Low-altitude object sensing method and device, low-altitude sensing system, readable storage medium and program product

Abstract

The invention relates to a low-altitude object sensing method and device, a low-altitude sensing system, a readable storage medium and a program product. The method comprises the following steps: in response to a perception demand request initiated for a low-altitude object perception task, determining a target perception device matched with the perception demand request in multi-source perception devices comprising at least two device types; sending a sensing data acquisition request to a target sensing device; the perception data acquisition request is used for indicating target perception equipment to perceive a low-altitude object to obtain multi-source perception data of at least two data types; receiving multi-source sensing data returned by the target sensing device, and performing data fusion processing on the multi-source sensing data to obtain a sensing data fusion result; outputting a perception result of the low-altitude object perception task according to the perception data fusion result; the sensing result is used for recording the sensing information of the low-altitude object. By adopting the method, the perception accuracy of the low-altitude object can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of integrated communication and perception, and in particular to a low-altitude object perception method, device, low-altitude perception system, computer-readable storage medium, and computer program product. Background Art

[0002] With the rise of the low-altitude economy, its application scenarios have gradually expanded to various vertical industries, such as express logistics, emergency services, geographic surveying and mapping, agricultural and forestry plant protection, urban planning, aerial photography, entertainment activities, etc.

[0003] However, the numerous mobile objects in the low-altitude airspace and the fixed infrastructure at high ground altitudes pose uncertainties and risks to each other. Therefore, low-altitude perception capabilities are needed to detect, identify and locate low-altitude objects, and provide communication services for some low-altitude objects. However, there are many low-altitude objects and the environment is complex. Low-altitude perception faces many difficulties and challenges, which makes the accuracy of low-altitude object perception in traditional technologies low. Summary of the invention

[0004] Based on this, it is necessary to provide a low-altitude object perception method, device, low-altitude perception system, readable storage medium and program product that can improve the accuracy of perception of low-altitude objects in response to the above-mentioned technical problems.

[0005] In a first aspect, the present application provides a low-altitude object perception method, comprising:

[0006] In response to a sensing requirement request initiated for a low-altitude object sensing task, determining a target sensing device matching the sensing requirement request from a multi-source sensing device including at least two device types;

[0007] Sending a perception data collection request to the target perception device; the perception data collection request is used to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device;

[0008] Receiving the multi-source perception data returned by the target perception device, and performing data fusion processing on the multi-source perception data to obtain a perception data fusion result;

[0009] According to the perception data fusion result, the perception result of the low-altitude object perception task is output; the perception result is used to record the perception information of the low-altitude object.

[0010] In one embodiment, determining a target sensing device matching the sensing demand request from among multiple source sensing devices including at least two types of devices includes:

[0011] Determining, according to the perception demand request, a perception accuracy requirement and a perception area requirement for the perception task of the low-altitude object;

[0012] Screening out the sensing device types that meet the sensing accuracy requirement and the sensing area requirement;

[0013] The environmental conditions of the area to be sensed are obtained, and the target sensing devices that meet the environmental conditions are screened out; the device type of the target sensing device is the sensing device type.

[0014] In one embodiment, the step of screening out the target sensing device that meets the environmental condition includes:

[0015] Acquire a target monitoring mode that matches the environmental conditions; the target monitoring mode is adapted to the light conditions, weather conditions, and sight distance conditions in the environmental conditions;

[0016] Among the multi-source sensing devices in the area to be sensed, a sensing device that adopts the target monitoring mode is selected as the target sensing device.

[0017] In one embodiment, the multi-source perception data is subjected to data fusion processing to obtain a perception data fusion result, including:

[0018] Obtaining a service request level of the perception demand request; the service request level is determined according to at least one of the perception accuracy requirement, real-time requirement, task urgency or environmental conditions of the low-altitude object perception task;

[0019] Determining a perception data fusion model that matches the service request level; the model performance of the perception data fusion model is positively correlated with the service request level;

[0020] The multi-source perception data is fused by the perception data fusion model to obtain the perception data fusion result.

[0021] In one embodiment, the multi-source perception data includes at least image data, radar signal data, and cellular signal data, and the multi-source perception data is fused by the perception data fusion model to obtain the perception data fusion result, including:

[0022] Inputting the image data into a pre-trained image feature extraction network to obtain image features in the multi-source perception data;

[0023] Obtaining radar signal features in the multi-source sensing data according to the radar signal data;

[0024] Acquire network signal features in the multi-source sensing data according to the cellular signal data;

[0025] The image features, radar signal features and network signal features are input into the perception data fusion model to obtain the perception data fusion result.

[0026] In one embodiment, outputting the perception result of the low-altitude object perception task according to the perception data fusion result includes:

[0027] Get a pre-trained perceptron neural network;

[0028] Inputting the perception data fusion result into the pre-trained perception neural network, and outputting at least one perception information of speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network;

[0029] At least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object is converted into a data packet to obtain a perception result of the low-altitude object perception task.

[0030] In a second aspect, the present application also provides a low-altitude object sensing device, the device comprising:

[0031] A response module, configured to respond to a perception demand request initiated for a low-altitude object perception task, and determine a target perception device that matches the perception demand request from a multi-source perception device including at least two device types;

[0032] A sending module, used for sending a sensing data collection request to the target sensing device; the sensing data collection request is used for instructing the target sensing device to sense the low-altitude object and obtain multi-source sensing data of at least two data types; the data type corresponds to the device type of the target sensing device;

[0033] A fusion module is used to receive the multi-source perception data returned by the target perception device, and perform data fusion processing on the multi-source perception data to obtain a perception data fusion result;

[0034] The output module is used to output the perception result of the low-altitude object perception task according to the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

[0035] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0036] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0037] In a fifth aspect, the present application also provides a computer program product, comprising a computer program, which implements the steps of the above method when executed by a processor.

[0038] The above-mentioned low-altitude object perception method, device, computer equipment, computer-readable storage medium and computer program product, by responding to a perception demand request initiated for a low-altitude object perception task, determine a target perception device that matches the perception demand request among multi-source perception devices including at least two device types, and send a perception data collection request to the target perception device so that the perception data collection request is used to instruct the target perception device to perceive the low-altitude object, thereby obtaining multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device; and then, by receiving the multi-source perception data returned by the target perception device, the multi-source perception data is processed. Perform data fusion processing to obtain perception data fusion results; output the perception results of the low-altitude object perception task based on the perception data fusion results; the perception results are used to record the perception information of the low-altitude objects; in this way, the advantages of multi-source perception equipment can be fully utilized to improve the overall perception and monitoring capabilities of the city's low altitude. On the premise of ensuring data security and privacy, the joint processing of multi-source perception data can be achieved to enhance the overall low-altitude perception capabilities. At the same time, it can also effectively enhance the perception capabilities in different scenarios, which is conducive to the implementation and promotion of low-altitude supervision programs, realize the manageability and controllability of low altitudes, and provide accurate perception services for low-altitude moving targets and stationary targets at higher altitudes on the ground. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 is an application environment diagram of a low-altitude perception method in an embodiment;

[0041] Figure 2 is a schematic diagram of a multi-source sensing device in one embodiment;

[0042] Figure 3 A schematic diagram of model training of a perception model in one embodiment;

[0043] Figure 4A schematic diagram of a low-altitude sensing method in an embodiment;

[0044] Figure 5 This is a diagram of an application scenario of a low-altitude perception method in an embodiment;

[0045] Figure 6 A schematic diagram of a device or algorithm selection module in one embodiment;

[0046] Figure 7 An application framework diagram of a low-altitude perception system in an embodiment;

[0047] Figure 8 is a timing diagram of a low-altitude sensing method in an embodiment;

[0048] Fig. 9 A schematic diagram of a selection strategy of a sensing device in one embodiment;

[0049] Fig.10 A schematic diagram of an output model of a monitoring method for a sensing object in one embodiment;

[0050] Fig.11 is a data processing flow chart of multi-source perception data in one embodiment;

[0051] Fig.12 A network diagram of a decision-level neural network in one embodiment;

[0052] Fig.13 is a flow chart of a low-altitude sensing method in another embodiment;

[0053] Fig.14 is a structural block diagram of a low-altitude sensing device in an embodiment;

[0054] Fig.15 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] The low-altitude sensing method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, a low-altitude sensing system (LASS) 102 communicates with a plurality of multi-source sensing devices 104 and service and application providers 106 through a network. In actual applications, the service and application provider 106 can send a perception demand request to the low-altitude perception system 102 for a low-altitude object perception task; the low-altitude perception system 102 receives the perception demand request; the low-altitude perception system 102 responds to the perception demand request initiated for the low-altitude object perception task, and determines a target perception device that matches the perception demand request from the multi-source perception devices 104 including at least two device types; the low-altitude perception system 102 sends a perception data collection request to the target perception device; the perception data collection request is used to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device; the low-altitude perception system 102 receives the multi-source perception data sent back by the target perception device, and performs data fusion processing on the multi-source perception data to obtain a perception data fusion result; the low-altitude perception system 102 outputs the perception result of the low-altitude object perception task based on the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

[0057] The sensing device may refer to a device used to sense various sensing information such as the position, speed, shape, size, etc. of a low-altitude target. Figure 2 As shown, the multi-source sensing device 104 may refer to the public IoT infrastructure widely deployed in the city, including existing sensing devices in urban rail transit, residential areas, commercial areas, office areas, industrial parks, cultural and entertainment areas, educational institutions, medical facilities, parks and other areas. These sensing devices have target perception capabilities in multiple dimensions such as image, video, radar (lidar, millimeter wave radar, ultrasonic radar), optics, infrared, radio frequency, optoelectronics, and electromagnetic.

[0058] In addition, the multi-source sensing device 104 may also include sensing devices specifically deployed for low-altitude monitoring, which can provide more comprehensive and accurate sensing capabilities for specific areas.

[0059] In order to achieve the continuity of low-altitude monitoring and increase the ability to monitor random and sudden targets, it is necessary to provide monitoring capabilities without blind spots in the monitoring target area. In addition to using public infrastructure for urban monitoring and low-altitude dedicated sensing equipment, personal users' private sensing devices will also be used at the same time, such as: mobile phones and other handheld terminal devices carried by residents, vehicle driving recorders, vehicle-mounted radars and other sensing devices are also used as data collection terminals for assistance. In summary, fixed and mobile sensing monitoring equipment (i.e., multi-source sensing equipment) are combined to form a low-altitude monitoring infrastructure covering a specific airspace.

[0060] In practical applications, the multi-source perception device 104 can run a pre-trained perception model; taking the multi-source perception device 104 as an image acquisition device, for example, the image acquisition device can be pre-configured with an image recognition algorithm; the image acquisition device can use the image recognition algorithm to perceive low-altitude objects. In practical applications, a federated learning approach can be used to pre-train the perception models used by the multi-source perception devices 104 on different perception device suppliers and data processing platforms, to achieve joint perception data processing between different perception device suppliers and data processing platforms, thereby avoiding direct data sharing and improving overall perception capabilities. Figure 3 As shown in the figure, the data collected by various sensing devices in the low-altitude monitoring area are stored in the local device or local server, and the original data is not uploaded centrally. Each sensing device or its local processing platform is responsible for the preprocessing and training of local data, and uploads the trained model update to the cloud for aggregation. The federated learning server deployed by the city management agency or cloud platform operator, as a coordinator, is responsible for collecting and updating the perception model from each sensing device, performing global aggregation of the perception model, and obtaining a global perception model. The global perception model is returned to each sensing device or its local processing platform. The data processing platform deployed in the cloud is used to process the global perception model, and further intelligently analyze the perception data provided by multiple devices, and integrate the perception data and perception results obtained by multi-dimensional and multi-angle monitoring to achieve high-precision and high-resolution detection, recognition and tracking of low-altitude targets. Through the application of federated learning, low-altitude multi-source perception data can be effectively integrated, joint perception data processing in a distributed environment can be realized, and low-altitude monitoring and perception capabilities can be improved. In view of the low latency requirements of some low-altitude application scenarios, a historical model library can be established in the cloud, and the global perception model can be updated through transfer learning, which can effectively save the training time of the global perception model.

[0061] In actual applications, for sensing devices in public infrastructure, dedicated low-altitude sensing devices, and personal private sensing devices located in different sensing areas and with different managers, distributed processing is often required. The specific workflow is as follows:

[0062] 1. Local data preprocessing: After the sensing device collects data, the data is first preprocessed locally on the sensing device to filter out noise and enhance effective information.

[0063] 2. Local model training: The perception device uses local data for model training. For example, through deep learning models or machine learning models, local training is performed on tasks such as target detection, classification, and trajectory tracking. During the training process, the device data of the perception device will not be leaked to the outside, protecting data privacy.

[0064] 3. Model update upload: Model updates (gradients or weights) after local training are uploaded to the federated learning server through a secure communication protocol. The collected data of the sensing device itself will not be uploaded.

[0065] 4. Global model aggregation: The federated learning server aggregates the model updates uploaded by multiple perception devices to generate a global perception model.

[0066] 5. Global perception model distribution: The global perception model will be distributed to each perception device. The device can use the updated global perception model to perform a new round of data processing and intelligent analysis to improve perception accuracy.

[0067] 6. Multi-source perception data processing and fusion: Through federated learning, the local perception model of the perception device can continuously integrate the data patterns from other devices to improve the comprehensiveness of perception. At the same time, the city's management platform can obtain the overall situational awareness of the low-altitude area without directly accessing the original data of each device.

[0068] Among them, the low-altitude perception system 102 and the service and application provider 106 can be independent physical servers, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0069] In an exemplary embodiment, Figure 4 As shown, a low-altitude sensing method is provided, which is applied to Figure 1 The low altitude sensing system 102 in FIG. 1 is taken as an example to illustrate, and the steps include the following steps S402 to S406. Among them:

[0070] Step S402, in response to a perception demand request initiated for a low-altitude object perception task, a target perception device matching the perception demand request is determined from among multi-source perception devices including at least two device types.

[0071] Among them, low-altitude objects can refer to stationary targets or moving targets in the low-altitude airspace. Specifically, low-altitude objects can also refer to low-altitude targets. Figure 5As shown in the figure, low-altitude objects can refer to a variety of low-altitude moving targets in the low-altitude airspace, such as "low, slow, and small" aircraft such as paragliders, drones, powered hot air balloons, various birds, and unpowered entertainment equipment such as kites and sky lanterns. It can be seen that these low-altitude moving targets often pose potential risks to urban operations and services. For example, in key areas such as natural gas or gas stations and places where flammable and explosive items are produced and stored, low-altitude targets can easily cause fires or equipment failures. For another example, in key infrastructure areas such as cultural heritage sites and observatories, low-altitude objects may cause damage or interference. In addition, urban buildings are getting taller and taller, and infrastructure such as high-voltage wires, communication towers, TV and radio antennas, and wind turbines can be seen everywhere. Although some moving targets have autonomous obstacle avoidance capabilities, they still face the risk of collision, especially in low visibility and bad weather conditions. Therefore, it is necessary to effectively perceive and monitor low-altitude objects.

[0072] The low-altitude object perception task may refer to a task of perceiving a certain low-altitude object in a certain low-altitude airspace, for example, a task of perceiving a drone flying in a certain area.

[0073] In a specific implementation, after the service and application provider obtains the low-altitude object perception task, the service and application provider can send a perception demand request for the low-altitude object perception task to the low-altitude perception system. After the low-altitude perception system receives the perception demand request for the low-altitude object perception task, the low-altitude perception system can respond to the perception demand request initiated for the low-altitude object perception task and determine a target perception device that matches the perception demand request from a multi-source perception device including at least two device types.

[0074] Specifically, the low-altitude perception system can determine the task requirements of the low-altitude object perception task based on the perception demand request; then, the low-altitude perception system can determine the target perception device that matches the task requirements of the low-altitude object perception task from the multi-source perception devices according to the task requirements of the low-altitude object perception task. In practical applications, the low-altitude perception system can select the target perception device that is suitable for the current perception scenario from the multi-source perception devices in the area to be perceived according to the service request level of the perception demand request. Of course, please refer to Figure 6 The low-altitude perception system can use the pre-trained device selection model to calculate the usage weights and data reporting cycles corresponding to different perception methods in the current perception scenario, using the perception accuracy requirements, perception area, weather conditions, real-time requirements and other information of the low-altitude object perception task as input parameters. The low-altitude perception system can select the target perception device suitable for the current perception scenario according to the usage weights and data reporting cycles corresponding to different perception methods.

[0075] In practical applications, the establishment of a device selection model requires the analysis of the characteristics of available multi-source perception devices, including but not limited to vision, radar, sensors, 4G / 5G cellular signals, Wi-Fi, Bluetooth, etc. Each perception method has its own unique perception range, accuracy, cost, energy consumption, and environmental adaptability. The following provides some exemplary characteristic information of perception methods:

[0076] Regarding image perception: it can obtain rich perception target information and the image processing algorithm is mature;

[0077] For radar sensing: Provides high distance resolution and speed measurement capabilities, suitable for long-distance target detection;

[0078] Regarding sensor perception: high perception accuracy, little affected by weather and light;

[0079] For 4G / 5G perception methods: provide communication capabilities while performing perception to achieve resource reuse;

[0080] Regarding Wi-Fi sensing methods: it has a certain penetration capability and is less affected by obstacles within the sensing range;

[0081] For Bluetooth sensing mode: suitable for indoor positioning and close-range target tracking.

[0082] Step S404: Send a perception data collection request to the target perception device.

[0083] Among them, the perception data collection request is used to instruct the target perception device to perceive the low-altitude objects and obtain multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device.

[0084] In a specific implementation, after the low-altitude perception system determines the target perception device, the low-altitude perception system can send a perception data collection request to the target perception device to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data of at least two data types.

[0085] Specifically, assuming that the target perception device includes an image acquisition device and a radar device in area A; the low-altitude perception system can send perception data acquisition requests to the image acquisition device and the radar device in area A respectively, to instruct the image acquisition device and the radar device in area A to perceive the low-altitude object, and obtain the image perception data and radar perception data of the low-altitude object as the above-mentioned multi-source perception data.

[0086] Step S406, receiving the multi-source perception data returned by the target perception device, and performing data fusion processing on the multi-source perception data to obtain a perception data fusion result.

[0087] In a specific implementation, the low-altitude perception system can receive multi-source perception data sent back by the target perception device. In practical applications, the low-altitude perception system can determine the multi-source perception data of the target perception device according to the real-time requirements of the low-altitude object perception task.

[0088] In practical applications, the low-altitude perception system can instruct the target perception device to report multi-source perception data periodically through a perception data collection request, and the low-altitude perception system can receive multi-source perception data sent back by the target perception device periodically. The low-altitude perception system can also instruct the target perception device to report multi-source perception data in real time through a perception data collection request, and the low-altitude perception system can receive multi-source perception data sent back by the target perception device in real time.

[0089] After acquiring multi-source perception data, the low-altitude perception system can use a pre-trained data fusion algorithm to perform data fusion processing on the multi-source perception data to obtain a perception data fusion result.

[0090] In practical applications, the low-altitude perception system can obtain a target data fusion algorithm that matches the service request level of the perception demand request from multiple candidate data fusion algorithms according to the service request level of the perception demand request; use the target data fusion algorithm to perform data fusion processing on multi-source perception data to obtain a perception data fusion result.

[0091] Of course, please see Figure 6 The low-altitude perception system can select a model through a pre-trained algorithm, and use the perception accuracy requirements, perception area, weather conditions, real-time requirements and other information of the low-altitude object perception task as input parameters to calculate the usage weights corresponding to different candidate data fusion algorithms. The low-altitude perception system can select a target data fusion algorithm suitable for the current perception scene according to the usage weights corresponding to each candidate data fusion algorithm; use the target data fusion algorithm to perform data fusion processing on multi-source perception data to obtain perception data fusion results. In practical applications, candidate data fusion algorithms can include at least a lightweight fusion algorithm, a high-precision fusion algorithm and a compressed fusion algorithm.

[0092] Step S408: output the perception result of the low-altitude object perception task according to the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

[0093] In a specific implementation, after obtaining the perception data fusion result, the low-altitude perception system can output the perception result of the low-altitude object perception task based on the perception data fusion result; the perception result is used to record the perception information of the low-altitude object. Specifically, the low-altitude perception system can input the perception data fusion result into a pre-trained perception neural network, and output at least one perception information of the speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network.

[0094] In order to facilitate the understanding of those skilled in the art, Figure 7 An exemplary application framework diagram of a low-altitude perception system is provided; Figure 7 As shown, the service and application providers are connected to the perception and infrastructure providers through network communication; the perception and infrastructure providers are used to control the low-altitude perception system and multi-source perception equipment. The low-altitude perception system can have the ability to dispatch and manage perception equipment, the ability to collect perception equipment data, the ability to process perception data, the ability to integrate multi-modal and collaborative perception, the ability to distribute perception, and the ability to provide low-altitude perception services; the multi-source perception equipment can include public Internet of Things perception equipment, dedicated low-altitude perception equipment, and private perception equipment; among them, the public Internet of Things perception equipment and dedicated low-altitude perception equipment can report their location, airspace coverage, perception performance, perception mode, etc. to the low-altitude perception system on a regular basis, on demand, or in real time according to different application scenarios; the low-altitude perception system provides interfaces with various platforms for information and data transmission and sharing. Auxiliary information such as city maps, weather conditions, and city asset information will help the low-altitude perception system better provide low-altitude perception services for service and application providers.

[0095] In order to facilitate the understanding of those skilled in the art, Figure 8 A timing diagram of a low-altitude sensing method is exemplarily provided; wherein the workflow of the low-altitude sensing method is:

[0096] 1: The service and application provider sends a perception requirement request for the low-altitude object perception task to the low-altitude perception system, and the low-altitude perception system receives the perception requirement request;

[0097] 2: The low-altitude perception system performs the perception device selection operation; the low-altitude perception system can select, allocate and dispatch one or more suitable perception devices, i.e., target perception devices, according to the low-altitude perception requirements, the perception task request location, environmental conditions and the perception capabilities of the perception devices in the area;

[0098] 3: The low-altitude perception system sends a perception data collection request to the target perception device, instructing the target perception device to perceive the low-altitude object and collect low-altitude data, which may include video, image, angle, speed, distance and other data;

[0099] 4: The low-altitude perception system receives multi-source perception data reported by the target perception device for processing, including data synchronization, data cleaning, data compression and feature extraction, and fuses data from different positions, angles and modes. For example, the low-altitude perception system can collect perception results from different angles related to the same low-altitude object for joint identification and positioning; for data from different locations or manufacturers, distributed processing can be performed, and the distributed results can be merged for collaborative perception to output the perception results of the low-altitude object perception task.

[0100] 5: The low-altitude perception system can provide feedback based on the perception results and use the perception results to provide low-altitude perception services (detection, identification, and positioning) capabilities to service and application service providers.

[0101] In the above-mentioned low-altitude perception method, by responding to a perception demand request initiated for a low-altitude object perception task, a target perception device matching the perception demand request is determined in a multi-source perception device including at least two device types, and a perception data collection request is sent to the target perception device so that the perception data collection request is used to instruct the target perception device to perceive the low-altitude object, thereby obtaining multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device; then, by receiving the multi-source perception data returned by the target perception device, the multi-source perception data is subjected to data fusion processing to obtain a perception data fusion result; according to the perception data fusion result, the perception result of the low-altitude object perception task is output; the perception result is used to record the perception information of the low-altitude object; in this way, the advantages of the multi-source perception device can be fully utilized, the overall perception and monitoring capabilities of the urban low-altitude can be improved, and the joint processing of the multi-source perception data can be realized on the premise of ensuring data security and privacy, and the overall low-altitude perception capability can be improved. At the same time, the perception capability in different scenarios can be effectively improved, which is conducive to the implementation and promotion of the low-altitude supervision plan, and the low-altitude can be managed and controlled, and accurate perception services can be provided for low-altitude moving targets and ground stationary targets with higher altitudes.

[0102] In an exemplary embodiment, among multi-source perception devices including at least two device types, a target perception device that matches the perception demand request is determined, including: determining the perception accuracy requirement and the perception area requirement of the perception task of the low-altitude object according to the perception demand request; screening out the perception device type that meets the perception accuracy requirement and the perception area requirement; obtaining the environmental conditions of the area to be perceived, and screening out the target perception device that meets the environmental conditions; the device type of the target perception device is the perception device type.

[0103] In a specific implementation, in the process of determining the target sensing device that matches the sensing demand request from the multi-source sensing devices including at least two types of devices, the low-altitude sensing system can determine the sensing accuracy requirements and sensing area requirements for the sensing task of the low-altitude object according to the sensing demand request; then, the low-altitude sensing system can filter out the sensing device types that meet the sensing accuracy requirements and sensing area requirements. The low-altitude sensing system can obtain the environmental conditions of the area to be sensed, and filter out the target sensing devices that meet the environmental conditions and whose device type is the sensing device type.

[0104] In practical applications, see Fig. 9 , the low-altitude perception system can select perception devices with obvious advantages or give higher weights to their perception results according to different environmental conditions. For example, the low-altitude perception system can preliminarily screen the type of perception devices suitable for the area to be perceived based on the perception accuracy requirements and perception area requirements. Then, according to the light conditions, weather conditions, line of sight conditions, etc. of the area to be perceived, select the appropriate perception device. In good light conditions, you can choose to monitor by image, video, etc., and in poor light conditions, you can choose to monitor by infrared, cellular, etc.

[0105] In practical applications, low-altitude perception systems can also establish deep learning models to intelligently and flexibly schedule and select perception devices. Fig.10 The low-altitude perception system can input information about factors such as the area to be perceived, performance requirements, weather conditions, light conditions, and line-of-sight conditions into the pre-trained neural network model, and output the type of perception device suitable for the area to be perceived through the pre-trained neural network model. Different types of perception devices correspond to different monitoring methods, such as sensors, vision, cellular, optoelectronics, radar and other monitoring methods.

[0106] The technical solution of this embodiment determines the perception accuracy requirements and perception area requirements for the perception task of low-altitude objects according to the perception demand request, and screens out the types of perception devices that meet the perception accuracy requirements and perception area requirements; obtains the environmental conditions of the area to be perceived, and screens out the target perception devices that meet the environmental conditions, so that the perception capability of the target perception device can effectively meet the task requirements of the low-altitude object perception task.

[0107] In an exemplary embodiment, screening out target sensing devices that meet environmental conditions includes: acquiring a target monitoring method that matches the environmental conditions; and selecting a sensing device that adopts the target monitoring method from among multiple source sensing devices in the area to be sensed as the target sensing device.

[0108] Among them, the target monitoring method is adapted to the light conditions, weather conditions, and line of sight conditions in the environmental conditions.

[0109] For specific implementation, please refer to Fig. 9 ,After the low altitude perception system has screened out the types of perception ,devices that meet the perception accuracy requirements and the perception area requirements, and in the ,process of screening out the target perception devices that meet the environmental conditions, the ,low altitude perception system can obtain the target monitoring method that matches the environmental conditions.

[0110] like Fig. 9 As shown, the low-altitude perception system can judge whether the light in the area to be perceived is good according to the environmental conditions; if the light in the area to be perceived is good, the image monitoring method and the video monitoring method are selected as the target monitoring method; if the light in the area to be perceived is not good, the infrared monitoring method and the cellular monitoring method are selected as the target monitoring method.

[0111] The low-altitude perception system can judge whether the weather in the area to be perceived is good based on environmental conditions; if the weather in the area to be perceived is good, image monitoring and video monitoring are selected as target monitoring methods; if the weather in the area to be perceived is not good, photoelectric monitoring and cellular monitoring are selected as target monitoring methods.

[0112] The low-altitude perception system can determine whether the line-of-sight path of the area to be perceived is good based on environmental conditions; if the line-of-sight path of the area to be perceived is good, the radio frequency monitoring method and the image monitoring method are selected as the target monitoring method; if the line-of-sight path of the area to be perceived is not good, the sensor monitoring method is selected as the target monitoring method.

[0113] The technical solution of this embodiment, by acquiring a target monitoring method that matches the environmental conditions and selecting a perception device that adopts the target monitoring method from the multi-source perception devices in the area to be perceived, as the target perception device, can realize adaptive selection of perception devices of appropriate device types according to the environmental conditions of the area to be perceived, thereby effectively improving the perception efficiency of low-altitude objects.

[0114] In an exemplary embodiment, multi-source perception data is subjected to data fusion processing to obtain a perception data fusion result, including: obtaining a service request level of a perception demand request; determining a perception data fusion model that matches the service request level; the model performance of the perception data fusion model is positively correlated with the service request level; and the multi-source perception data is subjected to fusion processing through the perception data fusion model to obtain a perception data fusion result.

[0115] The service request level is determined based on at least one of the perception accuracy requirements, real-time requirements, mission urgency or environmental conditions of the low-altitude object perception mission.

[0116] In practical applications, different service request levels of sensing demand requests can be predefined. Specifically, the sensing demand requests can be divided into different service request levels according to at least one of the sensing accuracy requirements, real-time requirements, task urgency or environmental conditions of the low-altitude object sensing task.

[0117] In practical applications, different perception accuracy requirements for low-altitude object perception tasks correspond to different data fusion strategies and the selection of high-precision model algorithms; different real-time requirements will affect the way data is used, that is, whether to use historical data reported periodically by the target perception device, or to initiate data requests to the target perception device in real time; different task urgency will affect the way data is used, data fusion strategies, and the selection of high-precision model algorithms.

[0118] In practical applications, the service request levels of perception demand requests can be divided into three levels: high, medium and low. The higher the service request level, the higher the model performance requirement of the perception data fusion model, that is, the model performance of the perception data fusion model is positively correlated with the service request level.

[0119] For high-level perception demand requests, which correspond to urgent tasks and require high precision and real-time performance, priority is given to perception methods such as radar and 5G, integrating modalities such as video and infrared, and selecting high-precision multimodal fusion models (e.g., high-precision perception data fusion models).

[0120] For mid-level perception demand requests, which correspond to routine tasks and require a balance between accuracy and cost, sensors, videos, optoelectronics and other modules can be selected. Different traditional algorithm combinations or lightweight multimodal neural network modules (such as lightweight perception data fusion models) can be used.

[0121] For low-level perception demand requests, the mid-level perception demand requests correspond to monitoring tasks, which require low cost. In order to solve the data storage cost, a single module and a traditional perception algorithm based on data compression (such as a perception data fusion model after data compression) can be selected.

[0122] In the specific implementation, when the low-altitude perception system performs data fusion processing on multi-source perception data and obtains the perception data fusion result, the low-altitude perception system can obtain the service request level of the perception demand request and determine the perception data fusion model that matches the service request level. Specifically, when the low-altitude perception system determines that the perception demand request is a low-level perception demand request, the low-altitude perception system can perform fusion processing on the multi-source perception data through the perception data fusion model after data compression to obtain the perception data fusion result; when the low-altitude perception system determines that the perception demand request is a medium-level perception demand request, the low-altitude perception system can perform fusion processing on the multi-source perception data through the lightweight perception data fusion model to obtain the perception data fusion result; when the low-altitude perception system determines that the perception demand request is a high-level perception demand request, the low-altitude perception system can perform fusion processing on the multi-source perception data through the high-precision perception data fusion model to obtain the perception data fusion result.

[0123] The technical solution of this embodiment obtains the service request level of the perception demand request and determines the perception data fusion model that matches the service request level. The multi-source perception data is fused through the perception data fusion model to obtain the perception data fusion result. This can adaptively select the appropriate perception data fusion model based on the importance of the low-altitude object perception task, thereby effectively saving computing resources and reducing operating costs.

[0124] In an exemplary embodiment, multi-source perception data are fused through a perception data fusion model to obtain a perception data fusion result, including: inputting image data into a pre-trained image feature extraction network to obtain image features in the multi-source perception data; identifying the distance and speed of low-altitude objects based on radar signal data to obtain radar signal features in the multi-source perception data; obtaining network signal features in the multi-source perception data based on cellular signal data; and inputting image features, radar signal features, and network signal features into the perception data fusion model to obtain a perception data fusion result.

[0125] Among them, the multi-source perception data includes at least image data, radar signal data and cellular signal data.

[0126] Among them, cellular signal data may refer to 4G / 5G signal data.

[0127] In the specific implementation, Fig.11 As shown, in the process of fusing multi-source perception data through a perception data fusion model to obtain a perception data fusion result, the low-altitude perception system can input image data into a pre-trained image feature extraction network (e.g., a visual feature extraction network, ResNet, Residual Neural Network) to obtain image features in the multi-source perception data.

[0128] The low-altitude perception system can identify the distance and speed of low-altitude objects based on radar signal data, and obtain radar signal features in multi-source perception data; the low-altitude perception system can input radar signal data into a signal feature extraction module (3D-FFT) to measure the distance, speed, and angle of the radar signal data, and identify the distance and speed of low-altitude objects; then, the result output by the signal feature extraction module is input into a radar feature processing network to obtain radar signal features in multi-source perception data.

[0129] The low-altitude perception system can also obtain network signal features in multi-source perception data based on cellular signal data; the low-altitude perception system can input cellular signal data into a signal feature extraction module (3D-FFT) to measure the distance, speed, and angle of the cellular signal data to obtain processed cellular signal data; then, the processed cellular signal data is input into a network signal feature processing network to obtain network signal features in multi-source perception data.

[0130] The low-altitude perception system can input image features, radar signal features, and network signal features into the perception data fusion model, such as Fig.11 The feature fusion network in the image is used to obtain the perception data fusion result. The low-altitude perception system can output the distance, speed, angle, number and type of low-altitude objects based on the perception data fusion result.

[0131] The technical solution of this embodiment obtains image features in multi-source perception data by inputting image data into a pre-trained image feature extraction network, and identifies the distance and speed of low-altitude objects based on radar signal data to obtain radar signal features in multi-source perception data; then obtains network signal features in multi-source perception data based on cellular signal data; inputs image features, radar signal features and network signal features into a perception data fusion model to obtain perception data fusion results, thereby realizing reasonable and effective fusion of multi-source perception data collected by multi-source perception devices, and accurately outputting perception information such as distance, speed, angle, quantity and type of low-altitude objects based on subsequent fusion results of the multi-source perception data.

[0132] In an exemplary embodiment, the perception result of the low-altitude object perception task is outputted based on the perception data fusion result, including: obtaining a pre-trained perception neural network; inputting the perception data fusion result into the pre-trained perception neural network, and outputting at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network; converting at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object into a data packet to obtain the perception result of the low-altitude object perception task.

[0133] In a specific implementation, when the low-altitude perception system outputs the perception result of the low-altitude object perception task according to the perception data fusion result, the low-altitude perception system can obtain a pre-trained perception neural network. The low-altitude perception system can input the perception data fusion result into the pre-trained perception neural network, and output at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network; the low-altitude perception system can convert at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object into a data packet to obtain the perception result of the low-altitude object perception task.

[0134] In practical applications, such as Fig.12 As shown, the decision-level neural network can include the above-mentioned perception data fusion model and perception neural network; in practical applications, the low-altitude perception system can input image data into a pre-trained image feature extraction network to obtain image feature information in multi-source perception data. The low-altitude perception system can input radar signal data and cellular signal data into a 3D-FFT processing module to output speed feature information, distance feature information, and angle feature information of low-altitude objects; the low-altitude perception system can input image features, speed feature information, distance feature information, and angle feature information into the perception data fusion model in the decision-level neural network to obtain perception data fusion results; the low-altitude perception system can input the perception data fusion results into a pre-trained perception neural network, and output at least one perception information of speed information, distance information, angle information, quantity information, or type information of low-altitude objects through the pre-trained perception neural network.

[0135] The technical solution of this embodiment is to input the perception data fusion result into a pre-trained perception neural network, output at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network, and convert at least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object into a data packet to obtain the perception result of the low-altitude object perception task, so that the perception data fusion result obtained by multi-source perception quantity fusion can be effectively used to quickly identify the perception information such as distance, speed, angle, quantity and type of the low-altitude object, thereby realizing the effective use of multi-source perception data for low-altitude perception.

[0136] In another embodiment, Fig.13 As shown, a low-altitude sensing method is provided, which is applied to Figure 1 Taking the low altitude sensing system 102 in the example as an example, the method includes the following steps:

[0137] Step S1302, in response to a perception requirement request initiated for a low-altitude object perception task, determining a perception accuracy requirement and a perception area requirement for the low-altitude object perception task according to the perception requirement request.

[0138] Step S1304, screening out the sensing device types that meet the sensing accuracy requirements and the sensing area requirements.

[0139] Step S1306, obtaining the environmental conditions of the area to be sensed, and filtering out target sensing devices whose device type is a sensing device type and that meet the environmental conditions.

[0140] Step S1308, sending a perception data collection request to the target perception device; the perception data collection request is used to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data of at least two data types; the data type corresponds to the device type of the target perception device.

[0141] Step S1310, receiving multi-source sensing data transmitted back by the target sensing device.

[0142] Step S1312, obtaining the service request level of the perception demand request; the service request level is determined based on at least one of the perception accuracy requirements, real-time requirements, task urgency or environmental conditions of the low-altitude object perception task.

[0143] Step S1314, determining a perception data fusion model that matches the service request level; the model performance of the perception data fusion model is positively correlated with the service request level.

[0144] Step S1316: fusing the multi-source perception data through the perception data fusion model to obtain a perception data fusion result.

[0145] Step S1318: output the perception result of the low-altitude object perception task according to the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

[0146] It should be noted that the specific limitations of the above steps can refer to the specific limitations of a low-altitude perception method above.

[0147] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0148] Based on the same inventive concept, the embodiment of the present application also provides a low-altitude sensing device for implementing the low-altitude sensing method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of a low-altitude sensing device provided below can refer to the limitations of a low-altitude sensing method above, and will not be repeated here.

[0149] In an exemplary embodiment, Fig.14 As shown, a low-altitude object sensing device is provided, comprising:

[0150] A response module 1410 is used to respond to a perception demand request initiated for a low-altitude object perception task, and determine a target perception device that matches the perception demand request from a multi-source perception device including at least two device types;

[0151] The sending module 1420 is used to send a sensing data collection request to the target sensing device; the sensing data collection request is used to instruct the target sensing device to sense the low-altitude object and obtain multi-source sensing data of at least two data types; the data type corresponds to the device type of the target sensing device;

[0152] The fusion module 1430 is used to receive the multi-source perception data returned by the target perception device, and perform data fusion processing on the multi-source perception data to obtain a perception data fusion result;

[0153] The output module 1440 is used to output the perception result of the low-altitude object perception task according to the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

[0154] In one embodiment, the response module 1410 is specifically used to determine the perception accuracy requirement and the perception area requirement of the low-altitude object perception task according to the perception demand request; filter out the perception device type that meets the perception accuracy requirement and the perception area requirement; obtain the environmental conditions of the area to be perceived, and filter out the target perception device that meets the environmental conditions; the device type of the target perception device is the perception device type.

[0155] In one of the embodiments, the response module 1410 is specifically used to obtain a target monitoring method that matches the environmental conditions; the target monitoring method is adapted to the light conditions, weather conditions, and visibility conditions in the environmental conditions; and among the multi-source sensing devices in the area to be sensed, a sensing device that adopts the target monitoring method is selected as the target sensing device.

[0156] In one embodiment, the fusion module 1430 is specifically used to obtain the service request level of the perception demand request; the service request level is determined according to at least one of the perception accuracy requirements, real-time requirements, task urgency or environmental conditions of the low-altitude object perception task; determine a perception data fusion model that matches the service request level; the model performance of the perception data fusion model is positively correlated with the service request level; and fuse the multi-source perception data through the perception data fusion model to obtain the perception data fusion result.

[0157] In one embodiment, the multi-source perception data includes at least image data, radar signal data and cellular signal data, and the fusion module 1430 is specifically used to input the image data into a pre-trained image feature extraction network to obtain image features in the multi-source perception data; obtain radar signal features in the multi-source perception data based on the radar signal data; obtain network signal features in the multi-source perception data based on the cellular signal data; input the image features, radar signal features and network signal features into the perception data fusion model to obtain the perception data fusion result.

[0158] In one embodiment, the output module 1440 is used to obtain a pre-trained perception neural network; input the perception data fusion result into the pre-trained perception neural network, and output at least one perception information of the speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network; convert at least one perception information of the speed information, distance information, angle information, quantity information or type information of the low-altitude object into a data packet to obtain the perception result of the low-altitude object perception task.

[0159] Each module in the above-mentioned low-altitude object sensing device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0160] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.15 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a low-altitude perception method is implemented.

[0161] Those skilled in the art will understand that Fig.15 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0162] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned low-altitude sensing method. Here, the steps of the low-altitude sensing method can be the steps of the low-altitude sensing method in each of the above-mentioned embodiments.

[0163] In one embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the processor executes the steps of the above-mentioned low-altitude sensing method. The steps of the low-altitude sensing method here can be the steps of the low-altitude sensing method in each of the above-mentioned embodiments.

[0164] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the processor executes the steps of the above-mentioned low-altitude sensing method. Here, the steps of the low-altitude sensing method can be the steps of the low-altitude sensing method in each of the above-mentioned embodiments.

[0165] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0166] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0167] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0168] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A low-altitude object perception method, characterized in that: The method comprises: In response to a sensing requirement request initiated for a low-altitude object sensing task, determining a target sensing device matching the sensing requirement request from a multi-source sensing device including at least two device types; Sending a perception data collection request to the target perception device; the perception data collection request is used to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data; Receiving the multi-source perception data returned by the target perception device, and performing data fusion processing on the multi-source perception data to obtain a perception data fusion result; According to the perception data fusion result, the perception result of the low-altitude object perception task is output; the perception result is used to record the perception information of the low-altitude object.

2. The method according to claim 1, characterized in that The step of determining, among the multi-source sensing devices including at least two types of devices, a target sensing device matching the sensing demand request comprises: Determining the perception accuracy requirement and the perception area requirement of the low-altitude object perception task according to the perception demand request; Screening out the sensing device types that meet the sensing accuracy requirement and the sensing area requirement; The environmental conditions of the area to be sensed are obtained, and the target sensing device that meets the environmental conditions is screened out; the device type of the target sensing device is the sensing device type.

3. The method according to claim 2, characterized in that The step of screening out the target sensing device that meets the environmental condition includes: Acquire a target monitoring mode that matches the environmental conditions; the target monitoring mode is adapted to the light conditions, weather conditions, and sight distance conditions in the environmental conditions; Among the multi-source sensing devices in the area to be sensed, a sensing device that adopts the target monitoring mode is selected as the target sensing device.

4. The method according to claim 1, characterized in that: The multi-source perception data is subjected to data fusion processing to obtain a perception data fusion result, including: Obtaining a service request level of the perception demand request; the service request level is determined according to at least one of the perception accuracy requirement, real-time requirement, task urgency or environmental conditions of the low-altitude object perception task; Determining a perception data fusion model that matches the service request level; the model performance of the perception data fusion model is positively correlated with the service request level; The multi-source perception data is fused by the perception data fusion model to obtain the perception data fusion result.

5. The method according to claim 4, characterized in that The multi-source perception data includes at least image data, radar signal data, and cellular signal data. The multi-source perception data is fused by the perception data fusion model to obtain the perception data fusion result, including: Inputting the image data into a pre-trained image feature extraction network to obtain image features in the multi-source perception data; Obtaining radar signal features in the multi-source sensing data according to the radar signal data; Acquire network signal features in the multi-source sensing data according to the cellular signal data; The image features, radar signal features and network signal features are input into the perception data fusion model to obtain the perception data fusion result.

6. The method according to claim 1, characterized in that Outputting the perception result of the low-altitude object perception task according to the perception data fusion result includes: Get a pre-trained perceptron neural network; Inputting the perception data fusion result into the pre-trained perception neural network, and outputting at least one perception information of speed information, distance information, angle information, quantity information or type information of the low-altitude object through the pre-trained perception neural network; At least one of the speed information, distance information, angle information, quantity information or type information of the low-altitude object is converted into a data packet to obtain a perception result of the low-altitude object perception task.

7. A low-altitude object sensing device, characterized in that: The device comprises: A response module, configured to respond to a perception demand request initiated for a low-altitude object perception task, and determine a target perception device that matches the perception demand request from a multi-source perception device including at least two device types; A sending module, used for sending a perception data collection request to the target perception device; the perception data collection request is used to instruct the target perception device to perceive the low-altitude object and obtain multi-source perception data; A fusion module is used to receive the multi-source perception data returned by the target perception device, and perform data fusion processing on the multi-source perception data to obtain a perception data fusion result; The output module is used to output the perception result of the low-altitude object perception task according to the perception data fusion result; the perception result is used to record the perception information of the low-altitude object.

8. A low-altitude perception system, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images