A vehicle personnel management and control method and system based on air-ground integrated data

Abstract

The application relates to the technical field of traffic control systems and discloses a vehicle and personnel management and control method and system based on air-ground integrated data, which comprises the following steps: a digital map taking a unique, geography-coordinate-independent topological road section identifier as a reference is constructed, a ground unit perceives and broadcasts a traffic flow state bound with the identifier, an air unit receives the state information and updates road section attributes on a preset topological map, and then, cooperative management is performed; through the conversion of space-time anchoring logic from unstable geometry coordinate measurement into identification of static topological logic relations, the application avoids systematic misalignment risks caused by physical world uncertainty, makes the system obtain original reliability and self-adaptive capacity in a complex signal environment such as a multi-layer overpass, and further opens a new way of low-cost and high-efficiency cooperative management.

Description

Technical Field

[0001] The present invention relates to a vehicle and personnel control method and system based on air-ground integrated data, belonging to the technical field of traffic control. Background Art

[0002] In the field of traffic control system technology, the existing air-ground integrated management and control solutions generally adopt a technical approach that requires aerial units such as drones and ground units such as cameras to first obtain the precise geographic coordinates of targets such as vehicles and personnel, and then uniformly transmit this coordinate data to a central server for data fusion, trajectory association, and risk decision-making.

[0003] However, when this positioning-first, fusion-later approach is applied to complex three-dimensional traffic environments such as urban multi-story interchanges or tunnel clusters, its over-reliance on precise coordinates constitutes a fundamental technical bottleneck. Specifically, due to physical uncertainty factors such as the attenuation of satellite signals between high-rise buildings and bridges, multipath reflection effects, and inherent errors in the conversion between different coordinate systems in the air and on the ground, coordinate information is inevitably inaccurate and lost.

[0004] This bottleneck directly translates into an objective information barrier and systemic risk at the application level. Specifically, when the central system fuses the aerial coordinates of a vehicle on a bridge with the ground coordinates of a vehicle under the bridge, their projected positions may overlap significantly, frequently leading to target confusion or trajectory breakage, ultimately resulting in erroneous congestion judgments and traffic diversion instructions. This causes the reliability and efficiency of the control system to plummet in the complex scenarios where it is most needed. Therefore, the technical problem to be solved by this invention is how to fundamentally break away from the reliance on precise geographic coordinate systems, mechanically break through the information fusion bottleneck caused by inaccurate coordinates, and eliminate the risk of systemic misjudgment in complex three-dimensional traffic environments. Summary of the Invention

[0005] The present invention provides a vehicle and personnel control method based on integrated air-ground data. Its main purpose is to solve the technical problem that the existing technology relies too much on precise coordinates, resulting in difficulty in integrating air-ground data in complex three-dimensional traffic environments, prone to misjudgment of the system, and unable to achieve reliable and efficient collaborative control.

[0006] To achieve the above objectives, the present invention provides a vehicle and personnel management method based on air-ground integrated data, the method comprising the following steps:

[0007] Step a: constructing a road network topology map. In the topology map, each minimum directed road segment is assigned a unique topology road segment identifier, which is independent of the geographic coordinates.

[0008] Step b: at least one ground unit senses the traffic flow state of the topological road section where it is located or monitored, where the traffic flow state includes one of unblocked, slow-moving and congested. The ground unit determines whether the traffic flow state meets the conditions for transitioning to a more congested state based on a set traffic flow characteristic threshold. If so, the ground unit monitors and receives the current traffic flow state of at least one upstream associated topological road section. Only when the current traffic flow state of at least one upstream associated topological road section meets the congestion condition or the slow-moving condition, the ground unit binds the determined traffic flow state to the corresponding topological road section identifier to form state-based topological information.

[0009] Step c: the ground unit broadcasts stateful topology information via a low-power wireless network;

[0010] In step d, at least one aerial unit receives stateful topological information. The aerial unit determines the current topological area based on its own imprecise location information and monitors the stateful topological information within the area. The aerial unit updates the visual mark of the road section in its preset topological map based on the received topological road section identifier and the corresponding traffic flow status. The aerial unit performs control operations related to the topological road section identifier based on the traffic status represented by the updated visual mark. The control operations include projecting visual diversion instructions or issuing voice prompts, thereby realizing collaborative control of vehicles and personnel based on discrete topological states.

[0011] Preferably, the traffic flow status in step b is determined based on the average vehicle spacing obtained by image recognition, or the number of vehicles passing per unit time collected by the coil sensor, and the average spacing or the number of vehicles passing is compared with a set threshold value. The control operations performed by the aerial unit based on the traffic flow status include: when the traffic flow status is congested, projecting a red light band or issuing a congestion warning prompt; when the traffic flow status is unobstructed, projecting a green light band or issuing a pass instruction.

[0012] Preferably, the method also includes: a traffic control facility as a ground unit, when it receives stateful topological information that its downstream or associated topological section is declared to be in a congested state, it automatically adjusts the green light time of the direction it controls according to the set adjustment rules to limit or guide traffic flow.

[0013] Preferably, in step b, when the ground unit determines that the traffic flow state meets the conditions for transitioning to a higher congestion state, if the traffic flow state of its upstream associated topological section does not meet the congestion or slow-moving conditions, the ground unit does not broadcast a higher congestion state, but instead broadcasts an auxiliary information indicating that a micro-disturbance event exists in the topological section.

[0014] Preferably, after receiving the auxiliary information indicating the micro-disturbance event, the aerial unit does not perform the congestion relief and control operation, but performs a lightweight warning action, which includes projecting a yellow pulse laser over the topological road section to alert the following vehicles to possible instantaneous abnormalities ahead.

[0015] Preferably, the method further comprises: the aerial unit projects a reference laser pattern onto the topological road section it monitors; capturing the actual image formed by the reference laser pattern on the road surface through the image sensor of the aerial unit; and calculating the edge diffusion index of the actual image. , The calculation formula is: ,in, is the pixel width of the actual imaging edge, The theoretical pixel width of the reference laser pattern under ideal conditions; when the edge diffusion index When the set threshold is exceeded, the laser projection system of the aerial unit automatically switches to high-contrast working mode, which includes dual-wavelength compensation and high-frequency pulse flashing mode.

[0016] Preferably, the method also includes: setting a rotatable polarizer in front of the image sensor of the aerial unit; controlling the rotatable polarizer to collect reflected light images of the topological section monitored by the aerial unit at at least two different polarization angles; determining the physical state of the road surface of the topological section by comparing the average brightness difference between the images collected at different polarization angles; and associating the physical state of the road surface with the topological section identifier, the physical state of the road surface including water accumulation or ice.

[0017] Preferably, the aerial unit adjusts the form of the visual evacuation instructions it projects based on the associated physical state of the road surface; when the physical state of the road surface is slippery, the projected green light band is modified to a dotted line or a slow flashing form.

[0018] Preferably, the topological road segment identifier is in text form, including the physical location information, direction information and lane information of the road segment.

[0019] A vehicle and personnel management and control system based on air-ground integrated data, characterized in that the system includes:

[0020] A map construction module is configured to construct a road network topology map, wherein each minimum directed road segment in the topology map is assigned a unique topology road segment identifier, and the identifier is independent of the geographic coordinates;

[0021] At least one ground sensing and information broadcasting unit is configured to sense the traffic flow state of the topological road section where it is located or monitored, the traffic flow state including one of unblocked, slow-moving and congested; and based on a set traffic flow characteristic threshold, determine whether the traffic flow state meets the conditions for transitioning to a more congested state; if it is determined to be satisfied, the ground sensing and information broadcasting unit monitors and receives the current traffic flow state of at least one upstream associated topological road section; only when the current traffic flow state of at least one upstream associated topological road section meets the congestion condition or the slow-moving condition, the ground sensing and information broadcasting unit binds the determined traffic flow state with the corresponding topological road section identifier to form stateful topological information; and is configured to broadcast the stateful topological information through a low-power wireless network;

[0022] At least one airborne receiving and control execution unit is configured to receive stateful topological information. The airborne receiving and control execution unit determines the current topological area based on its own non-precise location information and listens to the stateful topological information in the area; it is configured to update the visual mark of the road section in its preset topological map according to the received topological road section identifier and the corresponding traffic flow status; and it is configured to execute control operations related to the topological road section identifier based on the traffic status represented by the updated visual mark. The control operations include projecting visual diversion instructions or issuing voice prompts, thereby realizing vehicle and personnel collaborative control based on discrete topological states.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. The present invention constructs a road network topology map composed of unique topological road segment IDs that are independent of geographic coordinates, and uses this as the sole cornerstone for status information exchange between aerial and ground units, thereby transforming the spatiotemporal anchoring logic of the entire management and control system from dependence on unstable and continuously changing geometric measurement coordinates to the recognition of static and discrete topological logical relationships. This fundamental shift in approach enables the present invention to mechanically avoid the risk of system misalignment caused by inherent uncertainties in the physical world such as satellite signal obstruction, multipath reflection, or coordinate system conversion errors, ensuring that it can still maintain efficient and stable management and control capabilities in traditional signal black hole areas such as multi-layer interchanges and tunnel groups, and realizing the system's native immunity to complex spatial environments.

[0025] 2. The present invention constructs a distributed, self-organizing control network by using ground units to judge and actively broadcast the traffic flow status of the topological section where they are located, and aerial units or other control facilities to directly perform coordinated actions based on the received local status information. This architecture transforms the serial, long-chain response mode of central thinking and terminal execution in the existing technology into a parallel, adaptive response mode based on neighborhood information interaction. The decision on congestion relief is no longer the result of delayed calculation by the central server, but the orderly traffic flow that emerges at a macro level, thereby fundamentally improving the system's response speed and flexibility to emergencies, and making the system's operating logic qualitatively change from passive calculation to active emergence.

[0026] 3. By internalizing the complex spatial relationships of the road network into a one-time topological map and simplifying real-time traffic status judgment into discrete three-state classifications, the performance requirements for the system's hardware units are greatly reduced. This design allows both the air and ground units that make up the system to adopt mature, low-cost, and affordable components. This fundamentally avoids the existing technology's continued reliance on high-cost, high-precision sensors, high-computing fusion centers, and complex calibration processes, providing a new, economically viable technical path for the large-scale deployment of advanced integrated air-ground control capabilities into the capillary network of urban transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a timing diagram of the vehicle and personnel control method based on air-ground integrated data and the system interaction;

[0028] Figure 2 This is a graph showing the image contrast changes of the laser projection system of the present invention at different aerosol concentrations;

[0029] Figure 3 This is the architecture diagram of the vehicle and personnel management system based on air-ground integrated data of the present invention;

[0030] Figure 4 This is a relationship diagram between ESI value and atmospheric visibility of the present invention;

[0031] Figure 5 This is the vehicle and personnel control road network topology diagram based on air-ground integrated data of the present invention.

[0032] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0034] Taking a typical urban complex overpass application scenario as an example, the complete operation process of the vehicle and personnel control method based on air-ground integrated data provided by the present invention is elaborated in detail. The method mainly includes the construction and calibration stage of the road network topology map, the traffic status perception and broadcasting stage based on the collaboration of ground and air units, and the adaptive control execution stage of the air unit. First, in the initial stage of system deployment, the construction and calibration of the road network topology map are performed. The purpose of this step is to establish a stable and reliable information interaction benchmark independent of the geographic coordinate system, so as to fundamentally avoid the risk of systematic misjudgment caused by inaccurate satellite positioning signals in the existing technology. The specific procedures are as follows: for the physical road network that needs to be controlled, such as a three-story overpass area containing multi-layer ramps and main roads, it is logically disassembled into a series of minimized directed road sections with unique traffic directions. Each directed road section is regarded as an independent node in the topology map; then, a globally unique topological road section identifier encoded in text form is assigned to each such directed road section node. The identifier structure is designed to contain sufficient information for the system to clearly identify it. Its composition includes physical location information, direction information, and lane information. For example, the identifier of a directional road section about 200 meters long located in the second lane of the north-south main road of a Kaixuan Interchange can be deterministically encoded as Kaixuan-Interchange-SB-Lane2, where SB stands for Southbound. In this way, even if the road section overlaps with the underlying road in the vertical direction, its identifier is absolutely unique and distinguishable, thus laying the data foundation for subsequent air-ground information interaction. Furthermore, the system enters the continuous operation of traffic status perception and broadcasting. This stage is completed by the ground units and aerial units distributed on site in collaboration, aiming to build a decentralized, rapidly responsive distributed intelligent network. A typical ground unit, such as a camera with basic image processing capabilities installed on a road gantry or a coil sensor buried under the road surface, is responsible for continuously sensing the traffic flow status of a specific topological road section within its monitoring range.

[0035] To address the problem of inaccurate vehicle size recognition caused by factors such as lighting and shadows in image recognition, this solution does not rely on accurate speed measurement when determining traffic flow status, but relies on more robust traffic flow characteristic parameters. For example, for the image recognition unit, its core algorithm is configured to calculate the average distance between vehicles in the field of view, while for the coil sensor unit, it counts the number of vehicles passing per unit time; to convert these continuous physical quantities into discrete, broadcastable states, the system presets a set of traffic flow characteristic thresholds. The calibration process of this threshold is designed as follows: during the test period after the system is deployed, at least 72 consecutive hours of traffic flow data are collected and combined with the design of the road section. The traffic capacity is calculated, and the critical values ​​that can distinguish between the three states of smooth flow, slow flow and congestion are determined by traffic engineers. For example, for the section with the identifier Kaixuan-Interchange-SB-Lane2, the following rules can be calibrated: if the number of vehicles passing through the section per minute is less than ten, it is defined as smooth flow; if it is between ten and thirty, it is defined as slow flow; and if it is greater than or equal to thirty, it is defined as congestion; when a ground unit determines through the above rules that the traffic flow state of the monitored section meets the conditions for transitioning to a higher congestion state, such as from slow flow to congestion, the unit will not immediately broadcast this congestion state, but will trigger a pre-set logical judgment procedure, which is designed to distinguish real traffic waves Spread and isolated micro-disturbance events, specifically, the ground unit will temporarily enter the monitoring mode and receive the current traffic flow status broadcast of one or more upstream topological sections with which it is preset. Only when it monitors that at least one upstream section is also in a slow or congested state, the system will determine that the congestion is a conductive traffic event, and formally bind the determined congestion state with its own topological section identifier to form a state-based topological information such as Kaixuan-Interchange-SB-Lane2, congestion, and then periodically broadcast it to the surrounding area through a low-power wireless network; on the contrary, if all its upstream sections are broadcast as unblocked, the system will determine that the current event is For an isolated, non-conductive micro-disturbance, the ground unit will not broadcast the congestion status, but instead broadcast an auxiliary information indicating the presence of a micro-disturbance event on the topological road section. This design greatly improves the accuracy of the system's judgment and the macro-intelligence of the entire traffic flow response. Correspondingly, aerial units such as drones cruising in the air play a key role in receiving information and executing control. Based on the imprecise location information provided by the ordinary global positioning system module on board, the aerial unit only needs to determine that it is in the approximate topological area of ​​the Triumph Interchange, without knowing the precise latitude and longitude coordinates. Subsequently, it continuously monitors the status topological information broadcast by all ground units in the area.To further illustrate the core features of the topological road segment identifiers in this invention, it's important to note that their independence from geographic coordinates specifically refers to their independence from real-time dynamic geometric coordinates, acquired through global satellite navigation systems and other means, which are susceptible to environmental interference and inaccuracy. In contrast, the physical location information contained in the identifiers, such as the Triumph Interchange, is essentially a static semantic label used for human-computer interaction and initial system calibration. This transformation, replacing uncertain dynamic numerical measurements with deterministic static logical references, fundamentally ensures the robustness of the system in complex environments.

[0036] The onboard computer of the aerial unit is pre-installed with a road network topology map that is exactly the same as that of the ground end. When it receives a state-based topology information of congestion, such as Kaixuan-Interchange-SB-Lane2, it will immediately query its internal map and update the visual mark of the corresponding road section to red. The updated topology map intuitively presents the traffic situation in the area to the aerial unit; based on this, the aerial unit performs control operations related to the topology road section identifier according to the traffic status represented by the updated visual mark. If the status is congested, the aerial unit will fly over the road section and use its The laser projection device on board projects a striking red light band onto the road surface, and at the same time issues a voice prompt through the loudspeaker that there is congestion ahead and the driver should change lanes in advance. If the traffic is unobstructed, a green light band can be projected to instruct vehicles to pass safely. When the aerial unit receives auxiliary information indicating a micro-disturbance event, in order to avoid overreaction and interference with normal traffic flow, it will not perform the above-mentioned congestion relief and control operations, but will perform a lightweight early warning action, which is specifically reflected in the projection of a yellow pulse laser in the direction of oncoming vehicles above the topological section to alert the vehicles behind to possible instantaneous abnormalities ahead and guide them to actively and smoothly slow down and avoid them.

[0037] In addition, the present invention also includes a series of advanced procedures designed to improve the adaptability and reliability of the system in complex environments. One of them is an adaptive visual guidance adjustment mechanism for changes in environmental visibility. In view of the fact that low-visibility weather such as fog, haze or rainy days will seriously affect the clarity of laser projection instructions, the aerial unit is configured to periodically execute a self-inspection program called visual channel quality assessment when performing tasks. The specific operation process of this program is as follows: the aerial unit first projects a preset reference laser pattern with sharp edges, such as a crosshair, onto the road surface of the monitored road section directly below it; then, the actual image formed by the reference laser pattern on the road surface is captured by the onboard image sensor, and the image processing algorithm is called to calculate the pixel width of the actual imaging edge. At the same time, based on the current flight altitude and lens parameters, the system can calculate the theoretical pixel width of the laser pattern under ideal non-scattering conditions. ; By calculating the edge diffusion index The system can quantify the degree of scattering of light by the current atmosphere. When the index exceeds a preset threshold, such as 1.8, it means that the actual edge width of the light spot is eighty percent wider than the theoretical value. The system determines that the quality of the visual channel has deteriorated, and its laser projection system will automatically switch to a high-contrast working mode. This mode includes enabling dual-wavelength compensation, that is, simultaneously projecting red and green lasers to enhance chromatic aberration contrast, and enabling high-frequency pulse flashing mode, which can be set to flash at a frequency of 5 Hz, thereby maximizing the perceptibility of visual instructions in low-visibility environments. Another advanced procedure is the perception and response capability of the physical state of the road surface. In order to cope with the slippery hazards such as water accumulation or ice on the road surface after rain and snow, a set of motor-driven rotating polarizer systems is configured in front of the image sensor of the aerial unit. When conducting a safety assessment of a specific topological section, the aerial unit will perform the following detection process: First, control the rotatable polarizer to rotate to 0 degrees to collect a frame of reflection of the topological section. light image; then, the polarizer is quickly controlled to rotate 90 degrees to capture a second frame of image; because the reflection of light on water or ice has strong polarization characteristics, while dry asphalt pavement is mainly diffuse reflection with very weak polarization characteristics, the road condition can be effectively judged by comparing the average brightness difference between images collected at two orthogonal polarization angles. If the brightness difference exceeds a calibrated threshold, such as 45%, the system will determine the physical condition of the road surface of the topological section as slippery and associate this condition with the corresponding topological section identifier; when the aerial unit needs to provide a passage instruction for a section associated with a slippery state, even if the traffic flow state of the section is unobstructed, it will also adaptively adjust the projected visual diversion instruction form. Specifically, the continuous green light band originally used to indicate passage will be automatically modified to a dotted line or slow flashing form, warning of potential driving risks in a way that is not easy to alert the driver but can effectively convey a cautious driving signal.

[0038] Finally, the distributed control network constructed by the present invention also has a certain self-organizing and coordination ability. For example, a traffic control facility as a ground unit, such as a traffic light controller in front of a downstream entrance ramp, after receiving the stateful topological information that its downstream or associated topological section, such as Kaixuan-Interchange-SB-Main, is declared to be in a congested state, it will automatically adjust the green light time of the direction it controls according to the preset adjustment rules. For example, the original green light time of 60 seconds will be reduced by 30% to actively limit the traffic flow entering the congested area, thereby assisting the main road in diverting traffic. This collaborative control based on local information interaction greatly shortens the response chain of the entire traffic system, realizing a qualitative change from passive calculation to active emergence, and ultimately realizing a low-cost, high-efficiency, and natively reliable air-ground integrated vehicle and personnel collaborative control system in complex environments without relying on expensive high-precision sensors and central computing platforms.

[0039] Example 1: This example is used in a continuously running urban three-dimensional transportation hub control scenario, and its specific operation and synergy effects are described as follows. The scenario is a Kaixuan Interchange area with three layers of vertically overlapping road networks. Under low-visibility meteorological conditions of moderate rain at night, its inherent spatial structure and satellite signal shielding effect cause the performance of traditional control methods that rely on precise geographic coordinates to decline. In this environment, a set of air-ground integrated control systems deployed according to the present invention continues to operate. In the road network topology map that serves as the information interaction benchmark of this system, a key ramp on the top layer of the interchange has been uniquely marked as the topological road segment identifier Kaixuan-Interchange-SB-Lane2, and a ground unit has been deployed on this road segment, while an aerial unit cruises over this topological area; at a certain moment, a vehicle in the road segment is momentarily stopped due to a malfunction, and the traffic flow characteristic parameters monitored by the ground unit immediately meet the conditions for transitioning to a more congested state, but the The ground unit did not immediately declare congestion, but executed its built-in pre-logical judgment procedure, that is, monitoring the traffic flow status of the upstream topological section with which it was preset. Given that the upstream section was still unobstructed at this time, the system determined that this was an isolated micro-disturbance event. The ground unit then bound an auxiliary information carrying this judgment to the topological section identifier Kaixuan-Interchange-SB-Lane2 and broadcast it. Correspondingly, the aerial unit cruising in the interchange area, after receiving this auxiliary information indicating a micro-disturbance event, triggered its internal preset collaborative control rules. The system did not perform heavy congestion relief operations such as projecting red light bands to avoid triggering a chain reaction of vehicles behind on slippery roads, but instead initiated a lightweight early warning action. At the beginning of this action, the aerial unit first executed a visual channel quality assessment self-inspection procedure, which projected a reference laser pattern onto the road surface and calculated the edge diffusion index through the captured actual imaging. , the actual measured value of the index exceeded the system calibration threshold due to rain and fog scattering, so the laser projection system automatically switched to high-contrast working mode and projected a yellow laser with dual-wavelength compensation and high-frequency pulse flashing to the road surface in the direction of the oncoming vehicle upstream of the accident point; the ground unit refined the original congestion representation into a highly certain event type input based on the upstream state association judgment mechanism, while the adaptive visual guidance adjustment mechanism of the air unit ensured that the early warning information of the event was effectively deployed. The linkage of the two technical features achieved precise intervention in emergencies while avoiding secondary risks. In other words, as the faulty vehicle was not cleared in time, the micro-disturbance eventually evolved into real traffic congestion. After monitoring that the upstream section of the road had also entered a slow-moving state, the ground unit officially broadcast Kaixuan Interchange-SB-Lane2 is congested. Upon receiving this stateful topological information, a traffic control facility located at the downstream entrance ramp, acting as a ground unit, automatically reduces the green light time in the direction it controls, based on its pre-set adjustment rules designed to maintain the main road's capacity. This proactively limits the flow of traffic into the main road without relying on any central computing platform. This mechanism resolves the inherent contradiction between rapidly responding to local emergencies and maintaining road network efficiency. It replaces the need for a computationally intensive global optimal solution with a deterministic, local adjustment rule that depends solely on the downstream topological state. This internalizes the negative externalities of congestion at the directly connected entrance, thereby achieving distributed traffic self-balancing.

[0040] This design approach transforms the spatiotemporal anchoring logic from the measurement of unstable geometric coordinates to the identification of static topological logical relationships, and decomposes the collaborative decision-making process into a series of deterministic local response rules based on discrete states, thereby constructing a control network that is decentralized at the physical level and self-organized at the functional level. This network not only architecturally avoids the risk of systemic misalignment caused by the uncertainty of the physical world, but also transforms the operating logic of the entire system from a preset mode of passive calculation to an actively emergent adaptive mode, giving native reliability and macro-intelligence to traffic control in complex environments.

[0041] Example 2: To verify the effectiveness of the adaptive visual guidance adjustment mechanism of the aerial unit in a low visibility environment, the following quantitative test was performed. The test aims to compare the airborne unit with the edge diffusion index. The closed-loop control logic for visual channel quality assessment and automatic switching of working modes was verified. The test was conducted in a closed environmental simulation chamber that can precisely control the internal aerosol concentration to simulate different levels of foggy environments. The chamber was paved with standard asphalt pavement. A drone, serving as an aerial unit, was fixed on a gantry with precise height adjustment to ensure a constant geometric relationship between it and the ground observation point, thereby eliminating the interference of flight attitude disturbances on the measurement results. The key parameter of the test was the edge diffusion index. The activation threshold The threshold calibration follows the following experimental procedures: First, under the ideal condition of zero aerosol concentration, the airborne unit is operated at a standard operating altitude of ten meters. Project a crosshair-shaped reference laser pattern onto the ground, capture its image with an onboard image sensor, and calculate the ideal pixel width. As a benchmark, aerosol is gradually injected into the cabin, and the current atmospheric extinction coefficient is calibrated using a visibility meter. At each level of extinction coefficient, the system calculates the edge pixel width of the current image in real time. And find out , while a group of observers evaluated the laser pattern's recognizability level, and finally, the critical atmospheric state corresponding to the first time the recognizability fell below 95 percent confidence level was The measured value is determined as the activation threshold under this working condition According to this regulation, the threshold is calibrated to 1.8. At the beginning of the test, the environmental simulation chamber is clean air, and the laser projection system of the aerial unit operates in a standard single-wavelength continuous green laser mode. The crosshair pattern projected on the ground is clearly imaged. The system measures The value fluctuates around 1.02. As the aerosol generator is started, the fog concentration in the cabin increases at a linear rate. The aerial unit captures the laser pattern imaging in real time and continuously calculates It was observed that as the fog thickened, the edge of the laser pattern became blurred and diffused. Significant increase, leading to The value rises steadily when When the real-time calculated value reaches and exceeds the threshold of 1.8, the control logic of the aerial unit is triggered, and its laser projection system automatically switches from standard mode to high-contrast working mode. In this mode, the system simultaneously projects red and green wavelength lasers and pulses at a frequency of 5 Hz. The key performance data records during the entire test process are shown in Table 1.

[0042] Table 1: Aerosol injection timing and its impact on laser operating mode and image contrast.

[0043]

[0044] The test data showed that when the visibility dropped from more than 500 meters to 85 meters, The value shows a strong positive correlation with the degree of environmental degradation, while the image contrast, which is a quantitative indicator of visual instruction recognition, decays from 28.7 to 6.1, entering the difficult-to-recognize range. At the 60th second, After the value exceeds the 1.8 threshold and triggers the high contrast mode, the image contrast immediately jumps to 15.4. Even if the visibility deteriorates further to 50 meters, the contrast still remains at a usable level of 9.8. The direct reason for this phenomenon is that, based on At the critical point where the visual channel quality degrades, the closed-loop control mechanism actively enhances the saliency of the signal source and compensates for the information loss caused by the scattering of the transmission medium, thereby maintaining the recognizability of the command above the safety threshold.

[0045] Example 3: This example combines Figures 1 to 5 , a vehicle and personnel control method and system based on air-ground integrated data is described. Figure 1 As shown, first, after the wireless network receives the stateful topology information, the aerial unit queries the preset topology map and updates the road section markings. Then, the aerial unit calculates and obtains the ESI value (edge ​​diffusion index). If the ESI value is greater than 1.8, the system will switch to high contrast mode, start dual-wavelength compensation and enable 5Hz pulse flashing mode. Subsequently, the laser projection system projects a red light band onto the road surface to issue a congestion warning ahead. If the traffic flow is smooth, a green light band is projected and a pass instruction is issued. If the road surface is wet or in other physical conditions, the laser projection system will project a yellow laser according to relevant conditions for lightweight warning, reminding the following vehicles to pay attention to possible instantaneous abnormalities ahead.

[0046] like Figure 2 As shown in the figure, with the increase of aerosol injection time, the image contrast gradually decreases until it reaches a critical point. In the early stage of aerosol injection (0 to 30 seconds), the image contrast gradually decreases from about 30 to close to 15. During this process, the change in image contrast shows a gradually decreasing trend until the aerosol concentration reaches a certain level, and the image contrast further decreases. When the aerosol injection time exceeds 60 seconds, the image contrast stabilizes at a low level of about 10, which is close to the threshold of the minimum recognizable contrast. The dotted line in the figure represents the minimum recognizable contrast, which is 10 and is shown in the figure as a reference line for measuring changes in image quality. The solid line in the figure represents the actual trend of image contrast changing with aerosol concentration, indicating that as the aerosol concentration increases, the image recognizability gradually decreases and reaches a stable state.

[0047] like Figure 3As shown in the figure, the system includes multiple functional modules and works in conjunction with ground units and aerial units. In the road network topology map layer, various topological road sections are identified, such as Triumph-SB-Lane 1, Triumph-SB-Lane 2, Triumph-NB-Main Road, and Triumph-EB-Main Road. These road sections interact with the aerial unit through the ground unit. The ground unit includes an image head, coil sensor, and signal light. Its main function is to sense the traffic flow status (such as smooth, slow-moving, and congested) and broadcast traffic status information. The aerial unit (such as a drone) receives the information broadcast by the ground unit and conducts real-time control through a laser projection system and voice prompts based on the updated topological road section status information. The system also has three traffic status response mechanisms: smooth state, slow-moving state, and congested state. In the smooth state, the system guides vehicles by projecting green light strips; in the slow-moving state, the system emits yellow warning lights and voice prompts; and in the congested state, the system reminds vehicles to change lanes or slow down by projecting red light strips and voice prompts.

[0048] like Figure 4 As shown in the figure, as the atmospheric visibility decreases (from greater than 500 meters to 40 meters), the ESI value shows a gradual upward trend, indicating that when visibility decreases, the degree of scattering of image quality increases. The solid line in the figure represents the actual measured value of ESI, while the dotted line represents the set ESI threshold line with a threshold of 1.8. When the ESI value exceeds this threshold, the system triggers the high-contrast working mode. The data in the figure clearly reflects how the system adjusts the working mode of its visual guidance system according to the real-time measured ESI value under low visibility conditions to cope with image quality issues under adverse environmental conditions.

[0049] like Figure 5 As shown in the figure, the first-layer main road (solid line) represents the actual road section, including Triumph-Main Road-Eastbound Lane 1, and the second-layer sections (dashed lines) represent the virtual roads connecting the topological sections, such as Triumph-West Road-NB and Triumph-SB-Lane 2. The third-layer sections (dotted lines) show a more complex topological structure, including Triumph-SB-Lane 1 and Triumph-SB-Lane 3, etc. Each section is connected by different types of nodes, such as circular nodes (first-layer nodes), square nodes (second-layer nodes) and triangular nodes (third-layer nodes). These nodes represent connection points at different levels, forming a complete traffic control network. In addition, the aerial unit perceives the traffic flow status through the ground unit and interacts with it, transmits information between nodes, and implements collaborative control through laser projection and voice prompts.

[0050] Example 4: Before the present invention is officially put into daily management and control tasks, it is necessary to execute a systematic offline calibration and rule generation program. The program aims to transform the key parameters and logical rules within the system from a state that relies on manual experience to an engineering procedure that is based on field measured data and topological network relationships and has certainty and traceability. The execution scenario of this procedure is the debugging stage after the system deployment is completed and before it is officially put into use. The core challenge it faces is how to ensure that the system's built-in judgment thresholds and response logic have solid physical basis and self-consistent macro-intelligence when facing the changing environment and traffic conditions in the real world, so as to avoid misjudgments and secondary risks caused by parameter inaccuracies or rough rules. Specifically, the program first starts from the constructed road network. The topological map begins with deepening the logical relationship processing. For each topological road segment identifier in the map, the system management terminal executes an upstream reachability tracing algorithm. The algorithm takes the selected identifier as the starting point, reversely traverses all other road segments that are directly physically connected to it and can provide it with traffic input, and records the identifiers of these road segments to form an upstream association set bound to the starting point identifier. Correspondingly, a downstream association set can also be generated through forward traversal. In this way, the data structure of each topological road segment identifier not only contains its own definition, but also deterministically stores all its direct logical adjacencies in the entire traffic network, providing a list of upstream monitoring targets for direct query when subsequent ground units judge the conductivity of traffic waves.

[0051] Furthermore, the procedure enters the parameter calibration stage for advanced perception functions. Taking the road physical state detection function of the aerial unit as an example, it aims to determine an optimal decision threshold for the polarized light brightness difference analysis algorithm to distinguish dry and wet roads. This calibration process is designed as a standardized field experiment. The operation is as follows: a representative test section with completed topological calibration is selected. Under typical dry and rainless meteorological conditions, the aerial unit flies directly above the section. The rotatable polarizer in front of its image sensor collects road surface reflected light images at two orthogonal angles of 0 degrees and 90 degrees in turn. The system obtains a baseline reading by calculating the average brightness difference between the two frames of images. This process is repeated at least twenty times to obtain the statistical average value, which is recorded as the dry state reference value. Subsequently, a high-pressure sprinkler truck that meets road test standards was used to evenly wet the road section, simulating a flooded road surface after rain. The aerial unit repeated the exact same imaging and calculation process to obtain a reference value for the slippery state. , given that these two reference values ​​represent two extreme states under the current hardware and environment combination, in order to ensure the detection rate while suppressing false alarms caused by factors such as uneven road surface materials, the final business decision threshold is set to a value between the two that can be calculated using the following formula:

[0052] ,

[0053] in, It is an adjustable sensitivity coefficient with a value range between 0 and 1. The system default setting is 0.5. This dynamic fixed-point calibration method within the difference range completely replaces the practice of setting a fixed universal threshold, ensuring the adaptability of the judgment logic to specific deployment environments.

[0054] Finally, the core of the program is to generate a set of refined coordinated response rules for traffic control facilities as ground units. Taking the traffic light controller at the ramp entrance as an example, its goal is to adaptively adjust the green light release time according to the congestion conditions of different topological sections of the downstream main road. To this end, the system does not adopt a single reduction ratio, but instead constructs a green light time adjustment matrix. The row index of the matrix is ​​the identifier of all topological sections that may send congestion status to the controller and are defined in its downstream association set. The column index is the two traffic flow states of slow movement and congestion. Each element in the matrix is ​​a specific green light time modulation coefficient. Its value is calculated based on traffic engineering theory and is associated with static topological parameters such as the design capacity of the corresponding main road section and the ramp merging angle. For example, when a three-lane main road section is declared congested, its corresponding modulation coefficient is It may be set to 0.4, while the congestion coefficient of a single lane ramp may be 0.7. When the system is running, after the controller receives the stateful topology information, it only needs to query this matrix with the road segment identifier and traffic status as the index to obtain the corresponding value, and based on The new green light plan is calculated and implemented based on the formula.

[0055] Example 5: In this embodiment, between the topological area recognition and the subsequent visualization projection operation, in order to ensure that the system can achieve high-precision control of specific topological road sections without relying on precise geographic coordinates, the aerial unit will further combine the relative position relationship and geometric proportion information between the topological road sections recorded in its internal preset topological map after completing the approximate area judgment based on the non-precise location information, and execute a spatial fine-tuning strategy based on image matching through the visual images collected in real time by its downward-viewing image sensor. This strategy uses ground facilities such as lane lines, traffic signs or intersection contours as references, and dynamically adjusts the visual projection matrix of its topological map in the local computing module through the similarity calculation between the image matching results and the topological feature images in the preset map, thereby realizing the transition from coarse-grained topological area judgment to precise topological area judgment at the logical level. The transformation of fine-grained topological road segment positioning, this positioning mechanism can provide a sufficiently stable and repeatable target projection benchmark for the laser projection system without relying on an external coordinate system; specifically, the spatial fine-tuning strategy of the aerial unit in the topological area, the core of its image matching is to compare the road structure elements in the real-time video stream, such as the edge contours of the lane lines, the geometric shapes of the guide arrows or the visual primitives such as the road surface characters, with the vectorized features of the corresponding road segments contained in the preset topological map or the high-resolution aerial feature map. By continuously calculating the geometric transformation relationship between the real-time field of view and the preset features, such as affine transformation or perspective transformation, the system can correct its own logical projection position in the topological map in real time, thereby achieving high-precision and high-robustness target indication that does not rely on the external coordinate system, providing a solid foundation for the accuracy of subsequent laser projection.

[0056] During the scheduling and execution of the projection system, the core light source of the laser module is selected based on a continuous laser with a wavelength in the green and red range of visible light. Laser sources with central wavelengths of 532 nanometers and 635 nanometers are configured respectively. The output power is adjustable and the pulse state is switched at a frequency of 5 Hz through the internal control circuit. When performing dual-wavelength compensation, the two laser sources will be turned on synchronously and projected onto the surface of the target road section in an alternating flashing manner. During the spatial projection process, the light spot is limited to an elliptical pattern with a diameter of approximately 25 cm by the collimation device inside the laser module. This size design is based on the reverse deduction of the projection scaling relationship between the standard road width and the working height of the aerial unit to ensure sufficient visual significance and image sensor recognition capabilities in typical urban road environments.

[0057] The calculation principle of the edge diffusion index aims to evaluate the clarity level of the current visual channel by quantifying the degree of blur of the actual imaging edge of the laser pattern on the ground. The calculation process takes the image data collected by the image sensor as input, extracts the pattern boundary area through the edge detection algorithm and measures its pixel width in the image coordinate system. The result is defined as the pixel width of the actual imaging edge, and the corresponding value is recorded as Wa; at the same time, according to the current height of the aerial unit and the lens viewing angle parameters, the theoretical pixel width that the pattern should present under ideal conditions can be calculated through geometric deduction, recorded as W0, then the edge diffusion index can be defined as the ratio of Wa to W0. The higher the ratio, the more serious the diffusion of the light spot boundary. When the ratio exceeds the set threshold, the system determines that the quality of the current visual channel has deteriorated and switches to high-contrast working mode.

[0058] In the process of traffic flow status identification, if the ground unit adopts image recognition, its core processing flow includes image acquisition, foreground target extraction, vehicle size estimation and center point extraction. The system calculates the geometric center distance between consecutive vehicles in the same frame and averages the obtained distance within a time sliding window to obtain the average vehicle distance parameter; if the coil sensor method is adopted, the coil sensing signal triggers the event, and the system records the number of vehicles passing through per unit time, which corresponds to two typical traffic flow parameters of time density and spatial density respectively. These parameters will be compared with the defined boundaries under the preset threshold system to determine whether the current traffic state is one of the three states: smooth, slow or congested. This determination process is completed entirely locally in the ground unit and does not rely on external servers or complex computing support, ensuring that the system has real-time response capabilities.

[0059] In terms of the execution mechanism of the rotatable polarizer, a micro-stepping motor is used inside the aerial unit to drive the polarizer to rotate and lock its direction at two preset orthogonal polarization angles, respectively collecting reflected light images of the target topological road section. The image processing module then performs a pixel-by-pixel brightness comparison on the two images, extracts their average brightness difference and compares it with the preset threshold to determine whether there is obvious polarization reflection phenomenon in the road section. The system's default polarization difference judgment threshold is based on the intrinsic difference in reflection characteristics between dry road surfaces and typical wet road surfaces. The specific value can be dynamically adjusted during the deployment phase in combination with local road material and meteorological characteristics to adapt to actual usage needs in different environments.

[0060] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A vehicle and personnel management method based on air-ground integrated data, characterized in that: The method comprises the following steps: Step a: constructing a road network topology map. In the topology map, each minimum directed road segment is assigned a unique topology road segment identifier, which is independent of the geographic coordinates. Step b: at least one ground unit senses the traffic flow state of the topological road section where it is located or monitored, where the traffic flow state includes one of unblocked, slow-moving and congested. The ground unit determines whether the traffic flow state meets the conditions for transitioning to a more congested state based on a set traffic flow characteristic threshold. If so, the ground unit monitors and receives the current traffic flow state of at least one upstream associated topological road section. Only when the current traffic flow state of at least one upstream associated topological road section meets the congestion condition or the slow-moving condition, the ground unit binds the determined traffic flow state to the corresponding topological road section identifier to form state-based topological information. Step c: the ground unit broadcasts stateful topology information via a low-power wireless network; In step d, at least one aerial unit receives stateful topological information. The aerial unit determines the current topological area based on its own imprecise location information and monitors the stateful topological information within the area. The aerial unit updates the visual mark of the road section in its preset topological map based on the received topological road section identifier and the corresponding traffic flow status. The aerial unit performs control operations related to the topological road section identifier based on the traffic status represented by the updated visual mark. The control operations include projecting visual diversion instructions or issuing voice prompts.

2. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: The traffic flow status determination in step b is based on the average vehicle spacing obtained by image recognition, or the number of vehicles passing per unit time collected by the coil sensor, and is determined by comparing the average spacing or the number of vehicles passing with the set threshold. The control operations performed by the aerial unit based on the traffic flow status include: when the traffic flow status is congested, projecting a red light band or issuing a congestion warning prompt; when the traffic flow status is unobstructed, projecting a green light band or issuing a pass instruction.

3. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: The method also includes: a traffic control facility as a ground unit, when it receives stateful topological information that its downstream or associated topological road section is declared to be in a congested state, automatically adjusts the green light time of the direction it controls according to the set adjustment rules to limit or guide traffic flow.

4. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: In step b, when the ground unit determines that the traffic flow state meets the conditions for transitioning to a higher congestion state, if the traffic flow state of its upstream associated topological section does not meet the congestion or slow-moving conditions, the ground unit does not broadcast a higher congestion state, but instead broadcasts an auxiliary information indicating that a micro-disturbance event exists in the topological section.

5. The vehicle and personnel management method based on air-ground integrated data according to claim 4 is characterized in that: After receiving auxiliary information indicating a micro-disturbance event, the aerial unit does not perform congestion relief and control operations, but instead performs lightweight warning actions. The lightweight warning action includes projecting a yellow pulse laser over the topological road section to alert the following vehicles to possible transient anomalies ahead.

6. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: The method also includes: the aerial unit projects a reference laser pattern onto the topological road section it monitors; capturing the actual image formed by the reference laser pattern on the road surface through the image sensor of the aerial unit; and calculating the edge diffusion index of the actual image. , The calculation formula is: ,in, is the pixel width of the actual imaging edge, It is the theoretical pixel width of the benchmark laser pattern under ideal conditions. When the edge diffusion index ESI exceeds the set threshold, the laser projection system of the aerial unit automatically switches to the high-contrast working mode, which includes dual-wavelength compensation and high-frequency pulse flashing mode.

7. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: The method also includes: setting a rotatable polarizer in front of the image sensor of the aerial unit; controlling the rotatable polarizer to collect reflected light images of the topological road section monitored by the aerial unit at at least two different polarization angles; determining the physical state of the road surface of the topological road section by comparing the average brightness difference between the images collected at different polarization angles; and associating the physical state of the road surface with the topological road section identifier, the physical state of the road surface including water accumulation or ice.

8. The vehicle and personnel management method based on air-ground integrated data according to claim 7 is characterized in that: The aerial unit adjusts the form of the visual evacuation instructions it projects based on the associated road physical state; when the road physical state is slippery, the projected green light band is modified to a dotted line or slow flashing form.

9. The vehicle and personnel management method based on air-ground integrated data according to claim 1 is characterized in that: The topological road segment identifier is in text form and includes the physical location information, direction information, and lane information of the road segment.

10. A vehicle and personnel management and control system based on air-ground integrated data, characterized in that: The system comprises: A map construction module is configured to construct a road network topology map, wherein each minimum directed road segment in the topology map is assigned a unique topology road segment identifier, and the identifier is independent of the geographic coordinates; At least one ground sensing and information broadcasting unit is configured to sense the traffic flow state of the topological road section where it is located or monitored, the traffic flow state including one of unblocked, slow-moving and congested; and based on a set traffic flow characteristic threshold, determine whether the traffic flow state meets the conditions for transitioning to a more congested state; if it is determined to be satisfied, the ground sensing and information broadcasting unit monitors and receives the current traffic flow state of at least one upstream associated topological road section; only when the current traffic flow state of at least one upstream associated topological road section meets the congestion condition or the slow-moving condition, the ground sensing and information broadcasting unit binds the determined traffic flow state with the corresponding topological road section identifier to form stateful topological information; and is configured to broadcast the stateful topological information through a low-power wireless network; At least one airborne receiving and control execution unit is configured to receive stateful topological information. The airborne receiving and control execution unit determines the current topological area based on its own non-precise location information and listens to the stateful topological information in the area; is configured to update the visual mark of the road section in its preset topological map according to the received topological road section identifier and the corresponding traffic flow status; and is configured to execute control operations related to the topological road section identifier based on the traffic status represented by the updated visual mark, and the control operations include projecting visual diversion instructions or issuing voice prompts.

Images