An autonomous inspection system for unmanned aerial vehicles

Abstract

The present invention discloses an unmanned aerial vehicle (UAV) autonomous inspection system, comprising: a remote monitoring device that receives and stores information related to the traffic conditions of each road section interval from at least one big data platform, generates a task instruction for the UAV to perform autonomous cruising, and sends it to a cloud server; the cloud server that receives and analyzes the task instruction, generates a control instruction for the UAV to perform autonomous cruising, and sends the control instruction to a UAV control device according to the information on the position of the road section interval; the UAV control device that receives the control instruction from the cloud server, selects a UAV according to a predetermined rule, and sends the control instruction to the selected UAV; and the UAV that performs autonomous cruising based on the control instruction received from the UAV control device. According to the present invention, by statistically analyzing the traffic condition information from various channels, the accuracy and efficiency of the UAV autonomous inspection system for road network monitoring are improved. The present invention is applied to the construction of a large-scale mineral database in the Qinghai-Tibet Plateau.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic dispatching, and particularly relates to an unmanned aerial vehicle (UAV) autonomous inspection system. Background Art

[0002] In the urban traffic management system, camera monitoring devices or vehicle speed measurement devices are usually installed at key nodes to monitor the traffic flow conditions here in real time. For the road network of a certain area, considering the installation cost and maintenance cost, etc., the number of such camera monitoring devices and vehicle speed measurement devices is limited. In addition, the installation locations of the camera monitoring devices and vehicle speed measurement devices are fixed. Relying on this traditional mode to monitor road traffic flow and illegal acts, there are problems such as monitoring blind spots caused by fixed equipment positions and untimely fault repair. In addition, when due to some special reasons, such as vehicle traffic accidents, collapse of roads or bridges, fixed-installed devices cannot operate normally due to natural disasters or faults, etc., traffic congestion or even interruption may occur at uncertain locations, and there are situations where it is necessary to obtain on-site situations and relevant information as quickly as possible.

[0003] In recent years, the application of UAVs in the field of road traffic safety has been increasing, including daily traffic information collection, road traffic event detection, accident emergency rescue, etc. Compared with traditional traffic methods, UAVs have the advantages of convenient use, flexible application, fast flight speed, etc. Moreover, by carrying professional cameras, UAVs can clearly monitor a large area with the advantage of high altitude, greatly improving the efficiency of traffic management law enforcement, but at the same time also facing challenges in terms of technology, operation, etc.

[0004] Currently, for example, in applications such as road traffic event detection and accident emergency rescue, the task instructions of UAVs are usually determined by staff based on monitoring images from traffic monitoring management centers, alarm information from road vehicles, and traffic condition sharing information, etc., to determine whether to start UAV network inspections. However, generally, the data source channels of monitoring images, alarm information, and traffic condition sharing information, etc., are scattered, information aggregation takes time, and it also takes time and experience for staff to make accurate judgments based on these data. There is room for improvement in the response speed of issuing task instructions for UAV inspections. Summary of the Invention

[0005] Aiming at the above deficiencies in the prior art, the UAV autonomous inspection system provided by the present invention solves the problem of low accuracy and efficiency in starting the UAV autonomous inspection for existing road network monitoring.

[0006] To achieve the above invention purpose, the technical solution adopted by the present invention is: a UAV autonomous inspection system, including:

[0007] A remote monitoring device receives and stores information related to traffic conditions of each road section interval from at least one big data platform, generates a task instruction for a drone to perform autonomous cruise on the corresponding road section interval based on the received information related to traffic conditions, and sends the generated task instruction to a cloud server. The task instruction includes information indicating the location of the road section interval.

[0008] The cloud server receives the task instruction for the drone sent by the remote monitoring device, parses the task instruction to generate a control instruction for the drone to perform autonomous cruise, and sends the control instruction to at least one drone control device according to the information indicating the location of the road section interval obtained by parsing.

[0009] At least one drone control device receives the control instruction from the cloud server, selects a drone according to a predetermined rule, and sends the control instruction to the selected drone.

[0010] And at least one drone performs autonomous cruise based on the control instruction received from the at least one drone control device.

[0011] Furthermore: The at least one big data platform includes at least one of a traffic accident alarm platform, a road traffic monitoring platform, an automobile navigation system platform, and a natural disaster rescue platform.

[0012] Furthermore: The remote monitoring device includes:

[0013] A traffic information database stores the received information related to traffic conditions as traffic condition records according to road section intervals for each predetermined period. An item for recording an autonomous cruise warning value is set in the traffic condition records, and the autonomous cruise warning value is the cumulative value of warning values corresponding to emergency situations.

[0014] A task instruction generation unit generates a task instruction for autonomous cruise for the road section interval corresponding to this traffic condition record when the autonomous cruise warning value recorded in any of the traffic condition records exceeds a preset threshold.

[0015] And a communication unit sends the task instruction to the cloud server.

[0016] Furthermore: The cloud server parses the task instruction to generate the control instruction and sends the control instruction to the at least one drone control device according to the road section interval corresponding to the task instruction.

[0017] The at least one drone control device selects at least one drone according to the endurance mileage of the drone and sends the control instruction to it.

[0018] Furthermore, the specific calculation method of the autonomous cruise warning value is as follows:

[0019] In response to receiving traffic accident alarm information or traffic congestion information from the at least one big data platform, the control remote monitoring device extracts description information of the accident occurrence location or traffic congestion location and keywords of the accident situation or congestion situation from the traffic accident alarm information or traffic congestion information through any one of the pre-trained speech recognition neural network, text extraction neural network, and image recognition neural network, locates the accident occurrence location or traffic congestion location according to the description information of the accident occurrence location or traffic congestion location, searches for relevant traffic condition records according to the accident occurrence location or traffic congestion location, and accumulates the autonomous cruise warning value in the traffic condition records corresponding to the accident situation keywords or the congestion situation keywords;

[0020] In response to receiving natural disaster information from the at least one big data platform, the control remote monitoring device extracts description information of the natural disaster occurrence location and natural disaster situation keywords from the natural disaster information through any one of the speech recognition neural network, text extraction neural network, and image recognition neural network, locates the natural disaster occurrence location according to the description information of the natural disaster occurrence location, searches for relevant traffic condition records according to the natural disaster occurrence location, and accumulates the autonomous cruise warning value in the traffic condition records corresponding to the natural disaster situation keywords.

[0021] Furthermore, the drone control device is set near the drone library or set as an independent base station in the monitoring area including the service area of the highway and the area prone to natural disasters;

[0022] The drone library is used to store the at least one drone.

[0023] Furthermore, the road traffic monitoring platform includes traffic information collection units set at the entrances and exits of each road section interval;

[0024] The traffic information collection unit is used to collect traffic flow information of each road section interval, and send the location information, traffic flow information, and information collection time information of the road section interval as the information related to traffic conditions to the remote monitoring device;

[0025] The remote monitoring device saves the information related to traffic conditions to the traffic condition records of the road section interval according to the location information of each road section interval.

[0026] Furthermore, the cloud server also generates an inspection map based on the map data and the 3D model, and plans the inspection route, inspection points, and the poses of the drones at the inspection points during the construction of the inspection map.

[0027] Among them, the 3D model is established by drone scanning. During the scanning process, the lidar carried on the drone is used to obtain map features to ensure that the coordinates of each object in the 3D model are consistent with the real world. The 3D model and the inspection map share a coordinate system, which is used to mark or identify objects not reflected in the existing map.

[0028] Furthermore, the cloud server also performs object recognition on the objects based on the images received from the at least one drone, confirms each object and its status at the inspection point, and displays them in real time to the operator through a remote monitoring device.

[0029] The beneficial effects of the present invention are as follows: The drone autonomous inspection system of the present invention records the traffic conditions according to road sections, counts the information related to traffic conditions received from multiple big data platforms for each road section. When the autonomous cruise warning value in the record corresponding to any road section exceeds the preset threshold, a task instruction is automatically generated to indicate the drone to start the inspection of that road section. It can aggregate the data of big data platforms from multiple channels, thus preventing the omission of monitoring data, and can automatically generate a start instruction for the drone to start autonomous cruise according to the severity of traffic accidents, congestion, and natural disasters, improving the accuracy and efficiency of the drone autonomous inspection system for road network monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic structural diagram of a drone autonomous inspection system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0031] The following describes the specific embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

[0032] Such as Figure 1As shown, in an embodiment of the present invention, the UAV autonomous inspection system includes: a remote monitoring device that receives information related to traffic conditions of each road section interval from at least one big data platform and stores it, generates a task instruction for the UAV to perform autonomous cruise on the corresponding road section interval based on the received information related to traffic conditions, and sends the generated task instruction to the cloud server. The task instruction includes information indicating the location of the road section interval; a cloud server that receives the task instruction for the UAV sent by the remote monitoring device, parses the task instruction to generate a control instruction for the UAV to perform autonomous cruise, and according to the information indicating the location of the road section interval obtained by parsing, sends the control instruction to at least one UAV control device; at least one UAV control device that receives the control instruction from the cloud server, selects a UAV according to a predetermined rule, and sends the control instruction to the selected UAV; and at least one UAV that performs autonomous cruise based on the control instruction received from the at least one UAV control device.

[0033] The present invention is applied to the construction of a large-scale mineral database in the Qinghai-Tibet Plateau.

[0034] In an embodiment of the present invention, the UAV control device is arranged near the UAV hangar or is arranged as an independent base station in the monitored areas including service areas of expressways and areas prone to natural disasters. The UAV hangar is used to store at least one UAV. In an embodiment of the present invention, the UAV autonomous inspection system can be set for a predetermined range, and the UAV autonomous inspection system can monitor a preset area (such as a road) within this range. In the present invention, the size of the predetermined range is not limited, and it can be divided according to administrative regions or according to terrain, landform, etc. Among them, there can be multiple UAV hangars, which are respectively arranged in areas with dense monitored locations, such as service areas of expressways and areas prone to natural disasters (such as floods, mudslides, rolling stones, etc.), so as to cover the areas that need to be monitored. The UAV control device can be a transceiver device for wireless signals for UAV control. For example, it can be arranged near the UAV hangar or as an independent base station in an appropriate place so as to be able to cover the entire area that needs to be monitored. In the UAV control device, the identification information of at least one UAV is stored in association with its respective endurance mileage and model. When the UAV control device receives a control instruction to perform autonomous cruise from the cloud server, based on the information such as the endurance mileage and model of each stored UAV, it selects a UAV that can perform autonomous cruise and sends the control instruction to the selected UAV.

[0035] The mission instructions of existing drones for autonomous cruise on road traffic are issued by the staff of the traffic management center based on monitoring images, traffic accident alarm information, natural disaster alarm information, and traffic condition sharing information, etc. The data of monitoring images, traffic accident alarm information, natural disaster alarm information, and traffic condition sharing information, etc. come from multiple platforms or APPs, and the channels are scattered. The contribution degrees of various traffic conditions to the start of the drone's autonomous cruise are different. Therefore, it takes time to summarize the information and determine the necessity of starting the drone, and it is required that the staff have corresponding technical capabilities. Therefore, there are problems such as the staff missing information and not arranging the drone for autonomous cruise in time, or arranging the drone to cruise under unnecessary circumstances, resulting in low accuracy and efficiency of the start of the drone's autonomous inspection.

[0036] In one embodiment of the present invention, the remote monitoring device includes a communication unit, and this communication unit can be configured to connect to at least one big data platform through a communication network and receive information related to traffic conditions. The at least one big data platform includes at least one of a traffic accident alarm platform, a road traffic monitoring platform, an automobile navigation system platform, and a natural disaster rescue platform. In the present invention, the communication network can be either a wired network or a wireless network corresponding to mobile communication standards such as the fourth generation (4G), fifth generation (5G), and long term evolution (LTE).

[0037] In one embodiment of the present invention, as information related to traffic conditions, for example, the remote monitoring device receives traffic accident alarm information provided by the alarm person to the alarm platform through voice, pictures, etc. when a traffic accident occurs from the traffic accident alarm platform. Generally, the alarm person will describe the location information of the traffic accident and the severity of the traffic accident through voice, pictures, etc. The location where the traffic accident occurs can be described by, for example, the road name where the accident occurs, landmark buildings, traffic signal machines, etc. The severity of the traffic accident can include, for example, information such as whether a vehicle explosion or combustion occurs, and whether there are casualties.

[0038] In an embodiment of the present invention, for example, the remote monitoring device also receives information related to traffic conditions from the road traffic monitoring platform of the traffic management bureau. For example, the road traffic monitoring platform includes traffic information collection units arranged at the entrances and exits of each road section interval. The traffic information collection units are used to collect traffic flow information of each road section interval, and send the position information of the road section interval, the traffic flow information, and the information collection time information as the information related to traffic conditions to the remote monitoring device. Among them, the position information of each road section interval is represented by, for example, the longitude and latitude information of the interval entrance and exit. In an embodiment of the present invention, for example, the remote monitoring device determines whether the road section interval is in a congested state based on the number of vehicles entering and leaving each road section interval during a predetermined period. In an embodiment of the present invention, for example, the remote monitoring device is also connected to the vehicle navigation system platform through a communication network, and determines whether the road section interval is in a congested state based on the vehicle speeds of each road section interval fed back by the vehicle navigation system platform. In an embodiment of the present invention, for example, the remote monitoring device determines which level of high, medium, or low congestion level each road section interval is in based on the traffic flow information or vehicle speed information.

[0039] In an embodiment of the present invention, the remote monitoring device is also connected to the natural disaster rescue platform of the rescue center participating in the disaster relief work through a communication network. For example, in the case of road interruption caused by earthquakes, heavy rains, mudslides, etc., the remote monitoring device receives information related to natural disasters from the natural disaster rescue platform. Among them, the information related to the occurrence of natural disasters includes information related to the occurrence of disasters reported by the affected people through voice, pictures, etc. during the disaster. The information related to the occurrence of disasters includes the position information describing the occurrence of natural disasters and the severity of natural disasters. The occurrence location of natural disasters can be described by, for example, the road name where the natural disaster occurs, landmark buildings, landmark geographical landscapes, etc. The severity of natural disasters can include, for example, information related to whether landslides or mudslides occur, whether roads are washed out, whether bridges collapse, whether there are casualties, etc.

[0040] In an embodiment of the present invention, the remote monitoring device includes: a traffic information database, which stores the received information related to traffic conditions as traffic condition records according to road section intervals for each predetermined period. An item for recording the autonomous cruise warning value is set in the traffic condition records, and the autonomous cruise warning value is the cumulative value of the warning values corresponding to emergency situations; a task instruction generation unit, which generates a task instruction for autonomous cruise for the road section interval corresponding to the traffic condition record when the autonomous cruise warning value in any of the traffic condition records exceeds a preset threshold.

[0041] and a communication unit, which sends the task instruction to the cloud server.

[0042] The cloud server analyzes the task instruction to generate the control instruction, and according to the road section interval corresponding to the task instruction, sends the control instruction to the at least one drone control device;

[0043] The at least one drone control device selects at least one drone according to the endurance of the drone and sends the control instruction to it.

[0044] Preferably, in an embodiment of the invention, in the traffic information database, for each predetermined period, according to the position information of the starting points of each road section interval in the monitoring area, such as the longitude and latitude information of the starting points, the traffic condition information of each road section interval is stored as a traffic condition record. An option for an autonomous cruise warning value indicating the necessity of performing drone autonomous cruise is set in each traffic condition record, and the autonomous cruise warning value is the cumulative value of the warning values corresponding to the emergency situations of the cruise. That is to say, in the present invention, the autonomous cruise warning value represents the cumulative value of all values related to the demand for autonomous cruise in the road section interval corresponding to this record. Here, the values related to the demand for autonomous cruise include the values corresponding to the severity of traffic accidents, traffic jams, and natural disasters. In addition, whether it is a traffic accident, congestion, or natural disaster, only the alarm information at the moment when the event just occurs has practical significance for the start of drone autonomous cruise. Therefore, the above-mentioned predetermined period can be appropriately set considering storage costs, etc. For example, it can be set to any time length such as 10 minutes, 30 minutes, 1 hour, etc. The present invention does not limit the length of the predetermined period. When the predetermined period passes, the system resets the autonomous cruise warning value.

[0045] Hereinafter, the calculation method of the autonomous cruise warning value in the present invention will be described in detail. Those skilled in the art should know that the calculation method of the embodiment described here is only an example and is not used to limit the scope of protection of the invention. The present invention can also be implemented by equivalent variants of the embodiment.

[0046] In one embodiment of the present invention, in response to receiving traffic accident alarm information or traffic congestion information from at least one big data platform, the remote monitoring device is controlled to extract the description information of the accident occurrence location or traffic congestion location and the accident condition or congestion condition keywords from the traffic accident alarm information or traffic congestion information through any one of the pre-trained speech recognition neural network, text extraction neural network, and image recognition neural network. The accident occurrence location or traffic congestion location is located according to the description information of the accident occurrence location or traffic congestion location, relevant traffic condition records are searched according to the accident occurrence location or traffic congestion location, and the warning values weighted based on the accident condition keywords or congestion condition keywords are accumulated for the autonomous cruise warning values in the traffic condition records. Here, the speech recognition neural network, text extraction neural network, and image recognition neural network can be obtained through well-known training models, and the present invention does not limit this.

[0047] In one embodiment of the present invention, for example, when the extracted accident condition keywords include keywords indicating a relatively low severity of the accident, such as "minor scratch", "no personal injury", etc., or when the extracted congestion condition keywords include keywords indicating a low degree of road congestion, such as "slight congestion", a +1 process is performed on the autonomous cruise warning value; when the extracted accident condition keywords include keywords indicating a relatively high severity of the accident, such as "fire", "personal injury", etc., or when the extracted congestion condition keywords include keywords indicating a severe degree of road congestion, such as "completely blocked", "nowhere to go", etc., in order to quickly start the drone to start autonomous cruise to obtain more detailed information about the scene, an arbitrary value greater than 1 will be added to the autonomous cruise warning value. For example, it can be directly increased to the same value as the pre-set threshold.

[0048] In one embodiment of the present invention, in response to receiving natural disaster information from at least one big data platform, the remote monitoring device is controlled to extract the description information of the natural disaster occurrence location and the natural disaster condition keywords from the natural disaster information through any one of the speech recognition neural network, text extraction neural network, and image recognition neural network. The natural disaster occurrence location is located according to the description information of the natural disaster occurrence location, relevant traffic condition records are searched according to the natural disaster occurrence location, and the warning values corresponding to the natural disaster condition keywords are accumulated for the autonomous cruise warning values in the traffic condition records.

[0049] In an embodiment of the present invention, when the extracted natural disaster situation keywords include keywords indicating relatively minor natural disasters such as "light rain", "a little rockfall", "no casualties", etc., the autonomous cruise warning value is incremented by 1; when the extracted natural disaster situation keywords include keywords indicating relatively high severity of natural disasters such as "bridge collapsed", "road blocked", "huge rockfall", "mudslide", etc., in order to quickly start the drone to perform autonomous cruise to obtain more detailed information about the scene, an arbitrary value greater than 1 is added to the autonomous cruise warning value. For example, it can be directly increased to the same value as the pre-set threshold.

[0050] In an embodiment of the present invention, when the location information indicating the location of an accident or congestion received by the remote monitoring device from the road traffic monitoring platform is the road name, landmark building, traffic signal, etc. where the accident or congestion occurred, referring to the pre-stored map data, the received location information is converted into corresponding longitude and latitude information to search for the traffic condition records of the relevant road section interval based on the converted longitude and latitude information.

[0051] In an embodiment of the present invention, the remote monitoring device can generate a task instruction for the drone by any one of the following two methods:

[0052] (1) When the autonomous cruise warning value in any traffic condition record exceeds the pre-set threshold, a task instruction for autonomous cruise of the road section interval corresponding to this traffic condition record is generated, and the task instruction is sent to the cloud server;

[0053] (2) According to the monitoring area set by the staff operating the remote monitoring device, a task instruction for autonomous cruise corresponding to the monitoring area is generated, and the task instruction is sent to the cloud server.

[0054] In the case of (1), the generated task instruction for autonomous cruise can also be displayed on the display of the remote monitoring device, and when a confirmation instruction from the staff is received, the task instruction is sent to the cloud server. A predetermined time threshold can also be set, and when the task instruction is displayed on the remote monitoring device and the time of this time threshold has passed, the task instruction is sent to the cloud server.

[0055] In one embodiment of the present invention, at least one drone control device is stored in a cloud server in correspondence with the location information of a drone hangar near the drone control device. The cloud server parses the task instruction received from the remote monitoring device, and based on the information obtained from the parsing indicating the location of the road section where the drone rental cruise needs to be executed, the generated control instruction is sent to at least one drone control device corresponding to the drone hangar set in the road section. In the case where a drone hangar is not set for each road section, the control instruction can also be sent to the drone control device corresponding to the drone hangar set in the road section adjacent to the road section or the road section closest to the road section.

[0056] In one embodiment of the present invention, the staff can also select a road section for inspection through the remote monitoring device, and can set multiple inspection points in the map of the section (for example, sections with frequent accidents, sections with steep roads, sections with complex road conditions, etc.). The remote monitoring device sends inspection instructions to the cloud server, and the cloud server can select a suitable drone according to predetermined rules (for example, a drone on standby in the nearest hangar, or a drone returning nearby).

[0057] The cloud server generates an inspection map based on existing map data and three-dimensional models, and plans the inspection routes, inspection points, and the position of the drone when the inspection point locations are located during the process of constructing the inspection map;

[0058] Among them, the three-dimensional model is established through drone scanning. During the scanning process, the map features are obtained through the laser radar carried by the drone to ensure that the coordinates of each object in the three-dimensional model are consistent with the real world. The three-dimensional model and the inspection map share the coordinate system to mark or identify objects that are not reflected in the existing map.

[0059] In this embodiment, since the existing map data may be updated in a delayed or inaccurate manner, the present invention uses a drone to scan to pre-establish a local three-dimensional model, uses a drone control device to control the drone to scan the road network in a predetermined area without blind spots, obtains map features through the laser radar carried by the drone during the scanning process, and establishes a three-dimensional model of the area near the road network on the cloud server to ensure that the coordinates of each object in the three-dimensional model are consistent with the real world, and the three-dimensional model and the inspection map share the coordinate system. By pre-establishing a three-dimensional model, objects that are not reflected in the existing map, such as low-altitude obstacles such as wires and poles, woods, and signboards, can be marked or identified, and a more accurate and safe inspection route can be obtained to avoid collisions with obstacles when the drone is required to perform low-altitude monitoring. That is, by establishing and updating the three-dimensional model, the area where the drone can fly can be determined, and the drone can be autonomously approached when necessary to obtain more accurate and clear images or information.

[0060] It is also possible to update the 3D model in real time according to the map features obtained by the drone during daily inspections, so as to further avoid collisions between the drone and obstacles during low-altitude monitoring.

[0061] After the drone reaches the inspection point, it can stay at the inspection point temporarily and notify the operator. In the case where the operator needs to manually control the drone, switch to the manual control mode. The operator can set multiple inspection points in this small area through the remote monitoring device to conduct more detailed inspections or switch to the manual mode to control the drone, so as to check specific targets or target areas as desired.

[0062] After the inspection is completed, the drone can fly back to the drone hangar autonomously along the original path, or perform the next task according to further instructions.

[0063] The cloud server also generates an inspection map based on the map data and the 3D model, and plans the inspection route, inspection points, and the pose of the drone at the inspection point positions during the construction of the inspection map;

[0064] Among them, the 3D model is established by drone scanning. During the scanning process, the lidar carried on the drone is used to obtain map features to ensure that the coordinates of each object in the 3D model are consistent with the real world. The 3D model and the inspection map share the coordinate system, which is used to mark or identify objects not reflected in the existing map.

[0065] The cloud server also performs object recognition on the objects according to the images received from the at least one drone, confirms each object and its state at the inspection point, and displays them to the operator in real time through the remote monitoring device.

[0066] The cloud server also performs object recognition on the objects according to the received images, confirms each object and its state at the inspection point, and displays them to the operator in real time through the remote monitoring device, and issues early warnings. The method for issuing early warnings is as follows:

[0067] In a scenario of this embodiment, after a long period of rainfall in a mountainous section, the risk of rockfall is estimated to be relatively high. Therefore, the operator can set up multiple (e.g., 20) inspection points for this mountainous section with a relatively high risk (e.g., 10 kilometers) and conduct inspections at a predetermined time interval (e.g., once every 2 hours). During the inspection process, the cloud server performs target recognition on the received images and can detect information such as whether there are fallen rocks in the road and its surrounding areas, the size and location of the fallen rocks, whether there are stopped cars, and the license plate numbers of the vehicles. After detecting the above situations, the cloud server sends this information to the remote monitoring device. After receiving this information, the remote monitoring device can send a reminder or an alarm to the operator so that the operator can take countermeasures such as closing the road or providing rescue.

[0068] In a scenario of this embodiment, the drone can be set to fly along a certain section of the road to find vehicles with speeds lower than the threshold, locate the positions of vehicles where accidents may occur, collect relevant information about the vehicles, and transmit back images to determine whether it is necessary to dispatch assistance and conduct traffic control, etc.

[0069] In a scenario of this embodiment, during the passage of a flood peak, there may be risks such as piping or levee breaches. Therefore, the operator can set up multiple inspection points for this riverbank area with a relatively high risk and conduct inspections at a predetermined time interval (e.g., once every half hour). During the inspection process, the cloud server performs target recognition on the received riverbank images and detects whether there are situations such as piping and levee breaches in the riverbank and its surrounding areas and their severity. After detecting the above situations, the cloud server sends this information to the remote monitoring device. After receiving this information, the remote monitoring device can send a reminder or an alarm to the operator so that the operator can take corresponding countermeasures.

[0070] The drone is equipped with a radar ranging sensor, an infrared sensor, and a high-definition image sensor. Based on the equipped sensors, the drone can conduct close-range observations based on the data of the three-dimensional model to obtain clearer and more accurate data.

[0071] The beneficial effects of the present invention are as follows: The present invention provides an unmanned aerial vehicle (UAV) autonomous inspection system. By recording traffic conditions according to road sections, counting information related to traffic conditions received from multiple big data platforms for each road section, setting the cumulative value of the added autonomous cruise warning values corresponding to emergency situations according to the urgency of various traffic conditions, and automatically generating a task instruction indicating that the UAV starts to inspect the road section when the autonomous cruise warning value corresponding to any road section exceeds a preset threshold. Therefore, it can accurately count data from multiple big data platforms, thereby preventing the situation where the staff fails to arrange the UAV for autonomous cruise in a timely manner due to missing information or arranges the UAV to cruise under unnecessary circumstances, thus improving the accuracy of the UAV autonomous inspection startup and the system operation efficiency.

[0072] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as "center", "thickness", "upper", "lower", "horizontal", "top", "bottom", "inner", "outer", "radial", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of technical features. Therefore, the features defined by "first", "second", "third" may explicitly or implicitly include one or more of such features.

Claims

1. An autonomous inspection system for unmanned aerial vehicles, characterized in that: include: The remote monitoring device receives and stores information related to traffic conditions of each road section from at least one big data platform, generates a task instruction for the drone to perform autonomous cruising on the corresponding road section based on the received information related to traffic conditions, and sends the generated task instruction to a cloud server, wherein the task instruction includes information indicating the location of the road section; The cloud server receives the task instruction for the UAV sent by the remote monitoring device, analyzes the task instruction to generate a control instruction for the UAV to perform autonomous cruising, and sends the control instruction to at least one UAV control device according to the information indicating the position of the road section obtained by the analysis; at least one drone control device receives the control instruction from the cloud server, selects a drone according to a predetermined rule, and sends the control instruction to the selected drone; and at least one UAV that performs autonomous cruising based on the control instructions received from the at least one UAV control device; The remote monitoring device comprises: a traffic information database, for each predetermined period, storing the received information related to traffic conditions according to road sections as traffic condition records, wherein the traffic condition records are provided with an item for recording an autonomous cruise warning value, wherein the autonomous cruise warning value is an accumulated value of warning values ​​corresponding to emergency conditions; a task instruction generating unit, which generates a task instruction for autonomous cruising for a road section corresponding to the traffic condition record when the autonomous cruise warning value recorded in any of the traffic condition records exceeds a preset threshold; and a communication unit, for sending the task instruction to the cloud server; The calculation method of the autonomous cruise warning value is specifically as follows: In response to receiving traffic accident alarm information or traffic congestion information from the at least one big data platform, the remote monitoring device extracts description information of the accident location or traffic congestion location and accident condition or congestion condition keywords from the traffic accident alarm information or traffic congestion information through any one of a pre-trained speech recognition neural network, a text extraction neural network, and an image recognition neural network, locates the accident location or traffic congestion location according to the description information of the accident location or traffic congestion location, searches for the relevant traffic road condition records according to the accident location or traffic congestion location, and accumulates the warning values ​​corresponding to the accident condition keywords or the congestion condition keywords for the autonomous cruise warning values ​​in the traffic road condition records; In response to receiving natural disaster information from the at least one big data platform, the remote monitoring device extracts descriptive information of the location where the natural disaster occurred and natural disaster condition keywords from the natural disaster information through any one of the speech recognition neural network, the text extraction neural network, and the image recognition neural network, locates the location where the natural disaster occurred according to the descriptive information of the location where the natural disaster occurred, searches for the related traffic condition records according to the location where the natural disaster occurred, and accumulates the autonomous cruise warning values ​​in the traffic condition records that correspond to the natural disaster condition keywords.

2. The autonomous inspection system of unmanned aerial vehicle according to claim 1, characterized in that: The at least one big data platform includes at least one of a traffic accident alarm platform, a road traffic monitoring platform, a car navigation system platform, and a natural disaster rescue platform.

3. The autonomous inspection system of unmanned aerial vehicle according to claim 1, characterized in that: The cloud server parses the task instruction to generate the control instruction, and sends the control instruction to the at least one drone control device according to the road section corresponding to the task instruction; The at least one drone control device selects at least one drone according to the flight range of the drones and sends the control instruction to the drone.

4. The autonomous inspection system of unmanned aerial vehicle according to claim 1, characterized in that: The drone control device is arranged near the drone hangar, or as an independent base station in the area to be monitored including the service area of ​​the expressway and the area prone to natural disasters; The drone hangar is used to store the at least one drone.

5. The autonomous inspection system of unmanned aerial vehicle according to claim 2, characterized in that: The road traffic monitoring platform includes traffic information collection units arranged at the entrance and exit of each road section; The traffic information collection unit is used to collect traffic flow information of each road section, and send the location information, traffic flow information and information collection time information of the road section as the information related to the traffic conditions to the remote monitoring device; The remote monitoring device saves the information related to the traffic conditions into the traffic condition record of each road section according to the position information of the road section.

6. The autonomous inspection system of unmanned aerial vehicle according to claim 1, characterized in that: The cloud server also generates an inspection map based on the map data and the three-dimensional model, and plans the inspection route, inspection points, and the position of the drone when the inspection point location is constructed; Among them, the three-dimensional model is established through drone scanning. During the scanning process, the map features are obtained through the laser radar carried by the drone to ensure that the coordinates of each object in the three-dimensional model are consistent with the real world. The three-dimensional model and the inspection map share the coordinate system to mark or identify objects that are not reflected in the existing map.

7. The UAV autonomous inspection system according to claim 1, characterized in that: The cloud server also performs target recognition on objects based on the images received from the at least one drone, confirms each object and its status at the inspection point, and displays it to the operator in real time through the remote monitoring device.

Images