CN116828156B - Geospatial event acquisition method, system, equipment, medium and acquisition box - Google Patents

Abstract

The present invention proposes a method, system, device, medium and collection box for collecting geospatial events, and relates to the field of spatial information technology. It includes: executing a loop process until the preset condition that the object detection model reaches convergence is met; wherein the loop process includes: according to the real-time picture of the drone image transmission, controlling the drone to continuously collect first image data including the monitoring target; constructing and training the corresponding object detection model in the cloud server based on the sample data obtained by manual annotation according to the first image data. According to the converged object detection model, the second image data is calculated and inferred to obtain the spatial data corresponding to the monitoring target, wherein the second image data is the image data including the monitoring target collected by the drone after the object detection model converges. Generate the corresponding geospatial event according to the spatial data. This solution can ensure the real-time generation of geospatial event data under low coupling.

Description

Method, system, device, medium and collection box for collecting geospatial events

Technical Field

The present invention relates to the field of spatial information technology, and in particular to a method, system, device, medium and collection box for collecting geospatial events.

Background technique

With the continuous development of geospatial information theory and technology, the construction of smart cities is also gradually advancing. In this context, the society's demand for geospatial data has gradually increased from historical and current basic data to real-time data.

The commonly used method at present is to use drones to conduct field surveys first, and then use professional software to process the drone survey data in the computer room to obtain standard products such as DOM and DEM. Finally, these standard products are used for application and analysis. This method not only affects the real-time nature of the data, but also makes it difficult to form event-type data.

Summary of the invention

The purpose of the present invention is to provide a method, system, device, medium and collection box for collecting geospatial events, which can ensure the real-time generation of geospatial event data under low coupling.

The present invention is achieved in that:

In a first aspect, the present application provides a method for collecting geospatial events, comprising the following steps:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model reaches convergence, and the loop process includes: according to the real-time picture of the drone image transmission, the drone is controlled to continuously collect the first image data including the monitoring target; according to the first image data, manual annotation is performed to obtain the corresponding sample data; according to the sample data, the corresponding object detection model is constructed and trained in the cloud server. According to the converged object detection model, the second image data is calculated and inferred to obtain the spatial data corresponding to the monitoring target, wherein the second image data is the image data including the monitoring target collected by the drone after the object detection model converges. The corresponding geospatial event is generated according to the above spatial data.

Furthermore, based on the aforementioned scheme, the method further includes the following steps: performing spatial analysis on the above-mentioned geographic spatial events according to a GIS spatial analysis algorithm to obtain corresponding event analysis results.

Further, based on the aforementioned scheme, the above-mentioned spatial data includes at least one of geometric shape information, location information, event identification, event category and time information.

In a second aspect, the present application provides a method for collecting geospatial events, comprising the following steps:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model converges, and the loop process includes: planning the flight route of the drone based on the real-time images of the drone image transmission observed by the ground station, and sending the third image data including the monitoring target continuously collected by the drone during the flight route to the cloud server; performing manual annotation based on the third image data to obtain the corresponding sample data to train and verify the corresponding object detection model in the cloud server. After receiving the object detection model trained to convergence, the acquisition box uses the object detection model to detect the target object on the real-time images of the event area to be monitored newly collected by the drone, obtain the corresponding spatial data, and send the geospatial events generated based on the above spatial data to the cloud server for the cloud service to perform spatial analysis.

Further, based on the above-mentioned scheme, when the geospatial events generated based on the above-mentioned spatial data are sent to the cloud server, it includes: converting the above-mentioned spatial data into data that conforms to the OGC spatial data standard based on a predetermined GIS event model, and storing the data in a corresponding memory and sending it to the cloud server.

In a third aspect, the present application provides a collection box, comprising:

The receiving module is configured to: receive the object detection model trained to convergence by the cloud server, wherein the step of the cloud server training the above object detection model is: continuously receiving the fourth image data including the monitoring target collected in real time by the drone during the execution of the corresponding flight route, and training and verifying the corresponding object detection model based on the sample data obtained by the corresponding personnel continuously manually annotating the fourth image data, until the above object detection model converges. The processing module is configured to: use the received object detection model trained to convergence by the cloud server to perform target object detection on the real-time picture of the event area to be monitored newly collected by the drone to obtain the spatial data corresponding to the target event. The sending module is configured to: send the above spatial data to the cloud server and/or the target terminal device for the cloud service and/or the target terminal device to perform spatial analysis.

In a fourth aspect, the present application provides a real-time collection system for geospatial events, which includes:

The model training module is configured to: execute a loop process until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: controlling the drone to continuously collect first image data including the monitoring target according to the real-time picture transmitted by the drone; manually annotating the first image data to obtain corresponding sample data; and constructing and training the corresponding object detection model in the cloud server according to the sample data. The spatial data generation module is configured to: perform computational reasoning on the second image data according to the converged object detection model to obtain spatial data corresponding to the monitoring target, wherein the second image data is the image data including the monitoring target collected by the drone after the object detection model converges. The spatial event generation module is configured to: generate corresponding geospatial events according to the above spatial data.

In a fifth aspect, the present application provides a real-time collection system for geospatial events, which includes: a drone, a collection box, a ground station and a cloud server.

The above-mentioned UAV is used to receive the control signal of the ground station to execute the corresponding flight route in the event area to be monitored, and transmit the acquired real-time picture to the above-mentioned ground station for display, and send the collected image data including the monitoring target to the above-mentioned cloud service. The above-mentioned ground station is used to receive and display the real-time picture acquired by the UAV, and send the corresponding control signal to the UAV in response to the predetermined operation instruction. The above-mentioned acquisition box is used to receive the object detection model trained to convergence by the cloud server, and based on the object detection model, perform target object detection on the real-time picture of the event area to be monitored newly acquired by the UAV to obtain corresponding spatial data, and send the above-mentioned spatial data to the cloud server. The above-mentioned cloud server is used to continuously receive the fifth image data including the monitoring target collected in real time by the UAV in the execution of the corresponding flight route, and train and verify the corresponding object detection model based on the sample data obtained by the corresponding personnel continuously manually annotating the fifth image data until the above-mentioned object detection model converges; and send the object detection model trained to convergence to the above-mentioned acquisition box; and also receive the spatial data sent by the acquisition box for spatial analysis.

In a sixth aspect, the present application provides an electronic device comprising at least one processor, at least one memory and a data bus; wherein: the processor and the memory communicate with each other via the data bus; the memory stores program instructions to be executed by the processor, and the processor calls the program instructions to execute a method as described in any one of the first and second aspects above.

In a seventh aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of the first and second aspects above.

Compared with the prior art, the present invention has at least the following advantages or beneficial effects:

First, based on the real-time image transmitted by the drone, the drone is controlled to continuously collect image data including the monitored target. Then, after manually annotating the image data, the corresponding object detection model is trained in the cloud server, so that a converged object detection model related to the monitored target can be quickly obtained. Then, the converged object detection model can be used to perform computational reasoning on the image data acquired by the drone next, so as to obtain spatial data corresponding to the detection target that can be used to generate geospatial events. In other words, this technical solution can easily and conveniently obtain real-time geospatial events, which is convenient for monitoring sudden events.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

FIG1 is a flow chart of a method for collecting geospatial events provided by an embodiment of the present invention;

FIG2 is a flow chart of a method for collecting geospatial events provided by another embodiment of the present invention;

FIG3 is a structural block diagram of a collection box provided by an embodiment of the present invention;

FIG4 is a structural block diagram of a real-time collection system for geospatial events provided by an embodiment of the present invention;

FIG5 is a corresponding signaling diagram of an embodiment of a real-time collection system for geospatial events provided by an embodiment of the present invention;

FIG6 is a structural block diagram of an electronic device provided by an embodiment of the present invention.

Icon: 101, processor; 102, memory; 103, data bus.

Detailed ways

In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

Some embodiments of the present application are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and the features in the embodiments may be combined with each other. At the same time, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

Example 1

In the prior art, a model is pre-trained for the target object of interest, and then the model is set on a drone, which is then used to monitor the target object on site. However, this method cannot achieve real-time monitoring using drones, and is only suitable for monitoring general targets, but not for monitoring sudden events.

In order to address the above-mentioned problems, in a first aspect of an embodiment of the present application, a method for collecting geospatial events is provided, which can obtain real-time geospatial events by optimizing the processing flow, thereby facilitating the monitoring of sudden events.

Referring to FIG. 1 , the method for collecting geospatial events includes the following steps:

Step S101: Execute a loop process until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: according to the real-time picture of the drone image transmission, controlling the drone to continuously collect first image data including the monitoring target; manually annotating according to the first image data to obtain corresponding sample data; and constructing and training a corresponding object detection model in the cloud server according to the sample data.

In the above steps, when the drone is controlled to fly to the event area to be detected (including the area of the monitoring target), real-time image data including the monitoring target is collected, and then the collected image data including the monitoring target is used to understand the specific situation on the scene. Then, the real-time picture of the image transmission can be observed according to the monitoring target, and the drone can be controlled to execute the corresponding flight route, and the first image data including the monitoring target can be continuously collected on the way. Exemplarily, the operator can observe the real-time picture of the drone image transmission on the display of the ground station, and then manually observe the real-time picture of the image transmission, and then control the drone in real time to execute the corresponding flight route according to the observed situation, so that a large amount of first image data including the monitoring target can be accurately obtained by using the drone in combination with artificial factors.

After uploading the first image data to the cloud server, the corresponding staff can manually annotate the first image data in the cloud server, or use the client tool to continuously obtain the latest first image data from the cloud server and annotate it to obtain the corresponding sample data. Then, based on the obtained sample data, the corresponding object detection model can be constructed and trained in the cloud server using the sample data. During the training process, new sample data needs to be continuously obtained, and the model needs to be trained and verified until the model converges and exits the loop.

It should be noted that the reason why object detection models need to be trained is mainly because the monitoring targets are not determined in advance. The monitoring targets need to be selected according to the actual situation, so it is impossible to obtain pre-trained object detection models for detection and identification. For example, the monitoring target may be a specific object target temporarily determined in the case of natural disasters, traffic accidents, fires, public security incidents, etc. It is impossible to use a general object detection model to identify and detect it, and then monitor it.

Step S102: performing calculation and reasoning on the second image data according to the converged object detection model to obtain spatial data corresponding to the monitoring target, wherein the second image data is image data including the monitoring target collected by the drone after the object detection model converges;

Step S103: Generate corresponding geospatial events according to the spatial data.

After obtaining the converged object detection model, the drone can be further used to monitor the monitored area including the monitoring target, that is, the second image data obtained by the drone can be calculated and inferred to obtain the spatial data corresponding to the monitoring target, so as to further use the spatial data to generate the corresponding geospatial event. Among them, in some implementations of the present invention, the spatial data includes at least one of geometric shape information, location information, event identification, event category and time information. Of course, the specific spatial data can also be selected autonomously as needed, and the specific data included in the spatial data is not strictly limited here. For example, the spatial data obtained by identification may not only be coordinates, but also shapes. If the shape is a point, its expression form can be coordinates, but the shape may also be a line or a surface, and the corresponding expression form is adopted. In short, what the spatial data includes needs to be selected and applied according to the actual situation. Exemplarily, the subsequent processing can also be to use the image coordinates of the detected target object and the instantaneous POS data of the drone for coordinate processing, and then the monitored coordinate data is packaged with time and ID into EVENT events and sent to the cloud.

It should be noted that the first image data and the second image data are essentially image data including monitoring targets acquired by the drone, but the first image data is image data acquired in the process of training the object model, and the second image data is image data provided to the object detection model for target object detection after the converged object detection model is obtained. Therefore, if the acquisition process based on the first image data is: the operator observes the real-time image of the drone image transmission on the display of the ground station, and then manually observes the real-time image of the image transmission, and then controls the drone to execute the corresponding flight route, so that the drone can be combined with artificial factors to accurately acquire a large amount of first image data including monitoring targets; then correspondingly, the acquisition process of the second image data can be: after obtaining the converged object detection model, notify the corresponding operator, and then the operator will be able to start re-planning the flight route of the drone, and then acquire the second image data, and use the converged object detection model to detect the target object.

In addition, for geospatial events, in some embodiments, they can be understood as specific things that occur at a certain time and place, such as natural disasters, traffic accidents, fires, public security incidents, etc. At this time, it is a spatial data type organized and described by the geographic information spatial data model, generally with spatial characteristics, attribute characteristics and time characteristics. Among them, the spatial characteristics mainly describe its geographical location spatial coordinates, geometric shape, etc.; the attribute characteristics mainly describe its industry business information, such as name, type, etc.; the time characteristics describe the time information of its occurrence, including the specific time of the event, duration, and event sequence information.

Based on the above scheme, in some implementations of the present invention, after obtaining the geospatial event, the above method further includes the following steps: spatially analyzing the geospatial event according to the GIS spatial analysis algorithm to obtain the corresponding event analysis result. Thus, the event analysis result can be used to directly apply the event in the server, or it can be sent or copied to a stand-alone device to apply the event. That is, various application analyses can be performed on the corresponding server or terminal device in the manner of geographic information system spatial analysis.

As shown in FIG. 2 , in the second aspect of this embodiment, based on the same inventive concept, this embodiment of the present application further provides a method for collecting geospatial events, which includes the following steps:

Step S201: executing a loop process until a preset condition is met; wherein the preset condition is that the object detection model converges, and the loop process includes: planning the flight route of the drone based on the real-time image transmitted by the drone observed by the ground station, and sending the third image data including the monitoring target continuously collected by the drone on the flight route to the cloud server; manually annotating the third image data to obtain corresponding sample data, so as to train and verify the corresponding object detection model in the cloud server;

Step S202: After receiving the object detection model trained to convergence, the acquisition box uses the object detection model to detect the target object on the real-time image of the event area to be monitored newly acquired by the drone, obtains the corresponding spatial data, and sends the geospatial events generated based on the spatial data to the cloud server for the cloud service to perform spatial analysis. If the spatial data only includes the coordinates and geometric shape of the target object, it can be further packaged and combined with the attributes of the target object to generate a geospatial event.

On the basis of the method for collecting geospatial events disclosed in the first aspect, the ground station and the collection box are combined in the scheme of the second aspect, and the technical scheme is further optimized and limited. Among them, considering that there is no converged object detection model at the beginning, a certain amount of sample data is required to train the object detection model. However, if the drone only autonomously executes the established flight route to obtain image data including the monitored target in the monitored area, the accuracy of image data acquisition may be affected by the selection of the flight route, which is reduced, which is not conducive to the rapid convergence of the object detection model. Therefore, in the above embodiment, by planning the flight route of the drone based on the real-time picture of the drone image transmission observed by the ground station during the training of the object detection model, the drone can be controlled to obtain the third image data including the monitored target more accurately and effectively. It should be noted that the third image data here and the first image data in the previous text are essentially image data, but it is said that this is a description made to facilitate the distinction between the image data obtained by the drone under two different flight control conditions.

In addition, after the cloud server has trained and obtained a converged object detection model, the acquisition box can be used to download and receive the object detection model, and then the object detection model can be used on the acquisition box to perform target object detection on the real-time images of the event area to be monitored newly collected by the drone, thereby sharing the computational workload of target object detection to the acquisition box. This can improve the real-time performance of subsequent detection and tracking calculations of the monitored targets, and will not be affected by too many network fluctuations, which will lead to a decrease in timeliness.

Based on the aforementioned scheme, in some implementations of the present invention, sending the geospatial events generated based on the spatial data to the cloud server includes: converting the spatial data into data that complies with the OGC spatial data standard based on a predetermined GIS event model (including converting the structure and format of the spatial data), storing the data in a corresponding memory and sending it to the cloud server.

Example 2

As shown in FIG3 , based on the same inventive concept as in Example 1, the embodiment of the present application provides a collection box, which includes:

The receiving module is configured to: receive the object detection model trained to convergence by the cloud server, wherein the step of the cloud server training the object detection model is: continuously receiving the fourth image data including the monitoring target collected in real time by the drone during the execution of the corresponding flight route, and training and verifying the corresponding object detection model based on the sample data obtained by the corresponding personnel continuously manually annotating the fourth image data, until the object detection model converges;

The processing module is configured to: use the received object detection model trained to convergence by the cloud server to perform target object detection on the real-time picture of the event area to be monitored newly collected by the drone to obtain spatial data corresponding to the target event;

The sending module is configured to: send the spatial data to the cloud server and/or the target terminal device for the cloud service and/or the target terminal device to perform spatial analysis.

Based on the method for collecting geospatial events disclosed in Example 1, in the above Example 2, a collection box is disclosed, including a receiving module, a processing module and a sending model, so as to receive the object detection model trained to convergence by the cloud server, and based on this, perform target object detection on the real-time picture of the event area to be monitored newly collected by the drone, so as to obtain the spatial data corresponding to the target event, and then send it to the cloud server and/or the target terminal device for the cloud server and/or the target terminal device to perform spatial analysis. Among them, the target terminal device includes terminal devices such as computers, tablets and mobile phones.

It should be noted that the fourth image data here and the first image data and the third image data in the previous text are essentially image data. It is just that the description here is made to distinguish the image data obtained by the receiving drone in order to facilitate the understanding and distinction of the technical solution of the acquisition box.

In addition, in some implementations of the present invention, the collection box includes:

The receiving module is configured to: receive an object detection model trained to convergence by a cloud server, wherein the step of the cloud server training the object detection model is: executing a loop process until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: planning a flight route of the drone based on the real-time image transmitted by the drone observed by the ground station, and sending third image data including the monitored target continuously collected by the drone on the flight route to the cloud server; performing manual annotation based on the third image data to obtain corresponding sample data, so as to train and verify the corresponding object detection model in the cloud server.

The processing module is configured to: use the received object detection model trained to convergence by the cloud server to perform target object detection on the real-time images of the event area to be monitored newly collected by the drone to obtain spatial data corresponding to the target event.

The sending module is configured to: send the spatial data to the cloud server and/or the target terminal device for the cloud service and/or the target terminal device to perform spatial analysis.

For the specific implementation process of the above-mentioned collection box, please refer to the method for collecting geospatial events provided in Example 1, which will not be repeated here.

Example 3

Referring to FIG. 4 , based on the same inventive concept as the method for collecting geospatial events disclosed in the first aspect of Example 1, this embodiment of the present application provides a real-time collection system for geospatial events, which includes:

The model training module is configured to: execute a loop process until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: controlling the drone to continuously collect first image data including the monitored target according to the real-time image transmitted by the drone; manually annotating the first image data to obtain corresponding sample data; and constructing and training a corresponding object detection model in a cloud server according to the sample data;

The spatial data generation module is configured to: perform calculation and reasoning on the second image data according to the converged object detection model to obtain spatial data corresponding to the monitoring target, wherein the second image data is image data including the monitoring target collected by the drone after the object detection model converges;

The spatial event generation module is configured to generate corresponding geographic spatial events according to the spatial data.

For the specific implementation process of the above system, please refer to the method for collecting geospatial events provided in Example 1, which will not be repeated here.

Based on the same inventive concept as the method for collecting geospatial events disclosed in the second aspect of Example 1, as shown in FIG5 , the embodiment of the present application further provides a real-time collection system for geospatial events, which includes: a drone, a collection box, a ground station, and a cloud server;

The drone is used to receive control signals from the ground station to execute a corresponding flight route in the event area to be monitored, transmit the acquired real-time images to the ground station for display, and send the collected image data including the monitored target to the cloud service;

The ground station is used to receive and display the real-time images acquired by the UAV, and send corresponding control signals to the UAV in response to predetermined operation instructions;

The acquisition box is used to receive the object detection model trained to convergence by the cloud server, and based on the object detection model, perform target object detection on the real-time picture of the event area to be monitored newly acquired by the drone to obtain corresponding spatial data, and send the spatial data to the cloud server;

The cloud server is used to continuously receive fifth image data including monitoring targets collected in real time by the drone during the execution of the corresponding flight route, and to train and verify the corresponding object detection model based on sample data obtained by the corresponding personnel continuously manually annotating the fifth image data until the object detection model converges; and to send the object detection model trained to convergence to the acquisition box; and to receive spatial data sent by the acquisition box for spatial analysis.

As shown in FIG5 , the real-time collection system of the above-mentioned geospatial events includes a drone, a collection box, a ground station and a cloud server, wherein the drone is used to obtain image data including monitoring targets, and then the operator can control the drone in real time at the ground station to execute the corresponding flight route (whether it is the model training stage or the later stage of using the converged model, the drone can be controlled at the ground station to execute the corresponding flight route as needed), and the cloud server is used to receive the image data including the monitoring targets, and after manually annotating the image data, train and verify the corresponding object detection model until the model converges. Then, the collection box can be used to download the received converged object detection model, and the target object detection can be performed on the image data including the monitoring targets acquired by the drone next, so as to obtain the corresponding spatial data, so as to provide the cloud server with spatial analysis.

For the specific implementation process of the above system, please refer to the method for collecting geospatial events provided in Example 1, which will not be repeated here.

Example 4

Please refer to FIG6 , an embodiment of the present application provides an electronic device, which includes at least one processor 101, at least one memory 102 and a data bus 103; wherein: the processor 101 and the memory 102 communicate with each other through the data bus 103; the memory 102 stores program instructions that can be executed by the processor 101, and the processor 101 calls the program instructions to execute a method for collecting geospatial events. For example, the following is implemented:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model reaches convergence, and the loop process includes: according to the real-time picture of the drone image transmission, the drone is controlled to continuously collect the first image data including the monitoring target; according to the first image data, manual annotation is performed to obtain the corresponding sample data; according to the sample data, the corresponding object detection model is constructed and trained in the cloud server. According to the converged object detection model, the second image data is calculated and inferred to obtain the spatial data corresponding to the monitoring target, wherein the second image data is the image data including the monitoring target collected by the drone after the object detection model converges. The corresponding geospatial event is generated according to the above spatial data.

Or implement:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model converges, and the loop process includes: planning the flight route of the drone based on the real-time images of the drone image transmission observed by the ground station, and sending the third image data including the monitoring target continuously collected by the drone during the flight route to the cloud server; performing manual annotation based on the third image data to obtain the corresponding sample data to train and verify the corresponding object detection model in the cloud server. After receiving the object detection model trained to convergence, the acquisition box uses the object detection model to detect the target object on the real-time images of the event area to be monitored newly collected by the drone, obtain the corresponding spatial data, and send the geospatial events generated based on the above spatial data to the cloud server for the cloud service to perform spatial analysis.

Among them, the memory 102 can be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc.

The processor 101 may be an integrated circuit chip with signal processing capability. The processor 101 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

It is understood that the structure shown in Figure 6 is only for illustration, and the electronic device may also include more or fewer components than those shown in Figure 6, or have a different configuration than that shown in Figure 6. Each component shown in Figure 6 may be implemented by hardware, software, or a combination thereof.

Example 5

The present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by the processor 101, a method for collecting geospatial events is implemented. For example, the following is implemented:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model reaches convergence, and the loop process includes: according to the real-time picture of the drone image transmission, the drone is controlled to continuously collect the first image data including the monitoring target; according to the first image data, manual annotation is performed to obtain the corresponding sample data; according to the sample data, the corresponding object detection model is constructed and trained in the cloud server. According to the converged object detection model, the second image data is calculated and inferred to obtain the spatial data corresponding to the monitoring target, wherein the second image data is the image data including the monitoring target collected by the drone after the object detection model converges. The corresponding geospatial event is generated according to the above spatial data.

Or implement:

The loop process is executed until the preset conditions are met; wherein the preset conditions are that the object detection model converges, and the loop process includes: planning the flight route of the drone based on the real-time images of the drone image transmission observed by the ground station, and sending the third image data including the monitoring target continuously collected by the drone during the flight route to the cloud server; performing manual annotation based on the third image data to obtain the corresponding sample data to train and verify the corresponding object detection model in the cloud server. After receiving the object detection model trained to convergence, the acquisition box uses the object detection model to detect the target object on the real-time images of the event area to be monitored newly collected by the drone, obtain the corresponding spatial data, and send the geospatial events generated based on the above spatial data to the cloud server for the cloud service to perform spatial analysis.

If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program codes.

It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present application. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims (10)

1. A method for collecting geospatial events, characterized in that it comprises the following steps: The loop process is executed until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: according to the real-time picture transmitted by the drone, controlling the drone to continuously collect first image data including the monitoring target, wherein the monitoring target is not predetermined in advance; manually annotating according to the first image data to obtain corresponding sample data; and constructing and training the corresponding object detection model in the cloud server according to the sample data; In response to the convergence of the object detection model, according to the latest real-time image transmitted by the drone, re-plan the flight route and control the drone to execute the corresponding flight route based on the re-planned flight route; Performing computational inference on the second image data according to the converged object detection model to obtain spatial data corresponding to the monitoring target, wherein the second image data is image data including the monitoring target acquired by the drone after the object detection model converges; A corresponding geospatial event is generated according to the spatial data, wherein the geospatial event is a sudden event. 2. The method for collecting geospatial events according to claim 1, further comprising: The geographic space events are spatially analyzed according to the GIS spatial analysis algorithm to obtain corresponding event analysis results. 3. The method for collecting geospatial events as described in claim 1 is characterized in that the spatial data includes at least one of geometric shape information, location information, event identification, event category and time information. 4. A method for collecting geospatial events, characterized in that it comprises the following steps: The loop process is executed until a preset condition is met; wherein the preset condition is that the object detection model converges, and the loop process includes: planning the flight route of the drone based on the real-time image transmitted by the drone observed by the ground station, and sending the third image data including the monitoring target continuously collected by the drone on the flight route to the cloud server, wherein the monitoring target is not determined in advance; manually annotating the third image data to obtain corresponding sample data, so as to construct and train the corresponding object detection model in the cloud server according to the sample data; In response to the convergence of the object detection model, the object detection model trained to convergence is downloaded from the cloud server to the acquisition box, and according to the latest real-time image transmitted by the drone, the flight route is re-planned and based on the re-planned flight route, the drone is controlled to execute the corresponding flight route; After receiving the object detection model trained to convergence, the acquisition box uses the object detection model to detect target objects on the real-time images of the event area to be monitored newly acquired by the drone, obtains corresponding spatial data, and sends the geospatial events generated based on the spatial data to the cloud server for spatial analysis by the cloud service, where the geospatial events are sudden events. 5. The method for collecting geospatial events according to claim 4, wherein sending the geospatial events generated based on the spatial data to a cloud server comprises: The spatial data is converted into data that complies with the OGC spatial data standard based on a predetermined GIS event model, and the data is stored in a corresponding memory and sent to a cloud server. 6. A collection box, characterized in that it includes: The receiving module is configured to: receive an object detection model trained to convergence by a cloud server, wherein the step of the cloud server training the object detection model is: continuously receiving fourth image data including a monitoring target collected in real time by the drone during the execution of a corresponding flight route, and based on sample data obtained by continuous manual annotation of the fourth image data by corresponding personnel, construct and train a corresponding object detection model in the cloud server until the object detection model converges; wherein the monitoring target is not determined in advance; The processing module is configured to: use the received object detection model trained to convergence by the cloud server to perform target object detection on the real-time picture of the event area to be monitored newly acquired by the drone to obtain spatial data corresponding to the target event; wherein the real-time picture of the event area to be monitored newly acquired by the drone is the picture collected in real time when the operator responds to the convergence of the object detection model, re-plans the flight route according to the latest real-time picture transmitted by the drone, and controls the drone to execute the corresponding flight route based on the re-planned flight route; The sending module is configured to: send the spatial data to the cloud server and/or the target terminal device to generate sudden geospatial events based on the spatial data for the cloud service and/or the target terminal device to perform spatial analysis. 7. A real-time collection system for geospatial events, comprising: The model training module is configured to: execute a loop process until a preset condition is met; wherein the preset condition is that the object detection model reaches convergence, and the loop process includes: controlling the drone to continuously collect first image data including a monitoring target according to the real-time image transmitted by the drone, wherein the monitoring target is not predetermined in advance; manually annotating the first image data to obtain corresponding sample data; and constructing and training a corresponding object detection model in a cloud server according to the sample data; The route planning module is configured to: in response to the convergence of the object detection model, re-plan the flight route according to the latest real-time image transmitted by the drone, and control the drone to execute the corresponding flight route based on the re-planned flight route; The spatial data generation module is configured to: perform calculation and reasoning on the second image data according to the converged object detection model to obtain spatial data corresponding to the monitoring target, wherein the second image data is image data including the monitoring target collected by the drone after the object detection model converges; The spatial event generation module is configured to generate corresponding geographic spatial events according to the spatial data, wherein the geographic spatial events are sudden events. 8. A real-time collection system for geospatial events, characterized by comprising a drone, a collection box, a ground station and a cloud server; The drone is used to receive control signals from the ground station to execute a corresponding flight route in the event area to be monitored, transmit the acquired real-time images to the ground station for display, and send the collected image data including the monitoring target to the cloud service, wherein the monitoring target is not determined in advance; The ground station is used to receive and display the real-time images acquired by the drone, and send corresponding control signals to the drone in response to predetermined operation instructions; the control signals include signals for controlling the drone to continuously collect image data including the detection target according to the real-time images acquired by the drone, and also include signals for controlling the drone to execute the corresponding flight route based on the newly-planned flight route and the latest real-time images transmitted by the drone in response to the convergence of the object detection model; The acquisition box is used to receive the object detection model trained to convergence by the cloud server, and based on the object detection model, perform target object detection on the real-time picture of the event area to be monitored newly acquired by the drone to obtain corresponding spatial data, and send the spatial data to the cloud server; The cloud server is used to continuously receive fifth image data including monitoring targets collected in real time by the drone during the execution of the corresponding flight route, and to construct and train a corresponding object detection model in the cloud server based on sample data obtained by continuous manual annotation of the fifth image data by corresponding personnel until the object detection model converges; and to send the object detection model trained to convergence to the acquisition box; and to receive spatial data sent by the acquisition box for spatial analysis, wherein the spatial data is used to generate data of sudden geospatial events. 9. An electronic device, characterized in that it comprises at least one processor, at least one memory and a data bus; wherein: the processor and the memory communicate with each other through the data bus; the memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the method as described in any one of claims 1-5. 10. A computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method according to any one of claims 1 to 5 when executed by a processor.