Comprehensive sensing system and method of global intelligent platform

Abstract

The invention relates to the technical field of city management, in particular to a multi-global intelligent connection platform comprehensive sensing system and method. Comprising a high-low altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer and a closed-loop management module. The high-low altitude linkage sensing layer comprises a high-position video monitoring device deployed in an iron tower base station, a multi-rotor unmanned aerial vehicle in an unmanned aerial vehicle hangar, and an Internet of Things sensor array arranged on the ground, the high-position video monitoring device is coupled with a multispectral camera through a cradle head with an adjustable pitch angle, and covers a monitoring area with a radius not less than 500 meters; the unmanned aerial vehicle hangar is internally provided with a charging pile and a meteorological monitoring unit, the unmanned aerial vehicle carries an edge calculation module and a dual-light camera, and the Internet of Things sensor array comprises a voiceprint sensor, an air quality sensor and a vibration sensor; according to the design, geographical limitation can be broken through, and global seamless monitoring is realized, so that the timely discovery rate of events in a hidden area and the cross-department co-processing efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of urban management technology, specifically to a comprehensive sensing system and method for a fully intelligent interconnected platform. Background Technology

[0002] With the acceleration of new urbanization, urban and rural governance faces monitoring challenges in complex scenarios such as illegal high-altitude construction, water pollution, and forest fire risks. Existing urban and rural governance systems generally rely on a combination of fixed ground monitoring equipment and manual inspections, but significant blind spots exist in monitoring coverage in complex terrain areas. For example, visual obstruction caused by undulating mountainous terrain reduces the effective monitoring radius of traditional low-point cameras to less than 200 meters, making it difficult to detect precursors to landslides or illegal logging. In water scenarios, shoreline cameras cannot cover floating objects more than 300 meters from the shore, and image quality deteriorates sharply at night. More seriously, the deployment of existing sensing equipment is fragmented: security cameras, environmental sensors, and forestry monitoring equipment belong to different management systems, with incompatible data formats and communication protocols, forming "data silos." When cross-regional events occur, manual coordination is required to retrieve scattered data from multiple departments, resulting in event analysis taking more than two hours and missing the optimal window for response. While some current technologies attempt to compensate for the shortcomings of fixed equipment by increasing the frequency of drone inspections, limitations in endurance and communication stability mean that drones can only operate continuously for less than 25 minutes in mountainous and other scenarios, and the collected data cannot be transmitted in real time due to signal attenuation. Furthermore, the lack of a unified spatiotemporal reference between low-altitude monitoring equipment and ground sensors results in a multi-source data fusion error rate as high as 18%, severely impacting event location accuracy. These problems expose inherent deficiencies in the traditional urban and rural governance system regarding adaptability to complex terrain and the efficiency of multi-device collaboration.

[0003] Based on the above problems, there is an urgent need for a new technical solution that can overcome geographical limitations and achieve seamless monitoring across the entire area, so as to improve the timely detection rate of incidents in hidden areas and the efficiency of cross-departmental collaborative handling. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies, and to propose a comprehensive sensing system and method for a full-domain intelligent interconnection platform, comprising: A comprehensive sensing system for a global intelligent connectivity platform includes: The system comprises a high- and low-altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer, and a closed-loop management module. The high- and low-altitude linkage sensing layer includes high-position video surveillance equipment deployed at tower base stations, multi-rotor drones in drone hangars, and IoT sensor arrays deployed on the ground. The high-position video surveillance equipment is coupled with a multispectral camera through a gimbal with adjustable pitch angle, covering a monitoring area with a radius of not less than 500 meters. The drone hangar has built-in charging piles and meteorological monitoring units. The drones are equipped with edge computing modules and dual-light cameras. The IoT sensor array includes acoustic sensors, air quality sensors, and vibration sensors. The data link layer constructs a hybrid transmission network through a 5G communication module and an optical fiber transmission channel to realize real-time data transmission and command interaction between the high-position video surveillance equipment, drones and IoT sensors. The AI ​​scheduling layer includes a cloud server cluster and edge computing nodes. The cloud server cluster has a built-in dynamic path optimization model and a multi-source data fusion algorithm. The edge computing nodes are deployed in the drone hangar and are used to process the raw image data collected by the drone in real time.

[0005] Preferably, the application layer includes a panoramic view module, a cockpit monitoring module, and an event handling module. The panoramic view module dynamically integrates high-position video, UAV aerial footage, and sensor data based on a geographic information system to generate a three-dimensional visualization map. The cockpit monitoring module predicts the probability of equipment failure and generates maintenance work orders through a time series analysis model. The closed-loop management module includes an event recognition engine and a three-level feedback network. The event recognition engine extracts features from multi-source data through a convolutional neural network and matches them with a preset rule base to generate alarm events. The three-level feedback network pushes alarm events to town and street management departments, district-level command centers and municipal-level supervision platforms according to priority, and receives the processing results and sends them back to the system database.

[0006] By adopting the above technical solution, this claim constructs a full-domain intelligent interconnection system that includes a high- and low-altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer, and a closed-loop management module; it can solve the technical problems in traditional urban and rural governance, including: the sensing devices are scattered and independent, forming data silos and making cross-departmental collaboration difficult; the low-point monitoring coverage is limited and cannot cope with complex terrain scenarios such as mountains and water areas, and the event handling process is lengthy, with manual sorting being time-consuming and prone to missing key information.

[0007] More preferably, the data transmission rate of the data link layer satisfies the following constraints: , in, This indicates the minimum bandwidth requirement for a hybrid transmission network; This represents the data volume of a single frame in the i-th video stream; Indicates the maximum allowable latency for video frame transmission; This represents the amount of data sampled in a single run by the j-th type of sensor; This indicates the sensor sampling period.

[0008] The above technical solution limits the data transmission rate of the data link layer and defines a data transmission rate constraint formula to ensure the real-time transmission capability of multiple video streams and sensor data. It can solve the problems of network congestion when multiple devices transmit concurrently, resulting in extremely high data packet loss rate, and unreasonable resource allocation when low-frequency sensor data and high-frequency video data are transmitted together.

[0009] More preferably, the objective function of the dynamic path optimization model is: , in, This represents the estimated energy consumption of the k-th drone; This represents the signal strength of the k-th UAV at the m-th monitoring point; and Let be the weight coefficient, and satisfy... The constraints include the maximum flight time of the drone, no-fly zone avoidance, and data collection integrity thresholds.

[0010] The above technical solution proposes a UAV path optimization model that integrates energy consumption and signal strength for multi-objective decision-making. It can solve the problems of low single-use endurance of traditional UAVs, frequent return to recharge leading to low inspection efficiency, and the fact that traditional path planning ignores signal attenuation, resulting in signal loss in mountainous missions.

[0011] More preferably, the multi-source data fusion algorithm includes the following steps: Perform moving target detection on high-resolution video data and extract target position coordinates. and movement speed ; Super-resolution reconstruction of drone aerial images to extract target feature vectors ; Constructing a spatiotemporal matrix from sensor data And it is spatiotemporally aligned with the video data; The Kalman filter is used to correlate and match multi-source target information, and the fused target trajectory is output. .

[0012] The above technical solution defines a four-step processing flow for multi-source data fusion algorithms, which can solve the problems in traditional urban and rural governance systems, such as the inconsistency of spatiotemporal references between video and sensor data, resulting in a high fusion error rate, and the difficulty in identifying small-scale targets in low-resolution aerial images, leading to a high rate of missed detection.

[0013] More preferably, the edge computing node performs the following operations: The video stream is compressed in real time during the drone's flight, and the compression ratio is dynamically adjusted according to the link bandwidth. The acquired images undergo local feature enhancement processing, and image blocks containing moving targets are transmitted first. When a communication interruption is detected, a local caching mechanism is activated and data is stored in order of event priority.

[0014] By adopting the above technical solution, this solution can solve the problems of high key frame loss rate caused by the fluctuation of UAV image transmission bandwidth in complex environments such as mountainous areas, and the serious loss of important target details due to the fact that traditional edge nodes only support full frame compression.

[0015] More preferably, the deployment density of the drone hangar is determined according to the following formula: , in, Indicates the area under its jurisdiction; Indicates the event occurrence rate per unit area; This refers to the drone's cruising speed. This represents the maximum allowable response time for the system.

[0016] The above technical solution provides a formula for calculating the deployment density of drone hangars, which can solve the problems of resource waste caused by deploying drone hangars at fixed intervals, high coverage costs in mountainous areas, and excessively long arrival times for drones during emergency response, which may cause them to miss the optimal response window.

[0017] A method, applied to a comprehensive sensing system of a global intelligent connectivity platform as described in any of the above, includes: S1. Video data, aerial images and environmental parameters are collected synchronously through the high and low altitude linkage perception layer, and the raw data is transmitted to the AI ​​scheduling layer through the data link layer. S2. The cloud server calls the dynamic path optimization model to generate the optimal flight path for the drone swarm and sends image processing instructions to the edge computing nodes; S3. The application layer receives the processed data, merges it in the panoramic view module to generate a three-dimensional situation map, and marks abnormal events through the cockpit monitoring module; S4. The closed-loop management module initiates a tiered response mechanism based on the event type, pushes handling work orders to the three levels of management departments, and tracks the handling results to update the system database.

[0018] The above technical solution defines the four steps of the whole-domain intelligent connection method, which can solve the problems of rigid traditional manual inspection plans, a large number of inspection tasks without actual event transmission, and fragmented multi-dimensional data display, requiring decision-makers to compare information across systems.

[0019] Further preferably, the generation of the optimal flight path in S2 includes: constructing a probability map of hotspot areas based on historical work order data; using an ant colony algorithm to solve the multi-UAV coverage path problem; and dynamically adjusting the flight altitude to avoid meteorological risk areas.

[0020] The above technical solution provides three strategies for generating optimal flight paths, which can effectively solve the technical problems of 50% of UAV energy being wasted in low-value areas under the current fixed route mode, and 15% of flight missions being canceled midway due to sudden weather changes.

[0021] More preferably, the triggering condition for the hierarchical response mechanism described in step S4 is: When the sensor detects that the PM2.5 concentration exceeds the threshold for 3 consecutive hours, the environmental protection department's linkage process is automatically initiated. If the same coordinate point is identified as having illegal construction features in different data sources, the event level will be raised to level two. For work orders that are not processed within the specified time, an assessment report will be generated and sent to the supervisory department.

[0022] The above technical solution provides a trigger condition design for a graded response mechanism, which can solve the technical problems in traditional technical solutions, such as the underestimation of 30% of high-risk events due to subjective errors in human judgment of event levels, the lack of quantitative assessment of handling effectiveness, and insufficient motivation for departmental collaboration. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a block diagram of the comprehensive sensing system of the full-domain intelligent connectivity platform in this application; Figure 2 This is a flowchart of the integrated perception system of the whole-domain intelligent interconnection platform in this application. Detailed Implementation

[0025] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0026] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, operations, elements, components and / or collections thereof.

[0027] Please see Figure 1 Traditional urban and rural governance suffers from the following technical problems: sensing devices are scattered and independent, forming data silos and hindering cross-departmental collaboration; low-point monitoring has limited coverage and cannot handle complex terrain scenarios such as mountains and waterways; incident handling processes are lengthy, with manual sorting being time-consuming and prone to missing key information. Based on these issues, this application provides a comprehensive sensing system for a full-domain intelligent interconnection platform, including: The system comprises a high- and low-altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer, and a closed-loop management module. The high- and low-altitude linkage sensing layer includes high-position video surveillance equipment deployed at tower base stations, multi-rotor drones in drone hangars, and IoT sensor arrays deployed on the ground. The high-position video surveillance equipment is coupled with a multispectral camera through a gimbal with adjustable pitch angle, covering a monitoring area with a radius of not less than 500 meters. The drone hangar has built-in charging piles and meteorological monitoring units. The drones are equipped with edge computing modules and dual-light cameras. The IoT sensor array includes acoustic sensors, air quality sensors, and vibration sensors. The data link layer constructs a hybrid transmission network through a 5G communication module and an optical fiber transmission channel to realize real-time data transmission and command interaction between the high-position video surveillance equipment, drones and IoT sensors. The AI ​​scheduling layer includes a cloud server cluster and edge computing nodes. The cloud server cluster has a built-in dynamic path optimization model and a multi-source data fusion algorithm. The edge computing nodes are deployed in the drone hangar and are used to process the raw image data collected by the drone in real time.

[0028] The application layer includes a panoramic view module, a cockpit monitoring module, and an event handling module. The panoramic view module dynamically integrates high-position video, UAV aerial footage, and sensor data based on a geographic information system to generate a 3D visualization map. The cockpit monitoring module predicts the probability of equipment failure and generates maintenance work orders through a time series analysis model. The closed-loop management module includes an event recognition engine and a three-level feedback network. The event recognition engine extracts features from multi-source data through a convolutional neural network and matches them with a preset rule base to generate alarm events. The three-level feedback network pushes alarm events to town and street management departments, district-level command centers and municipal-level supervision platforms according to priority, and receives the processing results and sends them back to the system database.

[0029] It is worth mentioning that this embodiment constructs a full-domain intelligent connectivity system comprising a high- and low-altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer, and a closed-loop management module, including: Perception Layer: Integrates high-level tower monitoring equipment (multispectral camera + gimbal), drones (dual-light camera + edge computing), and ground sensor arrays (soundprint / vibration / air quality sensors) to form a full-area monitoring network covering a low altitude of 500 meters.

[0030] Data Link: Real-time transmission of multi-source data is achieved through a hybrid 5G and fiber optic network, meeting the concurrent transmission requirements of video streams and sensor data.

[0031] AI scheduling: It adopts a two-level processing architecture of cloud dynamic path optimization model and edge computing nodes. The cloud is responsible for global resource scheduling, and the edge nodes realize real-time compression and feature extraction of UAV images.

[0032] Application layer: Multi-dimensional data is integrated through GIS 3D visualization maps, and the cockpit module introduces time series prediction models to realize equipment failure prediction.

[0033] Closed-loop management: Based on a CNN-based event recognition engine and a three-level feedback network, a closed-loop process is achieved from event discovery to the tracking of handling results.

[0034] The technical effects of this embodiment include: ensuring comprehensive monitoring coverage by eliminating blind spots in traditional monitoring through the collaboration of high-position cameras and drones; improving data processing efficiency by enabling edge computing nodes to compress drone image processing latency to within 200ms, shortening cross-departmental event response time, and enabling major events to be quickly transmitted to the municipal platform through a three-level feedback network.

[0035] In traditional multi-device concurrent transmission scenarios, network congestion leads to a high data packet loss rate; furthermore, resource allocation is unreasonable when low-frequency sensor data and high-frequency video data are transmitted together. Therefore, the data transmission rate of the data link layer must meet the following constraints: , in, This indicates the minimum bandwidth requirement for a hybrid transmission network; This represents the data volume of a single frame in the i-th video stream; Indicates the maximum allowable latency for video frame transmission; This represents the amount of data sampled in a single run by the j-th type of sensor; This indicates the sensor sampling period.

[0036] In the above formula This represents the minimum threshold for the total system bandwidth requirement, which must meet the transmission needs of all concurrent data streams. By quantifying the instantaneous bandwidth requirements of video streams and sensor data, it prevents the loss of keyframes due to insufficient link capacity.

[0037] This represents the raw data size of each frame in a single video stream, and is related to the resolution and encoding format. Using H.265 encoding, the data size of a single frame in 4K video can be compressed from 15MB to 2MB, significantly reducing bandwidth pressure.

[0038] This represents the amount of data collected in a single sampling from a single type of sensor, including numerical values, timestamps, and geographic location information.

[0039] This represents the minimum time interval between two sensor samples, determining the data generation frequency. It is linked to the drone's path planning, dynamically increasing the sensor sampling frequency as the drone flies over the monitoring point.

[0040] To illustrate with an example: A town deploys 5 drones for inspection video, each encoded with H.265 at 20Mbps, and 50 air quality sensors (sampling period 60 seconds, single data 0.2KB). Then: , A 500Mbps link can meet the actual deployment requirements and reserve buffers for bursty traffic.

[0041] It is worth mentioning that the above formula defines a data transmission rate constraint formula, ensuring the real-time transmission capability of multiple video streams and sensor data. This embodiment provides dynamic bandwidth allocation, implementing a bandwidth preemption strategy for critical data streams such as fire alarms based on the video frame rate, ensuring their transmission latency is below 100ms; it can guarantee bandwidth utilization up to 92%, achieving optimal resource allocation for different data types through formula constraints; and the video stream transmission packet loss rate is reduced to below 0.5%, meeting the stringent requirements of law enforcement for image continuity.

[0042] Traditional drones, for example, have a single flight time of only 40 minutes, and the frequent return to home for charging leads to low inspection efficiency. Traditional path planning ignores signal attenuation, and most of the data in mountainous tasks becomes invalid due to signal loss. Based on this, the objective function of the proposed dynamic path optimization model is: , in, This represents the estimated energy consumption of the k-th drone; This represents the signal strength of the k-th UAV at the m-th monitoring point; and Let be the weight coefficient, and satisfy... Constraints include unmanned aerial vehicles (UAVs). Maximum flight time, no-fly zone avoidance, and data collection integrity thresholds.

[0043] In the above formula This represents the estimated energy consumption of the k-th UAV performing its mission, which is related to flight speed, payload, and airflow. It is based on the motor power formula. ,in For efficiency coefficient, Let A be the air density and A be the rotor area.

[0044] Let the signal strength of the k-th UAV at the m-th monitoring point be denoted by the signal strength, which is affected by the distance to the base station and obstruction by obstacles; the Log-distance path loss model is used for calculation. ; in For transmission power, Environmental degradation factor (3.5 for mountainous areas); This is the margin for shadow fading.

[0045] For the optimization weights of energy consumption and signal strength, satisfy Dynamically adjust strategy: Plains area: Prioritize reducing energy consumption.

[0046] Mountainous areas: The focus is on ensuring communication quality.

[0047] In severe weather: An additional 20% weighting is added to address signal fluctuations.

[0048] Traditional path planning only considers the shortest distance, leading to a 15% data loss during mountain flights due to weak signals. This model introduces a reciprocal term for signal strength. By penalizing low-signal areas in the objective function, the drone actively chooses the path closer to the base station, increasing the data integrity rate from 82% to 99.5%.

[0049] It is worth mentioning that this embodiment proposes a UAV path optimization model that integrates energy consumption and signal strength for multi-objective decision-making. This model can calculate the UAV's energy consumption curve based on motor power, flight speed, and payload; construct a signal strength heatmap by combining base station distribution maps to avoid communication blind spots; and increase the β weight (e.g., α:β=3:7) in complex scenarios such as mountainous areas to ensure communication stability. This achieves a 40% increase in UAV endurance utilization, extends the coverage radius of a single mission to 8 kilometers through energy consumption optimization, improves data collection integrity from 82% to 99.5%, and effectively avoids the risk of communication interruption through signal strength constraints.

[0050] In traditional urban and rural governance systems, the spatiotemporal reference of video and sensor data is inconsistent, resulting in a high fusion error rate; low-resolution aerial images struggle to identify small-scale targets (such as details of illegal construction sites), with a false negative rate exceeding 15%. Therefore, the multi-source data fusion algorithm includes the following steps: Perform moving target detection on high-resolution video data and extract target position coordinates. and movement speed ; Super-resolution reconstruction of drone aerial images to extract target feature vectors ; Constructing a spatiotemporal matrix from sensor data And it is spatiotemporally aligned with the video data; The Kalman filter is used to correlate and match multi-source target information, and the fused target trajectory is output. .

[0051] It is worth mentioning that this solution defines a four-step processing flow for the multi-source data fusion algorithm, specifically including: using the YOLOv5 model to perform real-time target recognition on high-resolution video data and outputting target coordinates. and movement speed It achieves a detection frequency of 30 frames per second; the ESRGAN algorithm is applied to drone aerial images to upscale the resolution from 720p to 4K, and SIFT feature vectors are extracted. Constructing a sensor spatiotemporal matrix The sensor data and video stream timestamps and geographic coordinates are aligned using Gaussian process regression, with the alignment error controlled within ±0.5 seconds. Kalman filters are used to fuse multi-source trajectory data, and a manual verification mechanism is triggered when the coordinate deviation between the high-position camera and the drone for the same target exceeds 5 meters.

[0052] The technical effects of this embodiment include: improving the accuracy of target association; the spatiotemporal alignment technology reduces the fusion error rate to less than 0.3%; and the super-resolution reconstruction reduces the effective target size of UAV images from 5cm×5cm to 2cm×2cm.

[0053] In complex environments such as mountainous areas, fluctuations in drone image transmission bandwidth can lead to a keyframe loss rate as high as 12%. Furthermore, traditional edge nodes only support full-frame compression, resulting in significant loss of important target details. Therefore, the edge computing node performs the following operations: The video stream is compressed in real time during the drone's flight, and the compression ratio is dynamically adjusted according to the link bandwidth. The acquired images undergo local feature enhancement processing, and image blocks containing moving targets are transmitted first. When a communication interruption is detected, a local caching mechanism is activated and data is stored in order of event priority. It is worth mentioning that this embodiment defines three core functions of an edge computing node, including: Based on the H.265 encoding standard, ROI (Region of Interest) compression is automatically enabled according to the current bandwidth (e.g., when bandwidth < 10Mbps), maintaining the original image quality of moving target blocks and increasing the compression ratio of background areas to 50:1; Histogram equalization is performed on the moving target region in the image, and the Laplacian operator is used to enhance edge features, thereby improving the confidence of subsequent AI recognition. When communication is interrupted, data is stored according to event priority (level 1 > level 2 > level 3), the buffer capacity supports continuous data storage for 8 hours, and automatic retransmission is performed after the link is restored.

[0054] This design can improve bandwidth utilization by 35%, maintain the image quality PSNR value of critical target blocks above 40dB, and reduce the buffered data loss rate during communication interruptions from 22% to 0.5%, ensuring complete evidence collection in disaster events.

[0055] Traditional fixed-space drone hangar deployments can easily lead to resource waste and increase the cost of covering mountainous areas; If the arrival time of drones during emergency response exceeds 15 minutes, the optimal window for handling the situation may be missed. Therefore, the deployment density of the drone hangar is determined according to the following formula: , in, Indicates the area under its jurisdiction; Indicates the event occurrence rate per unit area; This refers to the drone's cruising speed. This represents the maximum allowable response time for the system.

[0056] In the above formula This represents the area of ​​the administrative region to be covered, and a terrain correction factor needs to be considered.

[0057] Plains area: ; Hilly areas: Due to the increased path length caused by detours, a coefficient of 1.2 is required.

[0058] Mountainous areas: To avoid blind spots caused by mountain peaks, the coefficient needs to be multiplied by 1.5.

[0059] The average daily event occurrence rate per unit area is calculated by fitting historical work order data; specifically, the calculation method is to take the total number of events over the past 365 days. ,but Based on different quarters Adjustments were made to adapt to seasonal changes.

[0060] For the drone's cruising speed, a margin for headwind compensation needs to be reserved; for example, the DJI M300's flight speed is normally set at 12m / s, but it drops to 8m / s when there is a headwind of level 5.

[0061] The maximum permissible time from the occurrence of an incident to the arrival of the drone at the scene is set according to the incident level.

[0062] The specific grading criteria are as follows: Level 1 event: , Level 2 event: , Level 3 event: .

[0063] The traditional approach of deploying hangars at 10km intervals leads to resource idleness (utilization rate <50%) in sparsely populated areas. This formula quantifies the correlation between event density and response time, matching the number of hangars deployed to actual governance needs. For example, consider a mountainous town with an area of ​​150km². 2 , ,Require ,Pick ,but ; In practice, only 2 hangars need to be deployed (rounded up + redundant design), which reduces costs by 60% compared to the original solution (5 hangars required based on spacing).

[0064] It is worth mentioning that this embodiment provides a calculation formula for the deployment density of drone hangars. This design can reduce the construction cost of hangars and achieve a response time of ≤8 minutes in 95% of the area through density optimization. The drone reachability in mountainous areas is increased from 75% to 98%.

[0065] Traditional inspection methods suffer from rigid manual inspection plans, with 20% of inspection tasks resulting in no actual events occurring; multi-dimensional data displays are fragmented, requiring decision-makers to compare information across systems. Therefore, this embodiment provides a method applied to a comprehensive sensing system of a full-domain intelligent interconnection platform as described in any of the above, including: S1. Video data, aerial images and environmental parameters are collected synchronously through the high and low altitude linkage perception layer, and the raw data is transmitted to the AI ​​scheduling layer through the data link layer. S2. The cloud server calls the dynamic path optimization model to generate the optimal flight path for the drone swarm and sends image processing instructions to the edge computing nodes; S3. The application layer receives the processed data, merges it in the panoramic view module to generate a three-dimensional situation map, and marks abnormal events through the cockpit monitoring module; S4. The closed-loop management module initiates a tiered response mechanism based on the event type, pushes handling work orders to the three levels of management departments, and tracks the handling results to update the system database.

[0066] It is worth mentioning that this solution defines the four-stage process of the full-domain intelligent connectivity approach, including: Data acquisition: The high-position camera scans periodically at an elevation angle of 30°–60°, the drone cruises along a preset route, and the sensor array uploads environmental parameters once per minute; Path optimization: The cloud dynamically divides the drone responsibility grid based on historical heat maps, and a relay inspection mode is activated during peak mission periods; Situation generation: A 3D map is overlaid with a meteorological warning layer (e.g., areas with rainfall > 50 mm / h are marked in red) and fused with population density data; Closed-loop management: Repeated incidents (such as street vending at the same location 3 times a week) are automatically upgraded to special governance tasks.

[0067] The technical effects of this embodiment include: reducing invalid inspection tasks by 65% ​​and increasing the proportion of effective drone task time to 82%; and increasing the speed of cross-dimensional data correlation analysis by 4 times, supporting real-time rendering of more than 10 data layers.

[0068] The generation of the optimal flight path described in S2 includes: constructing a probability map of hotspot areas based on historical work order data; using the ant colony algorithm to solve the multi-drone coverage path problem; and dynamically adjusting the flight altitude to avoid meteorological risk areas.

[0069] It is worth mentioning that this solution provides three strategies for generating optimal flight paths, which can effectively solve the technical problems of 50% of UAV energy being wasted in low-value areas under the current fixed route mode, and 15% of flight missions being canceled midway due to sudden weather changes.

[0070] The triggering condition for the hierarchical response mechanism described in step S4 is: When the sensor detects that the PM2.5 concentration exceeds the threshold for 3 consecutive hours, the environmental protection department's linkage process is automatically initiated. If the same coordinate point is identified as having illegal construction features in different data sources, the event level will be raised to level two. For work orders that are not processed within the specified time, an assessment report will be generated and sent to the supervisory department.

[0071] It is worth mentioning that this solution provides a trigger condition design for a graded response mechanism, which can solve the technical problems in traditional technical solutions where 30% of high-risk events are underestimated due to subjective errors in human judgment of event levels, as well as the lack of quantitative assessment of response effectiveness and insufficient motivation for departmental collaboration.

[0072] Unless otherwise specified, the equipment components involved in the above embodiments are all conventional equipment components, and the connection methods and control methods involved are all conventional connection methods and control methods unless otherwise specified.

[0073] The present invention has been described in detail above with reference to the embodiments. However, those skilled in the art will understand that, without departing from the spirit of the present invention, various specific parameters in the above embodiments can be changed to form multiple specific embodiments, all of which are common variations of the present invention, and will not be described in detail here.

Claims

1. A comprehensive sensing system for a global intelligent interconnection platform, characterized in that, include: ‌ The system comprises a high- and low-altitude linkage sensing layer, a data link layer, an AI scheduling layer, an application layer, and a closed-loop management module. The high- and low-altitude linkage sensing layer includes high-position video surveillance equipment deployed at tower base stations, multi-rotor drones in drone hangars, and IoT sensor arrays deployed on the ground. The high-position video surveillance equipment is coupled with a multispectral camera through a gimbal with adjustable pitch angle, covering a monitoring area with a radius of not less than 500 meters. The drone hangar has built-in charging piles and meteorological monitoring units. The drones are equipped with edge computing modules and dual-light cameras. The IoT sensor array includes acoustic sensors, air quality sensors, and vibration sensors. The data link layer constructs a hybrid transmission network through a 5G communication module and an optical fiber transmission channel to realize real-time data transmission and command interaction between the high-position video surveillance equipment, drones and IoT sensors. The AI ​​scheduling layer includes a cloud server cluster and edge computing nodes. The cloud server cluster has a built-in dynamic path optimization model and a multi-source data fusion algorithm. The edge computing nodes are deployed in the drone hangar and are used to process the raw image data collected by the drone in real time.

2. The comprehensive sensing system for a full-domain intelligent interconnection platform according to claim 1, characterized in that, The application layer includes a panoramic view module, a cockpit monitoring module, and an event handling module. The panoramic view module dynamically integrates high-position video, UAV aerial footage, and sensor data based on a geographic information system to generate a 3D visualization map. The cockpit monitoring module predicts the probability of equipment failure and generates maintenance work orders through a time series analysis model. The closed-loop management module includes an event recognition engine and a three-level feedback network. The event recognition engine extracts features from multi-source data through a convolutional neural network and matches them with a preset rule base to generate alarm events. The three-level feedback network pushes alarm events to town and street management departments, district-level command centers and municipal-level supervision platforms according to priority, and receives the processing results and sends them back to the system database.

3. The comprehensive sensing system for a fully intelligent interconnected platform according to claim 1, characterized in that, The data transmission rate of the data link layer satisfies the following constraints: ， in, This indicates the minimum bandwidth requirement for a hybrid transmission network; This represents the data volume of a single frame in the i-th video stream; Indicates the maximum allowable latency for video frame transmission; This represents the amount of data sampled in a single run by the j-th type of sensor; This indicates the sensor sampling period.

4. The comprehensive sensing system for a full-domain intelligent interconnection platform according to claim 1, characterized in that, The objective function of the dynamic path optimization model is: ， in, This represents the estimated energy consumption of the k-th drone; This represents the signal strength of the k-th UAV at the m-th monitoring point; and Let be the weight coefficient, and satisfy... The constraints include the maximum flight time of the drone, no-fly zone avoidance, and data collection integrity thresholds.

5. The comprehensive sensing system for a full-domain intelligent interconnection platform according to claim 1, characterized in that, The multi-source data fusion algorithm includes the following steps: Perform moving target detection on high-resolution video data and extract target position coordinates. and movement speed ; Super-resolution reconstruction of drone aerial images to extract target feature vectors ; Constructing a spatiotemporal matrix from sensor data And it is spatiotemporally aligned with the video data; The Kalman filter is used to correlate and match multi-source target information, and the fused target trajectory is output. .

6. The comprehensive sensing system for a full-domain intelligent interconnection platform according to claim 1, characterized in that, The edge computing node performs the following operations: The video stream is compressed in real time during the drone's flight, and the compression ratio is dynamically adjusted according to the link bandwidth. The acquired images undergo local feature enhancement processing, and image blocks containing moving targets are transmitted first. When a communication interruption is detected, a local caching mechanism is activated and data is stored in order of event priority.

7. The comprehensive sensing system of the all-domain intelligent interconnection platform according to claim 1, characterized in that, The deployment density of the drone hangar is determined according to the following formula: ， in, Indicates the area under its jurisdiction; Indicates the event occurrence rate per unit area; This refers to the drone's cruising speed. This represents the maximum allowable response time for the system.

8. A method applied to a comprehensive sensing system for a global intelligent interconnection platform as described in any one of claims 1-7, characterized in that, include: S1. Video data, aerial images and environmental parameters are collected synchronously through the high and low altitude linkage perception layer, and the raw data is transmitted to the AI ​​scheduling layer through the data link layer. S2. The cloud server calls the dynamic path optimization model to generate the optimal flight path for the drone swarm and sends image processing instructions to the edge computing nodes; S3. The application layer receives the processed data, merges it in the panoramic view module to generate a three-dimensional situation map, and marks abnormal events through the cockpit monitoring module; S4. The closed-loop management module initiates a tiered response mechanism based on the event type, pushes handling work orders to the three levels of management departments, and tracks the handling results to update the system database.

9. A method according to claim 8, characterized in that, The generation of the optimal flight path described in S2 includes: constructing a probability map of hotspot areas based on historical work order data; using the ant colony algorithm to solve the multi-drone coverage path problem; and dynamically adjusting the flight altitude to avoid meteorological risk areas.

10. A method according to claim 8, characterized in that, The triggering condition for the hierarchical response mechanism described in step S4 is: When the sensor detects that the PM2.5 concentration exceeds the threshold for 3 consecutive hours, the environmental protection department's linkage process is automatically initiated. If the same coordinate point is identified as having illegal construction features in different data sources, the event level will be raised to level two. For work orders that are not processed within the specified time, an assessment report will be generated and sent to the supervisory department.

Images