An air-ground cooperative millimeter wave sensing and decision system

Abstract

The application discloses an air-ground cooperative millimeter wave sensing and decision system, and belongs to the technical field of communication and sensing integration, comprising a positioning sensing module, which is used for acquiring position positioning data of a preset unmanned aerial vehicle in a flight process and transmitting the position positioning data to a ground communication base station; a millimeter wave sensing module, which is used for acquiring surrounding environment information according to the position positioning data in response to the position positioning data; the surrounding environment information comprises distance, speed and angle parameters; and the millimeter wave sensing module comprises a millimeter wave radar. The application integrates millimeter wave sensing, data analysis, image acquisition and processing and other technical means, improves the flight safety and sensing reliability of the unmanned aerial vehicle in a complex environment, improves the intelligent level and data processing efficiency of the system, and ensures that high-quality sensing information can be acquired in different flight states.

Description

Technical Field

[0001] This invention relates to the field of integrated communication and sensing technology, and in particular to an air-ground collaborative millimeter-wave intelligent sensing and decision-making system. Background Technology

[0002] With the continuous development of wireless communication and the huge demand for sensing capabilities, integrated sensing and communication with edge computing has become a promising technology for future wireless communication frameworks, enabling communication, sensing and computing functions to complement each other.

[0003] Millimeter-wave communication, due to its advantages such as high bandwidth and high capacity, is widely used in low-altitude communication. Its core requirements focus on data transmission (data transmission) and image transmission (image transmission) between UAVs and ground stations. Data transmission requires the network to ensure continuous, stable, highly reliable, and low-latency communication between the UAV and the ground station during low-altitude operations, so that the UAV can receive operational commands, while the ground station can track the UAV's flight status in real time and provide immediate feedback. Image transmission focuses on efficiently transmitting high-definition images and video information captured by the UAV's camera to the ground. This type of transmission is characterized by a large demand for uplink bandwidth. For example, the uplink rate requirement for 1080P resolution images is approximately 5Mbps, while 4K resolution requires as much as 25Mbps.

[0004] In the field of low-altitude sensing, on the one hand, it is necessary to monitor and serve cooperative drones to ensure orderly and controllable flight; on the other hand, it is necessary to detect and counter non-cooperative drones to ensure airspace safety.

[0005] Regarding this research, application document CN201611153189.3 provides an air-ground integrated unmanned intelligent decision-making method. This technical solution solves the problem that unmanned intelligent decision-making technology is difficult to implement in embedded computing environments due to the requirements of computing resources and real-time performance through an air-ground collaborative working mode. It provides a new approach for the realization of unmanned intelligent decision-making technology and can be applied to autonomous tasks of unmanned systems such as UAVs, unmanned vehicles, and unmanned ships, as well as auxiliary decision-making systems in manned systems.

[0006] Another application, CN202211138037.1, provides a multi-target cooperative tracking method. This technical solution realizes effective cooperative optimization of base station tracking resources in the base station sensing working mode, achieving a reasonable match between the base station and the tracking target, and greatly improving the overall tracking efficiency of the network.

[0007] However, the aforementioned technical solutions suffer from dynamic perception limitations, especially when UAVs are flying at high speeds, their perception and communication systems need to adapt rapidly to changes in the dynamic environment. For example, existing perception models struggle to cope with the blurring and misalignment caused by high-speed imaging and motion artifacts, resulting in high false alarm and false negative rates and reducing the reliability of perception. Summary of the Invention

[0008] In view of the problems existing in the field of communication and sensing integration technology, the present invention is proposed.

[0009] Therefore, one of the objectives of this invention is to provide an air-ground collaborative millimeter-wave sensing intelligent perception and decision-making system. By integrating millimeter-wave sensing, data analysis, image acquisition and processing technologies, it improves the flight safety and perception reliability of UAVs in complex environments, while also enhancing the system's intelligence level and data processing efficiency, ensuring that high-quality perception information can be obtained under different flight conditions.

[0010] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0011] This invention provides an air-ground cooperative millimeter-wave intelligent sensing and decision-making system, comprising:

[0012] The positioning and sensing module is used to acquire the position and positioning data of the preset UAV during flight and transmit the position and positioning data to the ground communication base station;

[0013] A millimeter-wave sensing module, which responds to the location data and is used to sense and acquire surrounding environmental information based on the location data; the surrounding environmental information includes distance, speed, and angle parameters; the millimeter-wave sensing module includes a millimeter-wave radar;

[0014] The data analysis module is used to analyze and process the distance and angle parameters in the surrounding environment information, extract feature information, and transmit the feature information to the ground communication base station.

[0015] The data fusion processing module responds to the extracted feature information and is used to analyze the changes in the distance and angle parameters; the data fusion processing module includes a target recognition unit, a calculation unit, an analysis and prediction unit, and a decision-making unit.

[0016] The target identification unit is used to obtain the buildings around the preset UAV according to the distance parameter, and to identify the type of the buildings;

[0017] The computing unit responds to the buildings around the drone and uses them to calculate the distance between the preset drone and different buildings;

[0018] The analysis and prediction unit responds to the calculated distance and uses it to predict the probability of a collision between a pre-defined drone and a building based on the change in the distance; and presets a risk threshold based on the probability of collision.

[0019] The decision-making unit responds to the risk threshold. When the predicted probability of the preset drone colliding with the building exceeds the risk threshold, the system determines that the probability of the preset drone colliding with the building is high and issues a warning; otherwise, it does not make a determination.

[0020] In a preferred embodiment of the present invention, the method of analyzing and processing the distance parameters and angle parameters in the data analysis module includes data preprocessing of the distance parameters and angle parameters, wherein the data preprocessing is filtering the distance parameters and angle parameters to remove noise and outliers;

[0021] It also includes data processing and alignment, used to synchronize the acquired distance and angle parameters in time and space. The synchronization method includes a preset reference frame, through timestamp alignment and coordinate transformation, to integrate the distance and angle parameters into the reference frame for analysis and processing.

[0022] In a preferred embodiment of the present invention, when the system determines that the probability of a preset drone colliding with a building is high, the acquired buildings are divided into... in, This indicates the nth building being acquired, and image information of each building is acquired. Simultaneously, the speed parameters of the preset UAV corresponding to the acquired image information are calculated. When the speed parameters increase, the changes in image information of each building are collected. If the image information does not change with the speed parameters of the preset UAV, the system determines that the image corresponding to the building in the image information is misaligned; otherwise, no determination is made. Furthermore, if the image is determined to be misaligned, the probability of a collision between the preset UAV and the building is re-predicted.

[0023] As a preferred embodiment of the present invention, the following is provided: when the speed parameter increases, the changes in image information of each building are collected. The collection method includes collecting images by shooting from a fixed perspective of a preset drone. When shooting from a fixed perspective, the images of the endpoints of each building closest to the preset drone are acquired. When the preset drone is flying forward, if the images of the endpoints of buildings in front of the preset drone change while the images of the endpoints of buildings behind the preset drone do not change, the determination that the images corresponding to the buildings in the image information are misaligned is maintained; otherwise, the original determination is not maintained.

[0024] The data collection method also includes collecting data by taking and stitching images from multiple angles using a pre-set drone. During the flight, the pre-set drone takes pictures of the building from multiple angles and generates a panoramic image through image stitching technology. The multiple angles include front view, rear view, and side view.

[0025] From the stitched panoramic image, obtain images of 4 to 6 endpoints closest to the preset drone, mark the images as reference images, and update the endpoint images for a preset time period as the drone flies forward. Based on the changes in the endpoint images during the update time period, determine whether the images corresponding to buildings in the image information are misaligned.

[0026] As a preferred embodiment of the present invention, the method includes: acquiring changes in the image of the endpoint closest to the preset drone in the panoramic image at a time interval of one second; in the reference image, dividing the images of 4 to 6 endpoints closest to the preset drone into a rear image, a middle image, and a front image; when the preset drone is flying forward, if the rear image disappears in the reference image, the system determines that the stitched panoramic image has not been misaligned, otherwise, it determines that misalignment has occurred.

[0027] In a preferred embodiment of the present invention, when the system determines that the stitched panoramic image is not misaligned, the system calculates the correlation between the changes in the back-end image and the forward flight of the preset UAV, and calculates the correlation using the following formula:

[0028]

[0029] Where v represents the speed of the preset drone in the horizontal direction, f represents the shooting frequency of the preset drone, and Δ t This indicates the time interval at which the backend image changes; a change in the backend image refers to the change in which the backend image disappears from the reference image.

[0030] According to the above formula, in the time interval Δ t Within this range, the horizontal displacement of the preset drone is:

[0031] △ x =v·△ t ;△ x This indicates the horizontal displacement of the pre-defined drone.

[0032] In a preferred embodiment of the present invention, the horizontal displacement of the back-end image in the panoramic image is calculated based on the horizontal displacement of the preset drone using geometric relationships; if the height of the building is h and the horizontal distance between the building and the preset drone is d, then the horizontal displacement Δ of the back-end image in the panoramic image is... y The formula for calculating the geometric relationship is as follows:

[0033]

[0034] In a preferred embodiment of the present invention: the time corresponding to the calculated horizontal displacement is obtained, the horizontal displacement within the time is marked as a reference displacement, and based on the preset speed of the UAV flying horizontally, if the horizontal displacement of the image of the endpoint corresponding to a certain building in the panoramic image generated from multi-angle shots of buildings in a future time period is faster than the reference displacement, the system determines that the panoramic image is misaligned and calculates the probability of a collision between the preset UAV and the building.

[0035] 1. The system uses a millimeter-wave sensing module to perceive surrounding environmental information (such as distance, speed, and angle parameters) in real time, and combines this with a data analysis module and a data fusion processing module to predict collision probabilities. When the collision probability exceeds a preset risk threshold, the system can issue a timely warning to help the drone operator or the autopilot system take evasive action and reduce the risk of collision. Furthermore, the system can dynamically adjust the collision probability prediction based on changes in the drone's speed parameters and building image information to ensure the accuracy and timeliness of the warning.

[0036] 2. The data analysis module preprocesses and aligns distance and angle parameters, and the data fusion processing module performs comprehensive analysis of feature information, improving the accuracy and reliability of the perceived data. When the speed of the UAV changes, the system can detect whether the building image is misaligned and re-predict the collision probability based on the misalignment, avoiding misjudgments caused by image errors.

[0037] 3. Through the collaborative work of the target recognition unit, calculation unit, analysis and prediction unit, and decision-making unit, the system can automatically identify building types, calculate distances, predict collision probabilities, and make decisions, thus achieving intelligent perception and decision-making.

[0038] 4. The system can dynamically adjust the image acquisition and processing strategy based on the speed parameters and real-time perception data of the UAV, ensuring that high-quality perception information can be obtained under different flight conditions;

[0039] 5. The system provides multiple image acquisition methods, such as fixed-viewpoint shooting and multi-angle shooting and stitching, which can select the appropriate acquisition strategy according to actual needs, improving the flexibility and efficiency of image acquisition. Furthermore, by setting update time periods to update and analyze key images in panoramic images in real time, it can quickly detect image changes and respond, improving the real-time performance and adaptability of the system. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0041] Figure 1 This is a schematic diagram of the modular structure of the air-ground collaborative millimeter-wave intelligent sensing and decision-making system according to an embodiment of the present invention;

[0042] Figure 2 This is a schematic diagram of the process structure of an embodiment of the present invention;

[0043] The numbers in the diagram are: 110 - Positioning and sensing module; 120 - Millimeter wave sensing module; 130 - Data analysis module; 140 - Data fusion processing module; 1401 - Target recognition unit; 1402 - Calculation unit; 1403 - Analysis and prediction unit; 1404 - Decision-making unit. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0045] Due to limitations in current dynamic perception technologies, especially when drones are flying at high speeds, their perception and communication systems need to adapt rapidly to changes in the dynamic environment. For example, existing perception models struggle to cope with the blurring and misalignment caused by high-speed imaging and motion artifacts, resulting in high false alarm and false negative rates and reducing the reliability of perception.

[0046] Based on this, the present invention proposes an air-ground collaborative millimeter-wave sensing intelligent perception and decision-making system, which improves the flight safety and perception reliability of UAVs in complex environments by integrating millimeter-wave sensing, data analysis, image acquisition and processing technologies, while enhancing the system's intelligence level and data processing efficiency, ensuring that high-quality perception information can be obtained under different flight conditions.

[0047] The present solution will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0048] Reference Figures 1 to 2 As one embodiment of the present invention, this embodiment provides an air-ground cooperative millimeter-wave intelligent sensing and decision-making system, comprising:

[0049] The positioning and sensing module 110 is used to acquire the preset position and positioning data of the UAV during flight and transmit the position and positioning data to the ground communication base station;

[0050] In this embodiment, the positioning sensing module 110 includes a GPS satellite navigation positioning device or a Beidou satellite navigation positioning device;

[0051] The millimeter-wave sensing module 120 responds to position positioning data and is used to sense and acquire surrounding environmental information based on the position positioning data; the surrounding environmental information includes distance, speed, and angle parameters; the millimeter-wave sensing module includes a millimeter-wave radar;

[0052] In this embodiment, it can be used for target detection, tracking, and environment modeling;

[0053] The data analysis module 130 is used to analyze and process the distance and angle parameters in the surrounding environment information, extract feature information, and transmit the feature information to the ground communication base station.

[0054] The data fusion processing module 140 responds to the extracted feature information and is used to analyze the changes in distance and angle parameters. The data fusion processing module 140 includes a target recognition unit 1401, a calculation unit 1402, an analysis and prediction unit 1403, and a decision-making unit 1404.

[0055] The target recognition unit 1401 is used to obtain the buildings around the preset UAV based on the distance parameters and to identify the type of the buildings;

[0056] The computing unit 1402 responds to the acquired buildings around the drone and uses them to calculate the distance between the drone and different buildings.

[0057] The analysis and prediction unit 1403 responds to the calculated distance to predict the probability of a collision between the drone and a building based on the change in distance; and presets a risk threshold based on the probability of collision.

[0058] The decision-making unit 1404 responds to the risk threshold. When the probability of the preset drone colliding with the building exceeds the risk threshold, the system determines that the probability of the preset drone colliding with the building is high and issues a warning; otherwise, it does not make a determination.

[0059] In this embodiment, through the coordinated work of the positioning perception module, millimeter-wave perception module, data analysis module and data fusion processing module, the system can comprehensively perceive the surrounding environment of the UAV and make intelligent decisions, achieving high-precision perception of the surrounding environment and providing reliable data support for the autonomous flight of the UAV.

[0060] In the data analysis module, the methods for analyzing and processing distance and angle parameters include data preprocessing, which involves filtering the distance and angle parameters to remove noise and outliers.

[0061] It also includes data processing and alignment, which is used to synchronize the acquired distance and angle parameters in time and space. The synchronization methods include a preset reference frame, and the distance and angle parameters are integrated into the reference frame for analysis and processing through timestamp alignment and coordinate transformation.

[0062] In this embodiment, noise and outliers are removed by filtering, which improves the accuracy and reliability of the data. By aligning timestamps and transforming coordinates, data from different sensors are fused into the same reference frame, which solves the problem of temporal and spatial inconsistency of multi-source data. This not only improves the accuracy of data analysis and provides a high-quality data foundation for subsequent collision prediction and decision-making, but also enhances the robustness of the system, enabling it to better adapt to complex and ever-changing flight environments.

[0063] Based on the above, when the system determines that the probability of a pre-defined drone colliding with a building is high, the acquired buildings are divided into... in, This represents the nth building that has been acquired, and image information of each building is acquired. At the same time, the speed parameters of the preset UAV corresponding to the acquired image information are calculated. When the speed parameters increase, the image information of each building is acquired. If the image information does not change with the speed parameters of the preset UAV, the system determines that the image corresponding to the building in the image information is misaligned. Otherwise, no judgment is made. Furthermore, if the image is determined to be misaligned, the probability of the preset UAV colliding with the building is re-predicted.

[0064] In this embodiment, the image acquisition strategy can be dynamically adjusted according to the speed parameters of the UAV to ensure that effective image information can be obtained at different flight speeds;

[0065] By analyzing changes in image information, image misalignment can be detected and collision probability can be re-predicted, thus avoiding misjudgments caused by image errors.

[0066] Furthermore, as the speed parameter increases, the image information of each building changes. The acquisition method includes capturing images from a fixed perspective of a preset drone. When capturing images from a fixed perspective, the image of the endpoint of each building closest to the preset drone is obtained. If, during the forward flight of the preset drone, the image of the endpoint of a building in front of the preset drone changes while the image of the endpoint of a building behind the preset drone does not change, the determination that the image corresponding to the building in the image information is misaligned is maintained; otherwise, the original determination is not maintained.

[0067] The data collection methods also include collecting data by using a pre-set drone to take pictures of buildings from multiple angles during flight and then using image stitching technology to generate panoramic images. The multiple angles include front view, rear view and side view.

[0068] From the stitched panoramic image, obtain images of 4 to 6 endpoints closest to the preset drone, mark the images as reference images, and update the endpoint images for a preset time period as the drone flies forward. Based on the changes in the endpoint images during the update time period, determine whether the images corresponding to buildings in the image information are misaligned.

[0069] In this embodiment, panoramic images are generated by taking pictures from multiple angles, which can provide more comprehensive building information and is suitable for buildings with complex structures.

[0070] It is important to emphasize in this embodiment that the system acquires the changes of the image of the endpoint closest to the preset drone in the panoramic image at a time interval of one second. In the reference image, the images of 4 to 6 endpoints closest to the preset drone are divided into rear image, middle image and front image. When the preset drone is flying forward, if the rear image disappears in the reference image, the system determines that the stitched panoramic image has not been misaligned; otherwise, it determines that misalignment has occurred.

[0071] In this embodiment, key images in the panoramic image are updated in real time with an update period of one second, which improves the real-time performance and response speed of the system. By distinguishing between back-end, mid-end, and front-end images, it is possible to more accurately determine whether image misalignment has occurred. This improves the system's sensitivity to image changes and enables it to quickly detect and respond to image misalignment.

[0072] When the system determines that the stitched panoramic image is not misaligned, it calculates the correlation between changes in the backend image and the preset forward flight of the drone, based on the following formula:

[0073]

[0074] Where v represents the preset horizontal speed of the drone, f represents the preset shooting frequency of the drone, and △ t This indicates the time interval at which the back-end image changes; a change in the back-end image is the change in which the back-end image disappears from the reference image.

[0075] According to the above formula, in the time interval Δ t Within, the preset horizontal displacement of the drone is:

[0076] △ x =v·△t ;△ x This indicates the horizontal displacement of the pre-defined drone;

[0077] In this embodiment, the calculation parameters are dynamically adjusted according to the speed and shooting frequency of the UAV to ensure the accuracy of the calculation results and enhance the dynamic adaptability of the system, enabling it to adjust the calculation strategy in real time according to the flight status.

[0078] Based on the above, the horizontal displacement of the backend image in the panoramic image is calculated using geometric relationships, according to the preset horizontal displacement of the drone. If the height of the building is h and the horizontal distance between the building and the preset drone is d, then the horizontal displacement Δ of the backend image in the panoramic image is... y The formula for calculating the geometric relationship is as follows:

[0079]

[0080] In this embodiment, the horizontal displacement of the back-end image in the panoramic image is calculated through geometric relationships, which further improves the accuracy of displacement calculation. By combining the height and horizontal distance of the building, the displacement is comprehensively analyzed, which improves the reliability of the calculation results and enhances the system's environmental perception capability, enabling it to better adapt to buildings of different heights and distances.

[0081] In this embodiment, it is necessary to explain that the time corresponding to the calculated horizontal displacement is obtained, and the horizontal displacement within the time is marked as the reference displacement. Based on the preset speed of the UAV flying horizontally, if the horizontal displacement of the image of the endpoint corresponding to a certain building in the panoramic image generated from multi-angle photos of buildings in the future time period is faster than the reference displacement, the system determines that there is a misalignment in the panoramic image and calculates the preset probability of the UAV colliding with the building.

[0082] In this embodiment, the collision probability is dynamically calculated based on the change in horizontal displacement, which improves the accuracy and timeliness of collision warning. Furthermore, combining the image misalignment detection results with the collision probability calculation further optimizes the collision warning mechanism, improves the accuracy and reliability of collision warning, and reduces the possibility of false alarms and missed alarms.

[0083] In summary, this application improves the flight safety and perception reliability of UAVs in complex environments by integrating millimeter-wave sensing, data analysis, image acquisition and processing technologies, while also enhancing the system's intelligence level and data processing efficiency, ensuring the acquisition of high-quality perception information under different flight conditions.

[0084] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A coordinating air-ground millimeter-wave intelligent sensing and decision-making system, characterized in that, include: The positioning and sensing module is used to acquire the position and positioning data of the preset UAV during flight and transmit the position and positioning data to the ground communication base station; A millimeter-wave sensing module, which responds to the location data and is used to sense and acquire surrounding environmental information based on the location data; the surrounding environmental information includes distance, speed, and angle parameters; the millimeter-wave sensing module includes a millimeter-wave radar; The data analysis module is used to analyze and process the distance and angle parameters in the surrounding environment information, extract feature information, and transmit the feature information to the ground communication base station. The data fusion processing module responds to the extracted feature information and is used to analyze the changes in the distance and angle parameters; the data fusion processing module includes a target recognition unit, a calculation unit, an analysis and prediction unit, and a decision-making unit. The target identification unit is used to obtain the buildings around the preset UAV according to the distance parameter, and to identify the type of the buildings; The computing unit responds to the buildings around the drone and uses them to calculate the distance between the preset drone and different buildings; The analysis and prediction unit responds to the calculated distance and uses it to predict the probability of a collision between a pre-defined drone and a building based on the change in the distance; and presets a risk threshold based on the probability of collision. The decision-making unit responds to the risk threshold. When the predicted probability of the preset drone colliding with the building exceeds the risk threshold, the system determines that the probability of the preset drone colliding with the building is high and issues a warning. Conversely, no judgment is made; When the system determines that the probability of a pre-defined drone colliding with a building is high, the acquired buildings are divided into... , ,..., ,in, Indicates the obtained number The system collects image information of each building and calculates the speed parameters of the corresponding drone. When the speed parameters increase, the system collects changes in the image information of each building. If the image information does not change with the speed parameters of the drone, the system determines that the image corresponding to the building is misaligned. Otherwise, no misalignment is made. If the image is determined to be misaligned, the system re-predicts the probability of the drone colliding with the building. When the speed parameter increases, the image information of each building changes. The acquisition method includes acquisition by shooting from a fixed perspective of a preset drone. When shooting from a fixed perspective, the image of the endpoint of each building closest to the preset drone is acquired. When the preset drone is flying forward, if the image of the endpoint of the building in front of the preset drone changes while the image of the endpoint of the building behind the preset drone does not change, the determination that the image corresponding to the building in the image information is misaligned is maintained; otherwise, the original determination is not maintained. The data collection method also includes collecting data by taking and stitching images from multiple angles using a pre-set drone. During the flight, the pre-set drone takes pictures of the building from multiple angles and generates a panoramic image through image stitching technology. The multiple angles include front view, rear view, and side view. From the stitched panoramic image, obtain images of 4 to 6 endpoints closest to the preset drone, mark the images as reference images, and update the endpoint images for a preset time period as the drone flies forward. Based on the changes in the endpoint images during the update time period, determine whether the images corresponding to buildings in the image information are misaligned.

2. The air-ground collaborative millimeter-wave intelligent sensing and decision-making system as described in claim 1, characterized in that, In the data analysis module, the method of analyzing and processing the distance parameters and angle parameters includes preprocessing the distance parameters and angle parameters, wherein the data preprocessing is filtering the distance parameters and angle parameters to remove noise and outliers; It also includes data processing and alignment, used to synchronize the acquired distance and angle parameters in time and space. The synchronization method includes a preset reference frame, through timestamp alignment and coordinate transformation, to integrate the distance and angle parameters into the reference frame for analysis and processing.

3. The air-ground collaborative millimeter-wave intelligent sensing and decision-making system as described in claim 1, characterized in that, This includes acquiring changes in the image of the endpoint closest to the preset drone in the panoramic image at a time interval of one second. In the reference image, the images of 4 to 6 endpoints closest to the preset drone are divided into rear image, middle image and front image. When the preset drone is flying forward, if the rear image disappears in the reference image, the system determines that the stitched panoramic image has not been misaligned; otherwise, it determines that misalignment has occurred.

4. The air-ground collaborative millimeter-wave intelligent sensing and decision-making system as described in claim 3, characterized in that, When the system determines that the stitched panoramic image is not misaligned, it calculates the correlation between the changes in the back-end image and the preset forward flight of the drone, based on the following formula: ； in, This indicates the speed at which the preset drone flies horizontally. This indicates the preset shooting frequency of the drone. This indicates the time interval at which the backend image changes; a change in the backend image refers to the change in which the backend image disappears from the reference image. According to the above formula, at the time interval Within this range, the horizontal displacement of the preset drone is: ; This indicates the horizontal displacement of the pre-defined drone.

5. The air-ground collaborative millimeter-wave intelligent sensing and decision-making system as described in claim 4, characterized in that, Based on the preset horizontal displacement of the drone, the horizontal displacement of the backend image in the panoramic image is calculated using geometric relationships; if the height of the building is... The horizontal distance between the building and the preset drone is The horizontal displacement of the backend image in the panoramic image. The formula for calculating the geometric relationship is as follows: 。 6. The air-ground collaborative millimeter-wave intelligent sensing and decision-making system as described in claim 5, characterized in that, The system obtains the time corresponding to the calculated horizontal displacement, marks the horizontal displacement within the time as the reference displacement, and, based on the preset speed of the UAV flying horizontally, if the horizontal displacement of the image of the endpoint corresponding to a certain building in the panoramic image generated from multi-angle shots of buildings in a future time period is faster than the reference displacement, the system determines that the panoramic image is misaligned and calculates the probability of a collision between the preset UAV and the building.

Images