference is low, thereby improving the precise locking and image re-shooting effects of construction machinery.

[0045] The dynamic environmental parameter closed-loop correction mechanism reversely calibrates the environmental variable dependencies in millimeter-wave radar and acoustic ranging through PM2.5 concentration, temperature, humidity and pressure data, enabling the platform to have dynamic adaptive calibration capabilities, thereby improving data reliability and scenario generalization capabilities.

[0046] In a specific embodiment, the present invention further provides a cloud-edge coordination and control method, which is applied to the above-mentioned cloud-edge coordination and control platform, including: S100: Generate and send the UAV's cruise path, control the UAV to cruise along the cruise path, and scan the target area to collect terrain data to generate a rough DSM model. In this step, it should be noted that the generation of the cruise path is determined by the cloud-based control center based on factors such as historical flight paths, construction stage characteristics, regional construction frequency, and risk level, and is automatically sent to the UAV system controlled by the edge control layer. When the UAV performs a cruise mission, the millimeter-wave radar and infrared slave device on board work together. The millimeter-wave radar continuously scans the area to obtain terrain contour data, and the infrared slave device takes fixed-point details of areas with high heat source features. All perception data is initially filtered and fused at the edge to construct a rough DSM (digital surface model) model.

[0047] S200: The DSM rough model is pre-processed through the edge control layer and uploaded to the cloud control center. It should be noted that the edge control layer has a built-in lightweight Gaussian splatter image enhancement engine and anomaly detection module, which are used to perform image enhancement processing on the generated DSM rough model, improve image clarity and boundary features, and mark abnormal areas in the image where the signal-to-noise ratio is below a preset threshold (such as point cloud holes and blurred boundaries caused by haze). This step also includes registration, coordinate unification, and format standardization of multi-source perception data to ensure the structural integrity and algorithm compatibility of the uploaded data.

[0048] S300. Error correction is performed on the DSM coarse model through the intelligent processing algorithm for engineering perception data deployed by the cloud control center to generate a DSM fine model. In this step, it should be noted that the cloud control center is deployed with an intelligent processing algorithm for engineering perception data. After receiving the DSM coarse model uploaded by the edge, it first restores and interpolates the abnormal area, and then combines the ranging results of the acoustic wave and millimeter wave radar to perform error estimation to compensate for point cloud offset or noise distortion caused by dense fog, reflection or occlusion. This step also includes further refining the surface structure of the three-dimensional model based on historical data sets, multi-period comparison results, sensor internal reference models, etc., to achieve high-precision DSM fine model reconstruction.

[0049] S400: Generate a construction heat map based on the DSM model. This step, which requires explanation, uses the DSM model as its foundation. By comparing the volume differences between the current model and the historical model, the construction heat map calculates earthwork changes and performs spatiotemporal overlay analysis based on the heat source distribution (high-frequency mechanical activity areas). The resulting visualization output is a regional construction intensity distribution map. This heat map supports subsequent construction scheduling, work efficiency analysis, and risk warnings.

[0050] Implementing the technical solution of this embodiment has the following technical effects: With accurate and efficient data collection and modeling capabilities, drones can perform tasks without human intervention, can autonomously perform path scanning, quickly complete terrain data collection and preliminary modeling, and significantly improve inspection efficiency.

[0051] The stable perception capability in complex environments such as haze, through the collaborative perception of millimeter-wave radar and acoustic wave array, as well as edge-end image enhancement processing, effectively alleviates the interference of low-visibility environments on modeling accuracy and improves the quality of DSM rough models.

[0052] Cloud-based intelligent optimization ensures accuracy. After error correction by the cloud-based control center, the DSM rough model forms a high-precision DSM fine model, providing a solid foundation for the generation of construction heat maps and avoiding data distortion caused by noise errors in traditional modeling.

[0053] Data support for construction management decision-making has been enhanced. The construction heat map, as a layer that intuitively reflects the intensity and distribution of construction activities, provides reliable data support for construction units to optimize project scheduling, adjust resource allocation, and evaluate project quantities.

[0054] Furthermore, collecting terrain data includes: obtaining terrain contour data by scanning with a radar host, and obtaining heat source area images by scanning with an infrared slave.

[0055] In this embodiment, it's important to note that the collected terrain data not only includes conventional terrain contour information but also incorporates image information of heat source areas to achieve more comprehensive three-dimensional perception and scene understanding of the target area. Specifically, the radar master actively transmits millimeter-wave signals to perform high-precision detection of the surface contour, capturing structural features such as terrain boundaries and undulations. Simultaneously, the infrared slave uses a multispectral scanning method to capture thermal radiation from moving objects such as machinery and personnel, thereby locating heat source areas within the construction site. This multimodal perception mechanism overcomes the limitations of traditional single-view or radar perception, enabling the platform to maintain superior terrain reconstruction and dynamic object recognition capabilities even in low-visibility environments (such as haze, smoke, and dust).

[0056] Implementing the technical solution of this embodiment can effectively improve the integrity and perception accuracy of terrain data. On the one hand, millimeter-wave radar's perception of structural information is unaffected by visible light, ensuring clear terrain contours can be obtained even under complex weather conditions. On the other hand, the infrared slave camera supplements the ability to identify heat sources, making it particularly suitable for focused detection and detailed capture in high-intensity work areas. This helps capture the location of high-heat-intensity construction machinery, enriches the semantic information of the DSM rough model, and provides important input for subsequent construction heat map generation and flight path optimization.

[0057] Furthermore, the radar host includes a millimeter-wave radar and an acoustic wave array system. Accordingly, controlling the UAV to cruise along the cruise path in S100 includes: S110: Millimeter-wave radar transmits frequency-modulated continuous waves (FMCW) at the target area to acquire terrain contour data. It should be noted that millimeter-wave radar operates in the 30 GHz frequency band and above, offering excellent penetration. It can acquire coarse-grained contour data of the target area even in low-visibility environments (such as haze, dust, and low light). The radar periodically transmits frequency-modulated continuous waves (FMCW) and receives echo signals to measure the relative distance to obstacles and their morphological changes. The resulting echoes are pre-processed to construct a basic terrain point cloud, forming the initial contours of the DSM coarse model.

[0058] S120. Transmit ultrasonic pulses to the target area through the acoustic array system and receive echo signals after encountering obstacles. It should be noted that the acoustic array consists of multiple transmitting and receiving units, using directional transmission and multi-point reception to achieve small-scale, highly sensitive obstacle identification. This system is particularly suitable for detecting metal obstacles (such as crane cables, steel structures, and scaffolding). Even in areas where millimeter waves fail due to multipath scattering or attenuation, the acoustic system can still stably identify reflectors, significantly improving the robustness of overall obstacle identification.

[0059] S130. Calculate the spatial position of the obstacle based on the acoustic wave flight time and temperature and humidity correction model. In this step, it should be noted that in order to obtain high-precision acoustic ranging results, the system is based on the Time of Flight (TOF) principle and combines the ambient temperature and humidity correction formula: Where T represents the ambient temperature (°C) and H represents the relative humidity (%) to correct the sound wave propagation speed c. Then according to the distance calculation formula , accurately obtain the acoustic ranging value, and thus calculate the position of each obstacle point in space, which is especially suitable for compensating blind spots caused by occlusion or errors in millimeter wave scanning.

[0060] S140: Temporally and spatially align the obstacle location data measured by the acoustic array system with the millimeter-wave radar data, and calculate the ranging error. In this step, it should be noted that the system assigns a unified timestamp to the acoustic array and millimeter-wave radar data and establishes a coordinate transformation relationship in geometric space to achieve temporal and spatial alignment of the two sets of sensor data.

[0061] S150: Determine and correct the millimeter-wave radar positioning error based on the ranging error. It should be noted that the system dynamically compensates for the millimeter-wave radar positioning based on the ranging error. This adaptive filtering mechanism ensures the continuity and accuracy of millimeter-wave point cloud data even in inclement weather, overcoming the drawbacks of traditional radar systems that often experience point cloud holes or jumps in foggy and hazy environments.

[0062] Implementing the technical solution of this embodiment can realize a multimodal collaborative perception mechanism, combining the advantages of millimeter-wave radar and acoustic wave array systems, and complementary perception under different physical interference conditions to ensure the integrity and anti-interference capability of obstacle recognition.

[0063] Improve ranging accuracy and positioning robustness, enhance accuracy through acoustic ranging and environmental compensation models, and introduce a dynamic error correction mechanism to achieve real-time optimization processing of millimeter wave radar point cloud data.

[0064] To enhance operational stability in adverse weather conditions, the system is specially designed for typical meteorological environments on construction sites, such as fog, haze, and high humidity, significantly reducing the impact of the external environment on the drone's modeling accuracy and obstacle avoidance performance.

[0065] Providing reliable basic data for the subsequent DSM fine model construction, the corrected high-confidence point cloud lays the data foundation for the accuracy of the digital surface model, thereby improving the accuracy of subsequent construction thermal maps and volume calculations.

[0066] In this embodiment, determining and correcting the positioning error of the millimeter wave radar includes: S151: Timestamp the obstacle distance data obtained by the millimeter-wave radar and the acoustic array system, and convert the acoustic data coordinates to the millimeter-wave radar's coordinate system. It should be noted that due to differences in timing and spatial coordinate systems between the data collected by the millimeter-wave radar and the acoustic array system, effective alignment requires time synchronization. The system uses the drone's unified clock system or synchronization signal to accurately timestamp the two modal data. Then, based on the acoustic array's installation position relative to the millimeter-wave radar (including installation angle and baseline distance), a rigid coordinate transformation (such as Euler angle or quaternion conversion) is performed to uniformly map the acoustic coordinate system data to the millimeter-wave radar's coordinate system, achieving spatial alignment of the two modal data.

[0067] S152. Calculate the deviation between the obstacle distance data obtained by the millimeter wave radar and the acoustic wave array system:

[0068] In this step, the system calculates the difference between the distances measured by the millimeter-wave radar and the distances measured by the acoustic array system. By performing this difference analysis at key obstruction points or continuous point cloud segments, it can determine whether the millimeter-wave radar is experiencing interference from false alarms, missing values, or jumps in specific areas, thereby forming a dynamic basis for evaluating the radar's ranging accuracy.

[0069] S153. Obtain the PM2.5 concentration value at the corresponding time point and determine whether any of the following conditions are met: Exceeds the preset error threshold or PM2.5 concentration is higher than the preset fog environment standard. In this step, it should be noted that in order to improve the adaptability and intelligence of the compensation mechanism, the system not only relies on the ranging error between sensors, but also introduces the ambient air quality index as a supplementary judgment condition. Specifically, when the error If the distance exceeds a dynamically set threshold (e.g., 0.3m) or the monitored PM2.5 concentration (μg / m³) exceeds the dense fog level (e.g., >250), the system deems the current radar ranging unreliable and initiates a compensation process. This judgment mechanism combines environmental parameters with sensor status to form an intelligent judgment model with dynamic environmental perception capabilities.

[0070] S154. If any of the above conditions is met, the compensation mechanism is triggered: If the haze interference is serious, the acoustic wave coordinates are directly used to replace the corresponding point cloud coordinates; If the interference level is moderate, the radar and acoustic wave coordinates are weighted and fused according to the preset weights to generate the corrected point cloud coordinates: In this step, it should be noted that the system distinguishes different levels of interference and handles them differently: In heavy interference scenarios (such as PM2.5>400 or ) Under the condition of high-resolution scattering, the radar echo has serious scattering distortion, and the system directly replaces the original millimeter wave data points with the spatial coordinate points output by the acoustic array; Under moderate interference, the system adopts a fusion strategy and calculates the corrected coordinates as:

[0071] in, The fusion weight is dynamically set (such as 0.6~0.8). This method not only retains the overall distribution characteristics of the millimeter-wave radar point cloud, but also incorporates the local accuracy of acoustic ranging, significantly improving the model's precision and consistency.

[0072] In a specific embodiment, The value is 0.7.

[0073] S155. Use the corrected point cloud data as cruise coordinate data. It's important to note that after integrating millimeter-wave radar and acoustic array system data and performing error correction based on PM2.5 concentration and ranging errors, the resulting set of high-precision terrain point cloud data no longer serves solely as modeling input; it also serves as the foundational coordinate set for the drone's subsequent flight and cruise paths. This coordinate data boasts enhanced terrain conformance and obstacle avoidance capabilities, enabling direct use in path replanning, flight navigation, and dynamic obstacle avoidance strategy generation.

[0074] Implementing the technical solution of this embodiment can realize a highly robust adaptive calibration mechanism, integrate the multi-source perception capabilities of millimeter-wave radar and acoustic ranging, and achieve high-reliability correction and compensation of millimeter-wave ranging errors through the spatiotemporal alignment of point cloud data and dynamic environmental state judgment.

[0075] This solution specifically addresses the error accumulation problem caused by the degradation of millimeter-wave sensing performance in low-visibility environments. It sets a dual judgment mechanism of PM2.5 concentration threshold and ranging deviation threshold. Once a perception anomaly is detected, the sound wave-dominated compensation strategy is triggered, thereby significantly enhancing the system's stability and adaptability in complex environments.

[0076] By directly using the corrected and optimized point cloud data as the basis for cruise coordinates, spatial deviation problems caused by inaccurate ranging or modeling anomalies can be effectively avoided, ensuring that the UAV has higher path accuracy and obstacle avoidance capabilities in subsequent flight missions, and improving the spatial consistency and structural integrity of the DSM rough model from the source, thereby maximizing the flight safety of the UAV during cruise and reducing the risk of collision.

[0077] In an optional embodiment, pre-processing the DSM rough mold by the edge control layer in S200 includes: S210, abnormal area marking, by performing signal-to-noise ratio analysis on the image or point cloud data in the DSM rough model, identifying areas where the signal-to-noise ratio is lower than the preset threshold, and marking the abnormal area. In this step, it should be noted that the edge control layer uses the built-in Gaussian splash engine to first perform pixel-level or point cloud-level analysis on the DSM rough model uploaded from the drone, and extracts its signal strength to background noise ratio (SNR). The system sets a lower limit threshold for the signal-to-noise ratio (such as SNR < 10dB) as a judgment standard, and automatically identifies and marks areas below this threshold. Such abnormal areas are often caused by factors such as haze obstruction, sensor angle deviation, or unstable device posture. Identifying these areas in advance facilitates subsequent image enhancement or reshooting task planning.

[0078] S220: Data enhancement processing, performing quality optimization operations on the DSM rough model data, including but not limited to noise filtering, detail sharpening, and contrast adjustment. It should be noted that in this step, the system uses a variety of image / point cloud enhancement algorithms to improve the quality of the marked area or the entire DSM rough model image. For example, for image-based DSMs, non-local mean filtering (NLM) can be used to remove granular noise, and the Laplace operator can be applied to edge areas to enhance edge sharpness. For point cloud-based DSMs, voxel grid filtering (VoxelGrid) can be applied to reduce outliers, and curvature smoothing can be used to reconstruct the true terrain surface. The enhanced DSM data retains structural details while significantly reducing the interference of measurement noise on subsequent modeling.

[0079] S230: Data compression processing compresses the preprocessed DSM coarse model data. It should be noted that in this step, to reduce data upload bandwidth load and cloud processing latency, the system performs lossy or lossless compression on the preprocessed DSM coarse model data. For image-based DSMs, compression algorithms that support progressive transmission, such as JPEG2000, can be used. For point cloud data, Octree encoding or point index compression methods in PCL (Point Cloud Library) can be used. Compression processing ensures efficient data transmission and rapid cloud upload within limited resources, while also ensuring that key features required for model reconstruction are not compromised.

[0080] Implementing the technical solution of this embodiment can improve the reliability and availability of data input, actively mark abnormal areas through signal-to-noise ratio analysis, ensure that the system responds to low-quality data in a targeted manner, and improve overall modeling accuracy.

[0081] Enhance the visual clarity and geometric accuracy of DSM data, use image and point cloud enhancement technology to optimize the details of the original rough model data, significantly improve image contrast and edge clarity, and reduce interference caused by noise.

[0082] Improve edge computing efficiency and cloud processing fluency, significantly reduce the volume of uploaded data through a reasonable data compression mechanism, reduce dependence on communication links, and further improve edge-cloud collaborative processing efficiency.

[0083] In a specific embodiment, the intelligent processing algorithm for engineering perception data is a haze scattering error compensation algorithm. Accordingly, the error correction in S300 includes: S310: Collect real-time PM2.5 concentration and atmospheric humidity. It should be noted that in this step, after the cloud control center receives the coarse DSM model uploaded by the edge control layer, it first calls the environmental perception module deployed in the monitoring system to collect real-time PM2.5 concentration values ​​and atmospheric humidity parameters corresponding to the target area to achieve accurate error compensation. The collected data sources may include: environmental sensors onboard drones (such as integrated temperature, humidity, pressure, and fine particulate matter monitoring modules), as well as connected regional meteorological service interfaces (such as local weather stations or cloud-based meteorological APIs). This data serves as an important input variable for subsequent haze modeling and point cloud error estimation.

[0084] S320: Based on the haze forecast model, predict the height offset of the point cloud. Further, the height offset of the point cloud predicted in S320 includes: The point cloud height is corrected using the following error correction formula:

[0085] Where, is the construction site dust coefficient, which is used to characterize the long-term statistical characteristics of suspended particle concentration in the construction scene; is the scattering attenuation factor, which reflects the energy attenuation of the millimeter wave signal under different concentrations of particulate matter; The atmospheric humidity, construction site dust coefficient and scattering attenuation factor are obtained through historical data training, and have certain regional adaptability and model generalization capabilities. In this step, it should be noted that the system has a built-in haze impact prediction model trained based on historical point cloud offsets, which uses input environmental factors such as PM2.5 concentration and atmospheric humidity as variables and outputs a point cloud height offset estimate. The model can be constructed using polynomial regression, random forest regression or lightweight neural network, and is suitable for predicting the systematic error caused by haze conditions on the propagation distance of millimeter-wave radar signals. This offset reflects the terrain point cloud height estimation error caused by the prolonged radar signal propagation time due to the haze environment.

[0086] S330. Use the height offset of the point cloud to correct the height of the original point cloud and adjust the height data in the point cloud. In this step, it should be noted that the cloud control center corrects each terrain point cloud data in the DSM rough model point by point based on the predicted point cloud height offset. The correction methods include direct linear compensation or weighted adjustment. For example, if the average height deviation caused by haze in a certain area is predicted to be -12cm, the system will uniformly add 12cm to the height values ​​of all point clouds in the area. For areas with uneven impact, the system uses differential grid correction to ensure the continuity of the terrain model and the match between the real terrain. The corrected point cloud data is then used as input data for the generation of the DSM fine model, which significantly improves the geometric accuracy of the modeling.

[0087] Implementing the technical solution of this embodiment can enhance the robustness of the model to environmental disturbances. By integrating PM2.5 and humidity monitoring, the system can dynamically perceive the impact of atmospheric conditions on radar ranging and ensure the environmental adaptability of data processing.

[0088] Achieve accuracy compensation for point cloud modeling under haze conditions, use prediction models to quantitatively predict point cloud deviations, and make up for the shortcomings of traditional modeling in low-visibility environments where errors are uncontrollable.

[0089] The geometric accuracy and credibility of the DSM fine model are improved. The corrected point cloud data is closer to the actual terrain, ensuring the reliability of thermal map generation and earthwork calculation results, and providing a highly reliable data foundation for subsequent drone scheduling and engineering construction evaluation.

[0090] Furthermore, in a specific embodiment, the method further comprises: S500: Identify potential flight risk areas and obstacles within the construction area based on the latest terrain morphology and obstacle spatial distribution information contained in the DSM model. This step, by analyzing the highly precisely restored terrain height field and edge features in the DSM model, combined with indicators such as point cloud density, structural contour changes, and thermal zone concentration, clusters and labels sudden changes in terrain, areas of high elevation difference, and dynamic heat sources (such as operating construction machinery) within the construction site. This automatically identifies flight risk areas and obstacles that could interfere with low-altitude drone cruising.

[0091] In addition, abnormal fluctuations in heat sources (captured by infrared slaves) can be used to assist in identifying mobile devices with high thermal radiation, avoiding path conflicts caused by moving obstacles.

[0092] S600: Compare and analyze the DSM model with historical flight paths and obstacle avoidance records to generate an updated path planning solution. This step utilizes the constructed DSM model for multi-dimensional comparison and analysis with historical flight paths, obstacle avoidance point sets, and flight interruption records stored in a database. By constructing a multi-objective constraint graph (with terrain feasibility, flight efficiency, energy consumption, and safety margin as objective functions), a graph-based search (such as the A algorithm or RRT tree) is performed to output the optimal path under the current spatiotemporal conditions.

[0093] At the same time, the system automatically evaluates the possibility of historical obstacle avoidance events recurring in the current environment, plans preparatory avoidance strategies in advance, and marks them as "foreseeable risk segments" in the path, dynamically monitoring the risk level of this segment during subsequent flights.

[0094] S700: The path planning solution and corresponding obstacle avoidance strategy are distributed to the drone in the form of a template. This step encapsulates the path planning results and obstacle avoidance strategy in the form of a "flight strategy template," including but not limited to target coordinates, altitude constraints, flight speed, emergency avoidance instructions, and trigger thresholds for each flight segment. This template structure facilitates direct reading and execution by the drone's edge control layer, improving runtime responsiveness.

[0095] At the same time, the template supports real-time cloud-based delivery or offline embedding in the flight control system, and can automatically switch according to the on-site network status to enhance the platform's field adaptability.

[0096] The technical solution of this embodiment can dynamically construct a flight safety map and identify flight risk areas through high-precision DSM models. Compared with static map solutions, it has stronger real-time and accuracy and is particularly suitable for dynamic construction site environments.

[0097] It realizes intelligent coordination of path and obstacle avoidance strategies, and combines historical data to form a flight optimization model based on experience knowledge, which can significantly reduce obstacle avoidance response delay while ensuring flight efficiency.

[0098] To enhance the operational autonomy and deployment flexibility of the platform, the flight strategy templates can be transmitted and run locally in a standardized manner, with high versatility and low coupling characteristics, which is conducive to the rapid migration and deployment of the system between different drone models.

[0099] This embodiment further expands the application scenarios of DSM precision models in path planning and flight safety, forming a closed-loop data perception-graph construction-path optimization-command execution process, significantly improving the UAV's operational capabilities in complex environments.

[0100] The above-described embodiments merely represent several embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A cloud-edge coordination and management platform, characterized by: include: The drone is equipped with a radar host and an infrared slave. The radar host is used to scan the target area and obtain scanning data; the infrared slave is used to lock onto the heat source area generated by the mechanical equipment, take detailed photos, and obtain detailed data. An edge control layer, provided on the UAV device, has a built-in lightweight Gaussian splash engine for pre-processing the scan data and detail data, marking abnormal areas where the image signal-to-noise ratio is lower than a preset threshold, and generating a DSM rough model; The cloud-based control center is used to receive the DSM rough model and optimize the DSM rough model through the deployed engineering perception data intelligent processing algorithm to generate a DSM fine model; the cloud-based control center further calculates the earthwork volume of the target area based on the DSM fine model and generates a construction heat map.

2. The cloud-edge coordination and management platform according to claim 1 is characterized in that: The radar host is a millimeter-wave radar, which is used to penetrate suspended particles in the air and sense terrain contour information in haze or low-visibility environments. The millimeter-wave radar is combined with an acoustic wave array system, which includes multiple acoustic wave transmitting and receiving units for assisting in the calibration of the spatial positioning results of the millimeter-wave radar. The infrared slave is a multispectral imager that operates in the 8-14 μm band and is used to obtain image information in different bands. The multispectral imager has a built-in infrared reflectivity database of metal materials to improve the multispectral imager's ability to identify the reflective characteristics of different metal surfaces. The integrated PM2.5, temperature, humidity and pressure sensor is used to detect the concentration of fine particulate matter, temperature, humidity and air pressure parameters in the environment.

3. A cloud-edge coordination and control method, applied to the cloud-edge coordination and control platform according to claims 1-2, characterized in that: include: Generate and issue a UAV cruise path, control the UAV to cruise along the cruise path, and perform terrain scanning on the target area to collect terrain data and generate a rough DSM model; The DSM rough model is pre-processed through the edge control layer and uploaded to the cloud control center; Error correction is performed on the DSM rough model using the engineering perception data intelligent processing algorithm deployed by the cloud control center to generate a DSM fine model; A construction heat map is generated based on the DSM fine mold.

4. The cloud-edge coordination and management method according to claim 3 is characterized in that: The collecting of terrain data includes: obtaining terrain contour data by scanning with a radar host, and obtaining a heat source area image by scanning with an infrared slave.

5. The cloud-edge coordination and management method according to claim 3 is characterized in that: The radar host includes a millimeter wave radar and an acoustic wave array system. Accordingly, controlling the UAV to cruise along the cruise path includes: The millimeter wave radar transmits a frequency modulated continuous wave in the target area to obtain terrain contour data; Transmitting ultrasonic pulses to the target area through the acoustic wave array system and receiving echo signals after encountering obstacles; Calculate the spatial position of obstacles based on the sound wave flight time and temperature and humidity correction model; Performing spatiotemporal alignment of obstacle position data measured by the acoustic array system with millimeter-wave radar data to calculate ranging errors; Based on the ranging error, the positioning error of the millimeter wave radar is determined and corrected.

6. The cloud-edge coordination and management method according to claim 5 is characterized in that: The determining and correcting of the positioning error of the millimeter wave radar includes: Adding a unified timestamp to the obstacle distance data obtained by the millimeter-wave radar and the acoustic wave array system, and converting the coordinates of the acoustic wave data into the coordinate system of the millimeter-wave radar; Calculate the deviation between the obstacle distance data obtained by the millimeter wave radar and the acoustic wave array system: Get the PM2.5 concentration value at the corresponding time point and determine whether any of the following conditions are met: Exceeding a preset error threshold; The PM2.5 concentration is higher than the preset standard for dense fog; If any of the above conditions is met, the compensation mechanism will be triggered: If the haze interference is serious, the acoustic wave coordinates are directly used to replace the corresponding point cloud coordinates; If the interference level is moderate, the radar and acoustic wave coordinates are weighted and fused according to the preset weights to generate the corrected point cloud coordinates: The corrected point cloud data is used as the coordinate data for cruising.

7. The cloud-edge coordination and management method according to claim 3 is characterized in that: The pre-processing of the DSM block mold by the edge control layer includes: Abnormal area marking, by performing signal-to-noise ratio analysis on the image or point cloud data in the DSM rough model, identifying areas where the signal-to-noise ratio is lower than a preset threshold, and marking the abnormal area; Data enhancement processing, performing quality optimization operations on the DSM rough model data, including but not limited to noise filtering, detail sharpening and contrast adjustment; Data compression processing is to compress the pre-processed DSM rough model data.

8. The cloud-edge coordination and management method according to claim 3 is characterized in that: The engineering perception data intelligent processing algorithm is a haze scattering error compensation algorithm. Accordingly, the error correction includes: Collect real-time PM2.5 concentration and atmospheric humidity; Based on the haze forecast model, the height offset of the point cloud is predicted; The height offset of the point cloud is used to correct the original point cloud height and adjust the height data in the point cloud.

9. The cloud-edge coordination and management method according to claim 8, characterized in that: The height offset of the predicted point cloud includes: The point cloud height is corrected using the following error correction formula: Where, is the construction site dust coefficient, is the scattering attenuation factor, is the atmospheric humidity, and the construction site dust coefficient and the scattering attenuation factor are obtained through historical data training.

10. The cloud-edge coordination and management method according to claim 3, characterized in that: The method further comprises: Identify potential flight risk areas and obstacles within the construction area based on the latest terrain morphology and obstacle spatial distribution information contained in the DSM; Comparing and analyzing the DSM fine model with historical flight trajectories and obstacle avoidance records to form an updated path planning solution; The path planning scheme and the corresponding obstacle avoidance strategy are sent to the drone in the form of a template.

Images