WO2018195955A1 - 一种基于飞行器的设施检测方法及控制设备 - Google Patents

一种基于飞行器的设施检测方法及控制设备 Download PDF

Info

Publication number
WO2018195955A1
WO2018195955A1 PCT/CN2017/082501 CN2017082501W WO2018195955A1 WO 2018195955 A1 WO2018195955 A1 WO 2018195955A1 CN 2017082501 W CN2017082501 W CN 2017082501W WO 2018195955 A1 WO2018195955 A1 WO 2018195955A1
Authority
WO
WIPO (PCT)
Prior art keywords
flight
detection
aircraft
image
target facility
Prior art date
Application number
PCT/CN2017/082501
Other languages
English (en)
French (fr)
Inventor
李思晋
赵丛
封旭阳
张李亮
Original Assignee
深圳市大疆创新科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市大疆创新科技有限公司 filed Critical 深圳市大疆创新科技有限公司
Priority to CN202111071109.0A priority Critical patent/CN113791641A/zh
Priority to CN201780004504.2A priority patent/CN108496129B/zh
Priority to PCT/CN2017/082501 priority patent/WO2018195955A1/zh
Publication of WO2018195955A1 publication Critical patent/WO2018195955A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0094—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots involving pointing a payload, e.g. camera, weapon, sensor, towards a fixed or moving target
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/10—Simultaneous control of position or course in three dimensions
    • G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
    • G05D1/106—Change initiated in response to external conditions, e.g. avoidance of elevated terrain or of no-fly zones

Definitions

  • the present invention relates to the field of computer control technologies, and in particular, to an aircraft-based facility detection method and control device.
  • Embodiments of the present invention provide an aircraft-based facility detection method and control device, and implement inspection of a target facility by controlling the aircraft.
  • an embodiment of the present invention provides an aircraft-based facility detection method, including:
  • the aircraft flight is controlled in accordance with the flight rules to facilitate completion of detection of the target facility.
  • an embodiment of the present invention further provides a control device, including: a processor and a data interface;
  • the data interface is configured to exchange data with an aircraft
  • the processor configured to acquire an environment image including the target facility when the aircraft is located at a detection location for the target facility; and determine an image region to which the target facility belongs from the environment image a domain, and performing image segmentation on the image region to obtain a detection target for the target facility; acquiring a flight rule regarding the detection object according to an image location of the detection object in the environment image and the detection location;
  • the flight rules generate control commands and are transmitted to the aircraft through the data interface to control flight of the aircraft to facilitate completion of detection of the target facility.
  • Embodiments of the present invention when inspecting certain facilities, particularly for facilities with higher facilities or areas that are not easily accessible, can detect one or more of the facilities based on image recognition and automatic control of flight.
  • the object is inspected, which reduces the labor cost and safety of the inspection and improves the efficiency of the inspection.
  • FIG. 1 is a schematic flow chart of a method for detecting a target facility according to an embodiment of the present invention
  • FIG. 2 is a schematic flow chart of an aircraft-based facility detecting method according to an embodiment of the present invention
  • FIG. 3 is a schematic flow chart of another aircraft-based facility detecting method according to an embodiment of the present invention.
  • FIG. 4 is a schematic flow chart of a method for determining flight rules according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of a facility detecting apparatus according to an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a control device according to an embodiment of the present invention.
  • the embodiment of the present invention can detect and identify the target facilities that need to be detected at a distance, and automatically fly to the vicinity of the target facility to be detected. And through the image segmentation recognition technology, the parts in the target facility are divided and divided, and one or more detection objects of the target facility are obtained, and the carried sensors (such as cameras, thermal imagers, etc.) are used to identify the targets. Each test object is detected and recorded. After the test record is over, the aircraft such as the drone will automatically return to the aircraft or go to the next target facility to be tested in the vicinity of the power supply.
  • the carried sensors such as cameras, thermal imagers, etc.
  • FIG. 1 it is a schematic flowchart of a method for detecting a target facility according to an embodiment of the present invention.
  • the facility detection method of the embodiment of the present invention may be performed by a control device that can be configured on an aircraft. Further, in the embodiment of the present invention, the method will be described using an unmanned aircraft as an aircraft on which a sensor for detecting a target facility is mounted.
  • the main steps performed by the control device are as follows.
  • S101 Determine the location of the target facility.
  • GPS Global Positioning System, Global Positioning System (GPS) information, such as information, defines the location of the target facility, controls the drone to fly to the target facility autonomously, and the target facility appears within the scope of the drone.
  • the scope of the observation mainly refers to the drone.
  • the detection range of the mounted sensor such as the camera's shooting range.
  • the user may input location information of one or more target facilities on the user interface configured for the control device in advance, for example, in a interface displaying the map, by clicking and clicking, specifying one on the interface including the map. Or multiple location points, the control device can record these location points as location points of the target facility.
  • the control device can control the drone to fly based on the location point of each target facility when the drone is turned on the patrol mode, so as to fly to the corresponding one or more target facilities to monitor a target facility, or monitor a certain piece.
  • Multiple target facilities on the route for example, multiple connected towers.
  • S102 Perform obstacle avoidance processing based on image segmentation recognition and depth map.
  • active obstacle avoidance can be performed based on the image segmentation recognition and the acquired depth map to facilitate safe flight to an area where the target facility can be detected.
  • the depth map can be detected and calculated based on binocular vision detection to locate obstacles on the flight path.
  • the acquisition of the depth map can be obtained by binocular calculation matching, or can be calculated by a device based on structured light or infrared.
  • a device based on structured light or infrared can obtain a depth map with a higher relative quality.
  • the image segmentation technology can be further combined to determine and fly the obstacle. Avoidance. Since image segmentation does not need to be matched, it can also have a better recognition effect for areas with less texture, so it can be used together with the depth map. On the one hand, a better depth map can be obtained. On the other hand, a depth map can also be given. Each point in the field gives practical meaning, which helps the drone to carry out path planning and determine the flight rules of the drone flight.
  • the drone continuously estimates and corrects the relative position to the target facility.
  • the drone can use a visual tracking algorithm to lock the target facility within an image observable range, estimating the relative distance to the target facility from the size of the target facility in the image and the current flight speed.
  • the drone can also fly according to a specific trajectory to obtain the depth information of the object in the scene, and provide a reference for distance avoidance.
  • the patrol scene of the current detection may be determined based on the captured image identification, and the patrol scene may be specifically classified according to the target facility detected this time, for example, including: The scene of the inspection tower, the scene of the inspection bridge, etc., the determined inspection scene can provide reference information for flight obstacle avoidance and planned flight route.
  • the target facility detected this time for example, including: The scene of the inspection tower, the scene of the inspection bridge, etc., the determined inspection scene can provide reference information for flight obstacle avoidance and planned flight route.
  • the patrol scene is a scene of a patrol tower, since power lines are generally connected between adjacent towers in these scenarios, the connected parts are relatively fixed, and when the drone is setting a route, You can choose to bypass areas where power lines are dense.
  • the drone can be equipped with corresponding sensors to further improve the reliability of obstacle avoidance.
  • thermal imagers can be used to detect the presence of wires, providing a more robust obstacle avoidance for inspections that require close proximity to the power line.
  • Embodiments of the present invention may detect a target facility within an image range based on a visual detection method of an artificial feature. It is also possible to learn more massive and reliable image features from the data by learning the massive image data of the target facilities such as the electric tower based on the recognition algorithm of the deep neural network, thereby obtaining more accurate recognition results.
  • the detection algorithm runs on the image observed by the drone, and detects and locates the location of the target facility in the image. Once the target facility to be detected is found in the image, it is locked in the image and gradually flies to the target facility to be detected.
  • the tracking target algorithm can be used to lock the detected target facility, and the detected result is used as a reference to determine the flight path of the drone.
  • multiple feature points can be selected in the target facility.
  • the segmentation recognition based on the full image can enable the drone to select more stable image feature points according to the category of the target facility.
  • These image feature points generally need to be stably present on the target facility, do not move, and are easily detected and found.
  • the feature points selected on the electric tower are more stable than the feature points on the water surface.
  • the accuracy of SLAM Simultaneous localization and mapping
  • the drone can adjust the posture and route more flexibly.
  • the target facility to be detected may be selected that in the local path, the target facility to be detected does not appear within the field of view, and after passing the obstacle, the target facility to be detected is re-locked to the image according to the estimated relative position. in.
  • S104 Identify various components of the facility and perform targeted detection and recording.
  • the detection position of the target facility is reached, for example, when the distance from the target facility is somewhere in the area within the preset distance range, the detection object may be further identified from the target facility based on the image, for example, When the target facility is an electric tower, it is identified that the detection object to be detected this time is the entire tower head, or the components of the fixed power line.
  • the image segmentation algorithm will provide pixel level recognition and segmentation, and provide category information of each pixel in the image.
  • the role of the category information here mainly determines the detection object corresponding to the category information, and then determines the patrol strategy to be adopted. Direct the drone to fly.
  • the target facility to be detected may be segmented from the image, and a partial image region including only the target facility is determined, and then the partial image region is analyzed and identified, and different categories are identified for different locations of the target facility, and this time is obtained.
  • the key part that needs to be detected is the detection object, so that the key parts of the target facility can be detected and recorded in a targeted manner.
  • the object model for segmentation and recognition of the specific part of the target facility may be used.
  • the object model can also segment and identify the components of the target facility to obtain one or more detection objects.
  • each component of the target facility such as an electric tower can be identified, and the position of each component is marked in the image.
  • the user can specify a patrol policy for each component in advance.
  • the patrol strategy includes, but is not limited to, surround shooting, near-continuous video shooting, and fixed-point photography using a high-precision camera.
  • the input of the object model is an image, outputting pixel-level recognition results, representing a particular category to which each pixel belongs, may include background and various components in the refined target facility That is to say, each pixel can be a pixel of the background category or a pixel of a certain detection object category (for example, a tower head).
  • the object model can be configured according to the target facility that needs to be inspected. For example, for the electric tower, an object model such as a tower head or a tower foot can be configured to identify the tower head and the tower foot of the target facility in the image.
  • a specific inspection plan is generated based on the recognition result and the user's settings.
  • the generation of the inspection plan includes: generation of tracks, time allocation of each track, and the like.
  • the criteria for solving the trajectory will include information on how to reduce the flight time and how to choose a safer flight route, such as how to avoid the power line, and how to make the picture stable and reliable.
  • the characteristics of the drone such as the maximum and minimum acceleration of the drone, and the special effects of the detection device that performs the detection, such as the FOV of the camera, the optimal use distance of other sensors for detection, and the like.
  • the drone After the trajectory generation is completed, the drone performs the calculated trajectory, and patrols each component to be detected (detection object) of the target facility. At this point, the drone will continue to update the observations, and the real-time dynamic correction of the trajectory to ensure safe and effective inspection.
  • the aircraft automatically returns.
  • the drone can be executed according to the start
  • the departure point information recorded during the inspection mission and/or the recorded flight data are used to achieve automatic return.
  • the visual odometry and GPS information can be used to direct the drone to return.
  • the visual odometer estimates the trajectory of the drone in the execution of the mission by means of image feature matching. A visual odometer combined with an image segmentation algorithm can select better matching features for more accurate trajectory estimation.
  • the optimal return route to the takeoff point can be calculated by analyzing the flight trajectory when the mission is performed. For example, the flight trajectory of exploration, attempt, and repetition in the execution of a mission can be bypassed.
  • the flight trajectory that can be bypassed may be a partial trajectory marked in the flight trajectory obtained by implementing the visual odometer described above.
  • the flight trajectory estimated for return flight may also be corrected in conjunction with GPS coordinates.
  • the positioning sensor can be further used to achieve more accurate return after the drone reaches the vicinity of the takeoff point.
  • the drone can determine the best return route based on the visual odometer generated by the previous record, and at the same time turn on the obstacle avoidance function during the return flight.
  • the known 2D/3D map can be combined to determine obstacles on the flight path, such as determining obstacles such as buildings and mountains that have been identified on the map, and thus, when determining the flight path. Choose to bypass these obstacles.
  • a mark for identifying the detection object may be set on the target facility, and based on the mark and the captured image, the image area of the target facility is quickly segmented and positioned.
  • One or more test objects When acquiring the depth map, it can be acquired not only based on binocular vision, but also using a device like a laser radar.
  • Embodiments of the present invention when inspecting certain facilities, particularly for facilities with higher facilities or areas that are not easily accessible, can detect one or more of the facilities based on image recognition and automatic control of flight.
  • the object is inspected, which reduces the labor cost and safety of the inspection and improves the efficiency of the inspection.
  • FIG. 2 it is a schematic flowchart of an aircraft-based facility detection method according to an embodiment of the present invention.
  • the method in the embodiment of the present invention may be performed by a dedicated control device, and the control device may be configured in an unmanned manner. Machines and other aircraft.
  • the control device can also be used as a ground-end device to wirelessly interact with an aircraft such as a drone to complete the inspection of the target facility.
  • S201 Acquire an environmental image including the target facility when the aircraft is located at a detection location for the target facility.
  • the function of the detection location is mainly that the relevant processing of the target facility can be triggered to complete the inspection task for the target facility.
  • the detection position may refer to a position point located in a certain location area, and the location area may refer to an area within a preset distance range from the target facility.
  • the control device determines whether the aircraft has reached a detection location for the target facility based on the detected location of the aircraft (eg, GPS location coordinates) or the location reported by the aircraft, and based on the location of the target facility.
  • Whether or not the detection location is reached may also be determined by the aircraft.
  • the aircraft may perform analysis based on the captured image including the target facility, based on the actual size of the target facility, the size of the target facility in the image, and the target. The location of the facility in the image is used to estimate the distance between the aircraft and the target facility. If the distance is within a predetermined distance range, the aircraft can be considered to have detected the location of the target facility.
  • the detection location may also be a specific location, the image obtained at the location includes the target facility, or the target facility included in the image can be segmented at the location, for example, if the image is in the target facility When the occupied area satisfies the preset condition (the number of pixel points of the target facility is greater than a preset threshold, or the area of the image area occupied by the target facility is greater than a preset threshold), the position where the aircraft is located may be regarded as the detected position.
  • the preset condition the number of pixel points of the target facility is greater than a preset threshold, or the area of the image area occupied by the target facility is greater than a preset threshold
  • a camera such as a drone is equipped with a camera such as a camera.
  • the camera is triggered to capture an environment image.
  • the detection object is mainly used to determine the detection object, and the detection object is determined.
  • the flight rules used at the time are used at the time.
  • one or more facility location points to be detected may be configured on the interactive interface displaying the map; the facility corresponding to the selected one or more facility location points is determined as the target facility; according to the selected facility The location point controls the aircraft to fly to the target facility to facilitate flight to the detection location for the target facility. If the user selects multiple target facilities on the interactive interface, the aircraft can be controlled to perform the following steps to complete the inspection of multiple target facilities from near to near or from near to far. Or, based on the remaining energy of the aircraft, complete inspection of one or a part of the target facilities. If the control device is a smart terminal including a display, such as a smart phone, a tablet, etc., an interactive interface including a map can be directly displayed to the user.
  • control device If the control device is mounted on the aircraft, the control device can send an instruction to trigger the detection interface of the data for receiving the aircraft to display the interaction boundary through the built-in wireless communication interface or through the wireless communication interface on the aircraft. And receiving the location of the target facility determined on the interactive interface to control aircraft flight.
  • S202 Determine an image region to which the target facility belongs from the environment image, and perform image segmentation on the image region to obtain a detection target for the target facility. Determining the image area to which the target facility belongs may also be implemented based on image segmentation. The image segmentation may be performed based on the brightness and color of the pixels in the environment image, and the image regions to which the target facility belongs and the respective detection objects of the target facility in the environment image are obtained.
  • the detection object may be all of the target facilities, and the detection object is also a part of the target facility, such as a tower head of an electric tower, a fixed component of a fixed power line, and the like.
  • the detection object is specified in advance by the user. For example, if the user specifies that the entire electric tower needs to be inspected, then in S202, the entire electric tower in the image area can be used as the detection target after the image area of the electric tower is obtained.
  • the user can also specify the tower head portion of the inspection tower, and after obtaining the image area of the electric tower, the tower head is divided to obtain the detection target.
  • the image region is mainly segmented based on an object model preset for the target facility, and one or more detection objects are determined from a target facility of the image region.
  • the image segmentation is performed on the image region, and obtaining the detection object about the target facility may include: acquiring an object model preset for the target facility; and performing image segmentation on the image region according to the object model. And obtaining a detection object that satisfies the similarity condition with a shape similarity between the object models.
  • the object model is primarily used to identify a component on the target facility, for example, a predetermined object model for the tower head can assist in identifying the tower head portion of the tower.
  • a plurality of object models of different angles may be preset to correspond to one detection object, so that when the images about the target facility acquired at different angles are accurately segmented, the detection objects in the target facility can be accurately segmented.
  • the object model is configured with a model identifier, and the patrol policy associated with the detected object is obtained according to the model identifier.
  • the model identifier can be a name, for example, the model identifier of the object model for the tower head described above is “tower head”. After identifying a detection object based on the object model, the identifier of the detection object corresponds to the object model, and the identifier of the detection object may be the same as the model identifier of the object model.
  • a patrol policy associated with the identifier of the detection object may be determined from a preset mapping relationship library, and the patrol policy is mainly used to indicate how to patrol the detection object, including Surrounded by flying surround shooting, continuous video shooting, and the use of high-precision camera to take pictures.
  • the flight rule regarding the detection object may be further acquired based on the inspection rule.
  • S203 Acquire a flight rule regarding the detection object according to an image position of the detection object in the environment image and the detection position.
  • the image location may be a pixel location of the detection object in the image, and the orientation of the detection object relative to the aircraft may be determined based on the image location.
  • a flight rule capable of detecting the detection object from an upper direction, a lower direction, or the like may be determined, or a flight rule surrounding the detection object may be determined.
  • the flight rule may be a flight trajectory, for example, controlling a flight trajectory of a drone's flight, based on the flight trajectory,
  • S204 Control the aircraft flight according to the flight rule to facilitate completion of detection of the target facility. After the flight rules are determined, the aircraft is controlled to fly according to the flight rules, and the inspection of the inspection object can be completed.
  • the location information returned by the aircraft may be received in real time or periodically during the process of flying to the target facility to reach the detection location before the S201, or during the control of the flight of the aircraft, the location information includes: Distance information and direction information of the aircraft generated by the aircraft relative to a target object, or position coordinate information of the aircraft returned by the aircraft.
  • a relative position between the aircraft and the target facility is displayed in real time on the interactive interface based on the received location information and the location of the target facility.
  • Embodiments of the present invention when inspecting certain facilities, particularly for facilities with higher facilities or areas that are not easily accessible, can detect one or more of the facilities based on image recognition and automatic control of flight.
  • the object is inspected, which reduces the labor cost and safety of the inspection and improves the efficiency of the inspection.
  • FIG. 3 it is a schematic flowchart of another aircraft-based facility detection method according to an embodiment of the present invention.
  • the method in the embodiment of the present invention may be performed by a dedicated control device, and the control device may be configured in none.
  • the control device can also be used as a ground-end device to wirelessly interact with an aircraft such as a drone to complete the inspection of the target facility.
  • S301 Acquire an environmental image including the target facility when the aircraft is located at a detection location for the target facility.
  • the detection location is one of a location within a predetermined distance from the target facility.
  • the environmental image is an environmental image collected by a sensor such as a camera carried on the aircraft.
  • S302 Determine an image region to which the target facility belongs from the environment image, and perform image segmentation on the image region to obtain a detection target for the target facility. Image segmentation based The segmentation determines the image region, and analyzes the detection object that determines the target facility based on the preset object model.
  • S303 Acquire a flight rule regarding the detection object according to an image position of the detection object in the environment image and the detection position.
  • the role of the image position is to ensure that the detected object is always in the image. Determining a relative direction of the detection object according to the image position, and further, when detecting the detection object, if it is necessary to keep the detection object currently required to be detected in the middle of the screen, when generating a flight rule including a flight trajectory, Consider the image location. When only one detection object to be detected is included, only the flight rule for the detection object needs to be generated, for example, a flight trajectory surrounding the detection object is generated. If the detection object includes more than one, it is necessary to generate a flight rule, and based on the flight rule, multiple detection objects can be detected in sequence.
  • a flight rule may be generated, which starts from the detection position and first follows the flight in the flight rule. The trajectory goes up to the detection tower head, then flies down the flight trajectory to detect the tower body, and finally along the trajectory to the tower foot, thereby completing the "tower head", “tower body”, “tower” on multiple flight trajectories Foot” detection of three test objects.
  • the flight rule includes a flight trajectory
  • the S303 may specifically include: acquiring a patrol policy associated with the detection object; generating, according to an image position of the detection object in the environment image and the detection position Satisfy the flight trajectory of the inspection strategy.
  • a flight trajectory can be generated directly based on the inspection strategy corresponding to the detection object.
  • the detection object is in the middle position of the image collected by the aircraft.
  • the detection position may be generated as a starting point to generate a starting point.
  • a linear trajectory of the location of the target facility so that the aircraft can continuously capture the detection object of the target facility from far and near.
  • each of the detection objects is associated with a patrol inspection policy
  • the S303 may specifically include: acquiring a patrol inspection policy of each detection object; and according to each detection object in the environment image The image position and the detection position generate a flight rule; wherein the flight rule includes a flight trajectory that satisfies all the patrol strategies, or includes a plurality of flight trajectories, and each flight trajectory satisfies a part of the patrol strategy.
  • the limiting condition includes: a flight parameter based on a flight parameter and a detection parameter, the flight parameter includes a flight distance parameter, a flight duration parameter, a flight safety parameter, an energy loss parameter, Any one or more of the flight speed parameters.
  • the flight distance parameter is configured to a 1st effective value
  • the total length of the flight trajectory is further required to be the shortest, so that the flight distance of the aircraft The shortest to save energy.
  • the flight duration parameter is configured as a 1st effective value, it indicates that when generating a flight trajectory that satisfies one or more patrol strategies, the aircraft is further required to fly at a preset speed according to the adopted flight trajectory. The shortest time is to improve the efficiency of inspection.
  • the flight safety parameters are configured as 1 effective value, it indicates that the safe flight trajectory is preferentially selected, and some trajectories that may have obstacles are excluded, for example, when the electric tower is used as the target facility, the trajectory that may pass through the electric wire is excluded.
  • the energy loss parameter is configured as an effective value of 1 or so, it indicates that the trajectory with low energy consumption is preferentially selected as the final flight trajectory.
  • the method further includes: generating a detection parameter, the detection parameter is used to indicate a sensing parameter that the aircraft inspects the detection object during the control of the flight of the aircraft, and the sensing parameter includes: A shooting angle parameter of a sensor for detecting a detection target, and a shooting parameter of a camera for capturing an object to be detected.
  • the detection parameter may specifically be a parameter for controlling a pan/tilt angle, a parameter for controlling a focal length, a white balance, a shutter, and the like of the camera.
  • S305 Acquire a detected detection image during the process of controlling the flight of the aircraft according to the flight rule.
  • the process of controlling the flight of the aircraft according to the flight rule is a process of patrolling the detected object, and the captured image or the video generated according to the image may be immediately transmitted to the patrol user, and the patrol user generates the image by viewing the image or generating the image according to the image. Video to determine if one or more detected objects are normal.
  • the control device can further perform an analysis to determine the position of the currently detected object in the detected image. For example, when detecting the tower head of the electric tower, it is analyzed to determine the position of the tower head as the detection object in the detected image. Similarly, the image region in which the currently detected detection object is located may be determined according to the image segmentation technique, and the pixel position where the detection object is located may be further determined.
  • S306 Update the flight rule according to the position of the detection object in the detection image and the inspection strategy set for the detection object.
  • the update flight rule is mainly to ensure that the detection object that is currently required to be detected can be detected, for example, it is necessary to ensure that the position of the detection object is in the image center area of the detection image.
  • the S306 may be performed, and/or the update adjustment of the above-mentioned sensing parameters may be performed. Or adjust the sensing parameters first. If the preset detection requirements cannot be met, for example, the detection object cannot be guaranteed. The location is in the image center area of the detected image, and then the S306 is executed to update the flight rule (which may further be combined with the update adjustment of the sensing parameter) to meet the preset detection requirement.
  • S307 Control the flight of the aircraft according to the updated flight rule, so as to complete the detection of the target facility. That is to say, the detection object is continuously detected to obtain a corresponding detection image or a video generated based on the image, and is returned to the inspection user for viewing.
  • S308 After completing the detecting of the target facility, controlling the return of the aircraft according to a preset return trajectory; the preset return trajectory includes: recorded flight trajectory before the aircraft flies to the detection position .
  • Embodiments of the present invention when inspecting certain facilities, particularly for facilities with higher facilities or areas that are not easily accessible, can detect one or more of the facilities based on image recognition and automatic control of flight.
  • the object is inspected, which reduces the labor cost and safety of the inspection and improves the efficiency of the inspection. Moreover, it can automatically return to the air and avoid obstacles, further satisfying the automation and intelligent requirements of the inspection, and improving the safety of the inspection.
  • FIG. 4 it is a schematic flowchart of a method for determining a flight rule according to an embodiment of the present invention.
  • the method of the embodiment of the present invention includes the following steps.
  • S401 Acquire an initial flight rule regarding the detection object according to an image position of the detection object in the environment image and the detection position.
  • the initial flight rule includes one or more flight trajectories.
  • the above reference is further referred to when generating the initial flight rule. Limit parameters.
  • the remaining energy value includes data such as a remaining power value of the drone.
  • S403 Adjust the initial flight rule according to the remaining energy value, and use the adjusted rule as the flight rule for the detection object.
  • a distance capable of supporting the flight of the aircraft is determined according to the remaining energy value, and if the flight trajectory in the initial flight rule cannot be covered, the partial flight trajectory may be selected to obtain the flight rule that needs to be executed this time.
  • the initial flight rule is automatically recorded and the flight trajectory that has been executed is recorded, so as to continue to proceed from the part of the flight trajectory that has been executed on the basis of the initial flight rule for the next time. Redefine the adjustments on the initial flight rules to generate new flight rules including flight trajectories.
  • the flight trajectory and the like can be intelligently adjusted according to the battery power of the drone, etc., thereby further ensuring the safety of the inspection.
  • the embodiment of the invention further provides a computer storage medium, wherein the computer storage medium stores program instructions, and when the program instructions are executed, the corresponding method of the embodiment corresponding to FIG. 1, FIG. 2, FIG. 3 or FIG. 4 is implemented. .
  • FIG. 5 it is a schematic structural diagram of a facility detecting device according to an embodiment of the present invention.
  • the facility aircraft in the embodiment of the present invention may be installed in an aircraft capable of performing a patrol task, such as a drone.
  • the facility aircraft includes the following structure.
  • the obtaining module 501 is configured to acquire an environment image including the target facility when the aircraft is located at the detection location of the target facility, and the determining module 502 is configured to determine, from the environment image, an image region to which the target facility belongs, And performing image segmentation on the image region to obtain a detection object about the target facility; the processing module 503, configured to acquire, according to the image location of the detection object in the environment image and the detection location, A flight rule; a control module 504, configured to control the aircraft flight according to the flight rule to facilitate completion of detection of the target facility.
  • the flight rule includes a flight trajectory
  • the processing module 503 is specifically configured to acquire a patrol policy associated with the detection object; and according to the image location and location of the detection object in the environment image The detection location is generated to generate a flight trajectory that satisfies the inspection strategy.
  • the obtained detection object includes a plurality of detection objects, each of which is associated with a patrol inspection policy, and the processing module 503 is specifically configured to acquire a patrol inspection policy of each detection object; Generating a flight rule in the image position and the detection position in the environment image; wherein the flight rule includes a flight trajectory that satisfies all patrol strategies, or includes multiple flight trajectories, and each flight trajectory satisfies a part of the patrol strategy .
  • the generated flight rule further satisfies a preset limit parameter;
  • the limit parameter includes any one or more of a flight distance parameter, a flight duration parameter, a flight safety parameter, and an energy loss parameter.
  • the apparatus may further include: a generating module 505, configured to generate a detection parameter, the detection parameter is used to instruct the aircraft to sense the inspection object during the control of the flight of the aircraft
  • the parameter includes: a shooting angle parameter of a sensor for detecting the detection object, and a shooting parameter of the camera for capturing the detection object.
  • the processing module 503 is further configured to control the according to the flight rule During the flight of the aircraft, acquiring the detected detection image; updating the flight rule according to the position of the detection object in the detection image and the inspection strategy set for the detection object; controlling the flight of the aircraft according to the updated flight rule, so that The detection of the target facility is completed.
  • the processing module 503 is specifically configured to acquire an initial flight rule regarding the detection object according to an image position of the detection object in the environment image and the detection position; and acquire remaining energy of the aircraft a value; adjusting the initial flight rule according to the remaining energy value, and using the adjusted rule as the flight rule for the detection object.
  • the determining module 502 is specifically configured to acquire an object model preset for the target facility, and perform image segmentation on the image region according to the object model, so that the shape similarity between the object model and the object model is similar. The object of detection of the condition.
  • the object model is configured with a model identifier, and the patrol policy associated with the detected object is obtained according to the model identifier.
  • the device may further include: a setting module 506, configured to configure one or more facility location points to be detected on the interaction interface displaying the map; corresponding to the selected one or more facility location points The facility is determined to be the target facility; the aircraft is controlled to fly to the target facility based on the selected facility location point to facilitate flight to the detection location for the target facility.
  • a setting module 506 configured to configure one or more facility location points to be detected on the interaction interface displaying the map; corresponding to the selected one or more facility location points The facility is determined to be the target facility; the aircraft is controlled to fly to the target facility based on the selected facility location point to facilitate flight to the detection location for the target facility.
  • the device may further include: a receiving module 507, configured to receive location information returned by the aircraft, where the location information includes: distance information of the aircraft generated by the aircraft relative to a target object, and Direction information, or position coordinate information of the aircraft returned by the aircraft.
  • a receiving module 507 configured to receive location information returned by the aircraft, where the location information includes: distance information of the aircraft generated by the aircraft relative to a target object, and Direction information, or position coordinate information of the aircraft returned by the aircraft.
  • control module 504 is further configured to control the aircraft to fly to the target facility according to a specified rule during flight of the aircraft to the target facility, where the specified rule is used to indicate that the aircraft is flying to obtain a depth map. At least two shooting positions that can be taken at different angles; acquiring a depth map of the aircraft in a traveling direction based on at least two shooting positions; performing flight obstacle avoidance processing according to the acquired depth map.
  • control module 504 is further configured to: after the detecting the target facility is completed, controlling the return of the aircraft according to a preset return trajectory; the preset return trajectory includes: recording the location A flight trajectory before the aircraft flies to the detection position.
  • one of the facilities when performing inspections on certain facilities, especially for facilities with higher facilities or in areas that are not easily accessible, one of the facilities can be based on image recognition and automatic control of flight. Or a plurality of objects to be inspected are inspected, which reduces the labor cost and safety of the inspection, and improves the efficiency of the inspection. Moreover, it can automatically return to the air and avoid obstacles, further satisfying the automation and intelligent requirements of the inspection, and improving the safety of the inspection.
  • the control device of the embodiment of the present invention includes a power supply circuit, and the control device may be powered by a single battery or through a power supply interface. Powered by batteries from aircraft such as drones.
  • the control device may also include a processor 601, a data interface 602, and a memory 603.
  • the data interface 602 is primarily used to interact with the aircraft, or further, the data interface 602 can also interact with data from a monitoring device on the ground for receiving and displaying data such as images detected by the aircraft.
  • the memory 603 may include a volatile memory such as a random-access memory (RAM); the memory 603 may also include a non-volatile memory such as a flash memory. (flash memory), hard disk drive (HDD) or solid-state drive (SSD); the memory 603 may also include a combination of the above types of memories.
  • RAM random-access memory
  • non-volatile memory such as a flash memory.
  • flash memory flash memory
  • HDD hard disk drive
  • SSD solid-state drive
  • the memory 603 may also include a combination of the above types of memories.
  • the processor 601 can be a central processing unit (CPU).
  • the processor 601 may further include a hardware chip.
  • the hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof.
  • the PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a general array logic (GAL), or any combination thereof.
  • the memory 603 is further configured to store program instructions.
  • the processor 601 can invoke the program instructions to implement the facility detection method as shown in the embodiments corresponding to Figures 1, 2, 3 and 4 of the present application.
  • the processor 601 is configured to acquire an environment image including the target facility when the aircraft is located at a detection location for the target facility, and determine an image to which the target facility belongs from the environment image. a region, and performing image segmentation on the image region to obtain a detection target for the target facility; acquiring a flight rule regarding the detection object according to an image location of the detection object in the environment image and the detection location; The flight rule generates a control command, and The flight is sent to the aircraft through the data interface 602 to control the flight of the aircraft to facilitate completion of detection of the target facility.
  • the flight rule includes a flight trajectory
  • the processor 601 is configured to acquire a patrol policy associated with the detection object; according to an image position of the detection object in the environment image and the detecting Position, generating a flight trajectory that satisfies the inspection strategy.
  • the obtained detection object includes multiple, each detection object is associated with a patrol policy, and the processor 601 is configured to acquire a patrol policy of each detection object;
  • the image position and the detected position in the environment image generate a flight rule; wherein the flight rule includes a flight trajectory that satisfies all the patrol strategies, or includes a plurality of flight trajectories, and each flight trajectory satisfies a part of the patrol strategy.
  • the generated flight rule further meets a preset limit parameter;
  • the limit parameter includes any one or more of a flight distance parameter, a flight duration parameter, a flight safety parameter, and an energy loss parameter.
  • the processor 601 is further configured to generate a detection parameter, and send the detection parameter to the aircraft through the data interface 602, where the detection parameter is used to indicate the location during the process of controlling the flight of the aircraft
  • the sensing parameter comprising: a shooting angle parameter of a sensor for detecting the detection object, and a shooting parameter of the camera for capturing the detection object.
  • the processor 601 is further configured to: when detecting the flight of the aircraft according to the flight rule, acquire the detected detection image; according to the position of the detection object in the detection image and the tour set for the detection object Checking the strategy, updating the flight rules; controlling the aircraft flight based on the updated flight rules to facilitate completion of the detection of the target facility.
  • the processor 601 is configured to acquire an initial flight rule regarding the detection object according to an image position of the detection object in the environment image and the detection position; and acquire a residual energy value of the aircraft; The initial flight rule is adjusted according to the remaining energy value, and the adjusted rule is taken as the flight rule for the detection object.
  • the processor 601 is configured to acquire an object model preset for the target facility, perform image segmentation on the image region according to the object model, and obtain a similarity condition between the shape similarity with the object model. Test object.
  • the object model is configured with a model identifier, and the model identifier is obtained according to the model identifier. Joint inspection strategy.
  • the processor 601 is further configured to: configure one or more facility location points to be detected on the interaction interface that displays the map; determine, as the target, the facility corresponding to the selected one or more facility location points Facility; controls the aircraft to fly to the target facility based on the selected facility location point to facilitate flight to the detection location for the target facility.
  • the processor 601 is further configured to receive location information returned by the aircraft, where the location information includes: distance information and direction information of the aircraft generated by the aircraft with respect to a target object, or by Position coordinate information of the aircraft returned by the aircraft.
  • the processor 601 is further configured to: when the aircraft is flying to the target facility, control the aircraft to fly to the target facility according to a specified rule, where the specified rule is used to instruct the aircraft to fly to obtain a depth map. At least two shooting positions capable of being photographed at different angles; acquiring a depth map of the aircraft in a traveling direction based on at least two shooting positions; performing flight obstacle avoidance processing according to the acquired depth map.
  • the processor 601 is further configured to: after the detecting the target facility is completed, controlling the return of the aircraft according to a preset return trajectory; the preset return trajectory includes: recording the The flight path of the aircraft before it travels to the detection location.
  • Embodiments of the present invention when inspecting certain facilities, particularly for facilities with higher facilities or areas that are not easily accessible, can detect one or more of the facilities based on image recognition and automatic control of flight.
  • the object is inspected, which reduces the labor cost and safety of the inspection and improves the efficiency of the inspection. Moreover, it can automatically return to the air and avoid obstacles, further satisfying the automation and intelligent requirements of the inspection, and improving the safety of the inspection.

Landscapes

  • Engineering & Computer Science (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Image Analysis (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种设施检测方法及控制设备能够自动实现对某些设施巡检，节省了人力成本，提高了巡检效率，其中，设施检测方法包括：当飞行器位于针对目标设施的检测位置时，获取包括目标设施的环境图像(S201，S301)；从环境图像中确定出目标设施所属的图像区域，并对图像区域进行图像分割，得到关于目标设施的检测对象(S202，S302)；根据检测对象在环境图像中的图像位置和检测位置，获取关于检测对象的飞行规则(S203，S303)；根据飞行规则控制飞行器飞行(S304)，以便于完成对目标设施的检测(S204)。

Description

一种基于飞行器的设施检测方法及控制设备 技术领域
本发明涉及计算机控制技术领域，尤其涉及一种基于飞行器的设施检测方法及控制设备。
背景技术
某些设施，需要用户定期对这些设施进行巡检、维护，以便于确认这些设施的安全状态。例如，对于电塔、大桥、高楼等设施，需要定期巡检来确保这些设施的安全及正常运行。
然而，对于一些处于特殊位置的设施，特别是一些处于险峻位置处的电塔、大桥等设施，进行周期性的巡检变得十分困难。并且，此类设施通常数目较多，周期性的巡检需要耗费大量的人力。
发明内容
本发明实施例提供了一种基于飞行器的设施检测方法及控制设备，通过对飞行器的控制来实现对目标设施的巡检。
一方面，本发明实施例提供了一种基于飞行器的设施检测方法，包括：
当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；
从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；
根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；
根据所述飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
另一方面，本发明实施例还提供了一种控制设备，包括：处理器和数据接口；
所述数据接口，用于与飞行器交互数据；
所述处理器，用于当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；从所述环境图像中确定出所述目标设施所属的图像区 域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；根据所述飞行规则生成控制指令，并通过所述数据接口发送给所述飞行器以控制所述飞行器飞行，以便于完成对所述目标设施的检测。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。
附图说明
图1是本发明实施例的一种对目标设施进行检测的方法流程示意图；
图2是本发明实施例的一种基于飞行器的设施检测方法的流程示意图；
图3是本发明实施例的另一种基于飞行器的设施检测方法的流程示意图；
图4是本发明实施例的一种确定飞行规则的方法流程示意图；
图5是本发明实施例的一种设施检测装置的结构示意图；
图6是本发明实施例的一种控制设备的结构示意图。
具体实施方式
本发明实施例通过使用视觉技术，采集图像，可以在远处检测、识别出需要检测的目标设施，自动飞行到待检测的目标设施的附近。并通过图像分割识别技术，将目标设施中的各个部分分割区分开来，得到该目标设施的一个或者多个检测对象，有针对地使用携带的传感器(如相机，热成像仪等)对识别出的各个检测对象进行检测与记录。检测记录结束后，无人机等飞行器将自动返航或者在电量充足的情况下去临近的下一个需要检测的目标设施进行检测。
如图1所示，是本发明实施例的一种对目标设施进行检测的方法流程示意图。本发明实施例的设施检测方法可以由一个控制设备来执行，该控制设备可以配置在飞行器上。并且，在本发明实施例中，以无人机来作为飞行器对所述方法进行说明，该无人机上挂载有用于对目标设施进行检测的传感器。该控制设备所执行的主要步骤如下。
S101：确定目标设施的位置。可以利用GPS(Global Positioning System， 全球定位系统)信息等位置信息，大概定义目标设施的位置，控制无人机自主向该目标设施飞行，以至目标设施出现在无人机的观察范围之内，该观察范围主要是指无人机挂载的传感器的探测范围，例如相机的拍摄范围。用户可以预先在对控制设备进行配置的用户界面上，输入一个或多个目标设施的位置信息，例如，在一个显示有地图的界面中，通过触摸点击的方式在该包括地图的界面上指定一个或多个位置点，控制设备可以将这些位置点记录为目标设施的位置点。控制设备可以在无人机开启了巡检模式时，控制无人机基于各个目标设施的位置点飞行，以便于向对应的一个或者多个目标设施飞行，以监控一个目标设施，或者监控某条路线上的多个目标设施，例如，相连的多个电塔。
S102：基于图像分割识别以及深度图进行避障处理。在向目标设施飞行的过程中，可以基于图像分割识别以及获取的深度图进行主动避障，以便于安全飞行到可以检测目标设施的区域。
无人机自主飞行过程中需要检测飞行路径中的各种障碍物。避障的首要任务是检测出飞行方向上的障碍物。深度图可以基于双目视觉探测的方式探测并计算得到，从而定位飞行路径上的障碍。深度图的获取可以使用双目计算匹配得到，也可以用基于结构光或者红外的设备计算得到。基于结构光或者红外的设备可得到相对质量更高的深度图。
为了进一步提高深度图的精度，使得在纹理不丰富并且目标物过小时能够避免发生漏检与误检的情况，在本发明实施例中可以进一步结合图像分割技术，来进行障碍物的确定和飞行避障。由于图像分割不需要做匹配，对于纹理不丰富的区域也能有较好的识别效果，因此可以配合深度图一起使用，一方面可以得到更好的深度图，另一方面，也可以给深度图中的每一个点赋予实际意义，有助于无人机进行路径规划，确定无人机飞行的飞行规则。
在飞行的过程中，无人机不断估计、修正与目标设施的相对位置。在一个实施例中，无人机可以使用视觉跟踪算法，将目标设施锁定在图像可观测的范围内，通过目标设施在图像中大小的变化和当前的飞行速度，估计与目标设施的相对距离。无人机也可以按照特定轨迹飞行来大致获取场景中物体的深度信息，为距离避障提供参考。
在一个实施例中，可以基于拍摄到的图像识别确定出本次检测的巡检场景，巡检场景具体可以根据本次检测的目标设施来进行分类的，例如包括：巡 检电塔的场景、巡检大桥的场景等，确定出的巡检场景可以为飞行避障、规划飞行路线提供参考信息。例如，在巡检场景为巡检电塔的场景时，由于在这些场景下，相邻的电塔之间一般有电力线相连接，连接的部位相对比较固定，无人机在设定路线时，可选择绕开电力线密集的区域。
针对特定的巡检场景，无人机可以配备相应的传感器以进一步提高避障的可靠性。比如，针对电力系统的设备巡检，热成像仪可以用于检测电线的存在，对于需要靠近电力线进行巡检提供更鲁棒的避障。
S103：当检测到目标设施进入无人机的观察范围之后，检测目标设施的具体位置，并向目标设施飞行。本发明实施例可以基于人工特征的视觉检测方法来检测图像范围内的目标设施。也可以基于深层神经网络的识别算法通过对电塔等目标设施的海量图像数据学习，可以从数据中学到更加稳定可靠的图像特征，从而得到更精确的识别结果。
检测算法运行在无人机观察到的图像上，在图像中检测、定位目标设施的位置。一旦在图像中发现待检测的目标设施，在图像中将其锁定，并逐渐飞向待检测的该目标设施。此过程中，可以使用跟踪算法锁定检测到的目标设施，并利用检测的结果作为参考来确定无人机的飞行路线。
为了定位目标设施在图像中的位置，可以在目标设施中选择多个特征点。基于全图的分割识别可以使得无人机能够根据目标设施的类别选择更加稳定的图像特征点，这些图像特征点一般需要一直稳定地存在于目标设施上，不移动，容易被检测发现。比如，选择在电塔上的特征点比水面的特征点要更加稳定，基于这些图像特征点，可以提高为无人机计算SLAM(Simultaneous localization and mapping，同时定位与建图)的精度。因此，无人机可以更加灵活的调整姿态与路线。比如，为了使飞行路线更加安全，可以选择在局部路径中，待检测的目标设施不出现在视野范围之内，通过障碍物后，根据估计的相对位置，重新将待检测的目标设施锁定在图像中。
S104：识别设施的各个组成部分，并针对性的检测与记录。在到达了对目标设施的检测位置时，例如，与目标设施的距离在预设的距离范围内的区域中的某个位置时，可以进一步地基于图像从目标设施中识别出检测对象，例如，目标设施为电塔时，识别出本次需要检测的检测对象为整个塔头，或者固定电力线的部件。
图像分割算法将提供像素级别的识别与分割，提供图像中每一个像素的类别信息，这里的类别信息的作用主要在于确定对该类别信息对应的检测对象，进而确定出需要采用的巡检策略，指导无人机飞行。
可以将待检测的目标设施从图像中分割出来，确定出仅包括目标设施的局部图像区域，再对该局部图像区域进行分析识别，对目标设施的不同的位置识别出不同的类别，得到本次需要检测的关键部分，该关键部分即为检测对象，从而可以对目标设施的关键部分进行针对性的检测与记录。为了给巡检提供更加准确的信息，在与目标设施的距离小于预设的距离阈值时，可以使用针对目标设施的特定部位进行分割、识别的对象模型。
这种对象模型除了能将目标设施与背景分割开之外，还可以细化分割、识别目标设施的组成部分，得到一个或多个检测对象。对于巡检过程中，可以对例如电塔等目标设施的各个组成部分进行识别，在图像中标记出各组成部分的位置。用户可以事先指定针对每个组成部分的巡检策略。在一个实施例中，该巡检策略包括但不限于：环绕拍摄，远近持续视频拍摄，以及定点使用高精相机拍照。与远程识别的图像分割模型类似，所述对象模型的输入是图像，输出像素级别的识别结果，其中表示了每个像素所属的特定类别，可以包括背景以及细化的目标设施中各种组成部件，也就是说每个像素点可以为背景类别的像素点，或者为某个检测对象类别(例如塔头)的像素点。所述对象模型可以根据实际需要巡检的目标设施来配置，例如，对于电塔，可以配置塔头、塔脚等对象模型以便于识别出目标设施在图像中的塔头和塔脚。
根据识别结果以及用户的设定，生成具体的巡检方案。巡检方案的生成包括：轨迹的生成，每一段轨迹的时间分配等。求解轨迹的准则会考虑的信息包括：完成所有检测对象的巡检的前提下，如何更可能的减少飞行时间，如何选择更加安全的飞行路线，比如包括如何避开电力线，如何使得拍摄画面稳定可靠，无人机的特性，比如无人机的最大最小加速度，以及进行检测的检测设备的特效，例如相机的FOV，其它用于检测的传感器的最佳使用距离等。
当轨迹生成完成后，无人机执行该计算出的轨迹，对目标设施各个待检测部件(检测对象)进行巡检。此时，无人机仍然会不断的更新观察，实时动态的修正轨迹以保证能安全有效的进行巡检。
S105：检测完毕，自动返航。待完成检测任务，无人机可以根据开始执行 巡检任务时记录的出发点信息和/或记录的飞行数据来实现自动返航。在一个实施例中，可以结合视觉里程计(visual odometry)与GPS信息指导无人机返航。视觉里程计通过图像特征匹配的方式估计无人机在执行任务中的轨迹。结合图像分割算法的视觉里程计可以选择更好的匹配特征从而实现更精准的轨迹估计。
可以通过分析执行任务时的飞行轨迹，计算出返回起飞点最优的返航路线。比如，执行任务中探索、尝试、重复的飞行轨迹可以绕过。其中，可以绕过的飞行轨迹可以是在实现上述的视觉里程计得到的飞行轨迹中标记出的部分轨迹。
在一个实施例中，还可以结合GPS坐标，纠正为返航估计的飞行轨迹。并且可以进一步地通过定位传感器，在无人机到达起飞点附近之后，实现更加精准的返航。
在一个实施例中，无人机可以根据之前的记录所生成的视觉里程计确认最佳的返航路线，同时在返航的过程中开启避障功能。
另外，在实现飞行避障时，可以结合已知的2D/3D地图来确定飞行路线上的障碍物，例如确定在地图上已经标识的建筑物、山脉等障碍物，进而可以在确定飞行路径时选择绕过这些障碍物。为了更快地从目标设施中确定出检测对象，可以在目标设施上设定用于标识检测对象的标记，基于这些标记以及拍摄的图像，快速地从目标设施所在的图像区域中分割并定位出一个或多个检测对象。在获取深度图时，不仅可以基于双目视觉的方式获取，也可以使用类似激光雷达之类的装置来获取。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。
再请参见图2，是本发明实施例的一种基于飞行器的设施检测方法的流程示意图，本发明实施例的所述方法可以由一个专用的控制设备来执行，该控制设备可以配置在无人机等飞行器上。该控制设备也可以作为一个地面端设备，通过无线的方式与无人机等飞行器交互数据，进而完成对目标设施的巡检任务。
S201：当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像。该检测位置的作用主要在于：可以触发对目标设施的相关处理以便于完成对目标设施的巡检任务。
所述检测位置可以是指位于某个位置区域中的位置点，该位置区域可以是指与目标设施的距离在一个预设的距离范围内的区域。控制设备基于检测到的所述飞行器的位置(例如GPS位置坐标)或所述飞行器上报的位置，并根据目标设施的位置，来确定飞行器是否到达针对目标设施的检测位置。
是否到达检测位置也可以由飞行器自行判断，在一个实施例中，飞行器可以根据拍摄到的包括目标设施的图像进行分析，基于预设的目标设施的实际大小、目标设施在图像中的大小、目标设施在图像中的位置来估计飞行器与目标设施之间的距离，如果该距离在一个预设的距离范围内，则可以认为飞行器到底了对目标设施进行检测的检测位置。
该检测位置还可以是一个特定的位置，在该位置上得到的图像中包括所述目标设施，或者在该位置上能够对图像中包括的目标设施进行分割，例如，如果图像中，目标设施所占区域满足预设的条件(目标设施的像素点的个数大于预设的阈值、或者目标设施所占图像区域的面积大于预设的阈值)时，飞行器所在的位置即可以认为是检测位置。
无人机等飞行器上配置了摄像机等拍摄装置，到达检测位置后，触发拍摄装置采集环境图像，对于在S201中采集到的环境图像，主要用于确定出检测对象，并确定针对检测对象进行检测时所使用的飞行规则。
在一个实施例中，可以在显示地图的交互界面上配置一个或者多个待检测的设施位置点；将被选中的一个或者多个设施位置点所对应的设施确定为目标设施；根据选中的设施位置点控制飞行器向目标设施飞行，以便于飞行到针对目标设施的检测位置。如果用户在交互界面上选择了多个目标设施，可以控制飞行器由远及近或由近及远先后执行下述的步骤完成多个目标设施的巡检。或者基于飞行器的剩余能量，完成其中的一个或者其中的部分目标设施的巡检。如果控制设备为一个包括显示器的智能终端，例如智能手机、平板电脑等，则可以直接显示一个包括地图的交互界面给用户。如果控制设备挂载在飞行器上，则所述控制设备可以通过自带的无线通信接口，或者通过所述飞行器上的无线通信接口，发送指令以触发用于接收飞行器的数据的检测端显示交互界 面，并接收在所述交互界面上确定的目标设施的位置，以控制飞行器飞行。
S202：从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象。确定目标设施所属的图像区域也可以是基于图像分割实现的。可以基于所述环境图像中像素的亮度及颜色进行图像分割，得到环境图像中所述目标设施所属的图像区域和目标设施的各个检测对象。
所述检测对象可以是所述目标设施的全部，所述检测对象也是所述目标设施的部分部件，例如电塔的塔头、固定电力线的固定部件等等。检测对象由用户预先指定。例如，用户指定需要巡检整个电塔，那么在S202中，可以在得到电塔的图像区域后，将该图像区域中的整个电塔作为检测对象。用户也可以指定巡检电塔的塔头部分，在得到电塔的图像区域后，再分割得到塔头作为检测对象。
在对所述图像区域进行分割时，主要是基于为所述目标设施预设的对象模型来对所述图像区域进行分割，从图像区域的目标设施中确定一个或者多个检测对象。在一个实施例中，所述对该图像区域进行图像分割，得到关于所述目标设施的检测对象具体可以包括：获取为所述目标设施预设的对象模型；按照对象模型对图像区域进行图像分割，得到与所述对象模型之间的形状相似度满足相似度条件的检测对象。所述对象模型主要用于识别出目标设施上的某个组成部分，例如，预设的关于塔头的对象模型能够协助识别出电塔的塔头部分。可以预设不同角度的多个对象模型来对应一个检测对象，以便于在不同角度获取到的关于目标设施的图像时，均能够准确地分割确定出目标设施中的检测对象。进一步地，所述对象模型配置有模型标识，根据模型标识获取对应检测对象关联的巡检策略。该模型标识可以为一个名称，例如上述的关于塔头的对象模型的模型标识即为“塔头”。在基于对象模型识别出一个检测对象后，该检测对象的标识与对象模型对应，检测对象的标识可以与对象模型的模型标识相同。基于检测对象的标识，可以从预先设置的映射关系库中确定出与该检测对象的标识关联的巡检策略，这些巡检策略主要用于指示如何对检测对象进行巡检的巡检规则，包括环绕飞行环绕拍摄，远近持续视频拍摄，以及定点使用高精相机拍照等规则。在执行下述的S203时，可以进一步地基于巡检规则来获取关于所述检测对象的飞行规则。
S203：根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则。图像位置可以是检测对象在图像中的像素位置，基于图像位置可以确定检测对象相对于飞行器的方位。可以以所述检测位置为起点，确定出能够从上方、下方等方位对检测对象进行检测的飞行规则，或者确定出环绕所述检测对象飞行的飞行规则。该飞行规则可以为一个飞行轨迹，例如控制无人机分飞行的飞行轨迹，基于该飞行轨迹，能够实现对
S204：根据所述飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。在确定出飞行规则后，控制飞行器按照该飞行规则飞行，即可完成对检测对象的巡检。
在所述S201之前向目标设施飞行以到达检测位置的过程中、或者在S204控制所述飞行器飞行的过程中，可以实时或周期性地接收所述飞行器返回的位置信息，所述位置信息包括：由所述飞行器生成的所述飞行器相对于目标对象的距离信息和方向信息，或由所述飞行器返回的所述飞行器的位置坐标信息。根据接收的位置信息和所述目标设施的位置，在交互界面上实时显示所述飞行器和目标设施之间的相对位置。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。
再请参见图3，是本发明实施例的另一种基于飞行器的设施检测方法的流程示意图，本发明实施例的所述方法可以由一个专用的控制设备来执行，该控制设备可以配置在无人机等飞行器上。该控制设备也可以作为一个地面端设备，通过无线的方式与无人机等飞行器交互数据，进而完成对目标设施的巡检任务。
S301：当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像。所述检测位置是在离所述目标设施在预设的距离范围内的其中一个位置。所述环境图像时所述飞行器上携带的拍摄头等传感器采集到的环境图像。
S302：从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象。基于图像分割技术 分割确定出所述图像区域，并基于预设的对象模型来分析确定出目标设施的检测对象。
S303：根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则。所述图像位置的作用在于：需要确保检测对象一直位于图像中。根据所述图像位置能够确定检测对象的相对方向，进一步地，在对该检测对象进行检测时，如果需要将当前需要检测的检测对象保持在画面中间，则在生成包括飞行轨迹的飞行规则时，考虑该图像位置。当需要检测的检测对象仅包括一个时，只需要生成针对该检测对象的飞行规则，例如生成环绕该检测对象的飞行轨迹。如果检测对象包括多个，则需要生成一个飞行规则，基于该飞行规则能够先后对多个检测对象进行检测。例如，对电塔的检测包括“塔头”、“塔身”、“塔脚”三个检测对象时，可以生成飞行规则，该飞行规则包括从检测位置开始，先沿着飞行规则中的飞行轨迹往上飞检测塔头，接着沿着该飞行轨迹往下飞检测塔身，最后沿着轨迹再到塔脚，从而在多段飞行轨迹上完成对“塔头”、“塔身”、“塔脚”三个检测对象的检测。
其中，所述飞行规则包括飞行轨迹，所述S303具体可以包括：获取与所述检测对象关联的巡检策略；根据所述检测对象在所述环境图像中的图像位置和所述检测位置，生成满足所述巡检策略的飞行轨迹。在检测对象仅为一个时，直接基于该检测对象对应的巡检策略即可生成一个飞行轨迹。例如，在一个简单的实施例中，检测对象在飞行器采集到的图像的中间位置，当巡检策略为远近持续视频拍摄，则可以生成以所述检测位置为起始点，生成一条从起始点到所述目标设施所在位置的一条直线轨迹，以便于飞行器能够由远及近地持续拍摄所述目标设施的检测对象。
如果得到的所述检测对象包括多个，则每一个检测对象均关联了巡检策略，所述S303具体可以包括：获取每一个检测对象的巡检策略；根据各检测对象在所述环境图像中的图像位置和所述检测位置，生成飞行规则；其中，所述飞行规则中包括满足所有巡检策略的飞行轨迹，或者包括多段飞行轨迹，每一段飞行轨迹满足部分巡检策略。
在上述生成飞行规则时，还进一步基于预设的限制条件对飞行规则的生成进行约束。所述限制条件包括：基于飞行参数和检测参数设置的条件，所述飞行参数包括：飞行距离参数、飞行时长参数、飞行安全参数、能量损耗参数、 飞行速度参数中的任意一种或多种。
在一个实施例中，如果飞行距离参数被配置为1等有效数值时，表明在生成满足一个或者多个巡检策略的飞行轨迹时，进一步还要求飞行轨迹的总长度最短，使得飞行器的飞行距离最短，以节省能耗。当飞行时长参数被配置为1等有效数值时，表明在生成满足一个或者多个巡检策略的飞行轨迹时，进一步还要求飞行器按照所采用的飞行轨迹时以预设的速度飞行时，所耗费的时间最短，以提高巡检效率。当飞行安全参数被配置为1等有效数值时，表明优先选择安全的飞行轨迹，将一些可能存在障碍物的轨迹排除，例如在电塔作为目标设施时，排除掉可能会穿过电线的轨迹，以确保飞行安全。当能量损耗参数被配置为1等有效数值时，表明优先选择能耗低的轨迹作为最终的飞行轨迹。
S304：根据所述飞行规则控制所述飞行器飞行。在一个实施例中，还包括：生成检测参数，所述检测参数用于在控制所述飞行器飞行的过程中指示所述飞行器对检测对象进行巡检的感测参数，所述感测参数包括：用于对检测对象进行检测的传感器的拍摄角度参数、用于对检测对象进行拍摄的拍摄机的拍摄参数。例如，对于通过云台挂载拍摄机的无人机，检测参数具体可以是用于控制云台角度的参数，对拍摄机焦距、白平衡、快门等进行控制的参数。
S305：在根据所述飞行规则控制所述飞行器飞行的过程中，获取检测得到的检测图像。根据所述飞行规则控制所述飞行器飞行的过程即为对检测对象进行巡检的过程，拍摄到图像或者根据图像生成的视频可以即时传递给巡检用户，巡检用户通过查看图像或者根据图像生成的视频来确定一个或者多个检测对象是否正常。
对于采集到的图像，控制设备还可以进一步地进行分析，确定当前检测的对象在检测图像中的位置。例如当前在检测电塔的塔头时，分析确定作为检测对象的塔头在检测图像中的位置。同样可以根据图像分割技术来确定当前检测的检测对象所在的图像区域，并进一步地确定该检测对象所在的像素位置。
S306：根据检测图像中检测对象的位置和为检测对象设置的巡检策略，更新所述飞行规则。更新飞行规则主要为了保证能够检测到当前需要检测的检测对象，例如需要保证检测对象的位置是在检测图像的图像中心区域。在本发明实施例中，可以执行所述S306，和/或对上述提及的感测参数进行更新调整。或者先调整感测参数，如果无法满足预设的检测需求，例如无法保证检测对象 的位置是在检测图像的图像中心区域，则执行所述S306，更新飞行规则(还可以进一步结合对感测参数的更新调整)，以满足预设的检测需求。
S307：根据更新后的飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。也就是说继续对检测对象进行检测得到对应的检测图像或者基于图像生成的视频，并返回给巡检用户查看。
S308：在完成对所述目标设施的检测后，根据预设的返航轨迹控制所述飞行器返回；所述预设的返航轨迹包括：记录的在所述飞行器飞行至所述检测位置之前的飞行轨迹。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。并且还能够自动返航和避障，进一步地满足了巡检的自动化、智能化需求，提高了巡检的安全性。
再请参见图4，是本发明实施例的一种确定飞行规则的方法流程示意图，本发明实施例的所述方法包括如下步骤。
S401：根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的初始飞行规则。初始飞行规则包括一段或者多段飞行轨迹，在一个实施例中，在生成初始飞行规则时，除了考虑所述检测对象在环境图像中的图像位置和所述检测位置外，还进一步参考了上述提及的限制参数。
S402：获取所述飞行器的剩余能量值。所述剩余能量值包括无人机的剩余电量值等数据。
S403：根据所述剩余能量值对所述初始飞行规则进行调整，将调整后得到的规则作为关于所述检测对象飞行规则。根据剩余能量值确定能够支撑飞行器飞行的距离，如果不能够覆盖所述初始飞行规则中的飞行轨迹，则可以选择执行部分飞行轨迹，得到本次需要执行的飞行规则。在执行部分飞行轨迹对检测对象进行巡检后，自动记录所述初始飞行规则并记录已经执行的飞行轨迹，以便于下一次继续在该初始飞行规则的基础上，从已执行的部分飞行轨迹，重新确定在初始飞行规则上进行调整，生成新的包括飞行轨迹的飞行规则。
可以根据无人机的电池电量等情况来对飞行轨迹等进行智能的调整，进一步确保了巡检安全。
本发明实施例还提供了一种计算机存储介质，该计算机存储介质中存储有程序指令，在执行这些程序指令时，实现上述图1、图2、图3或图4所对应实施例的相应方法。
下面对本发明实施例的设施飞行器及控制设备进行描述。
请参见图5，是本发明实施例的一种设施检测装置的结构示意图，本发明实施例的所述设施飞行器可以设置到无人机等可飞行的能够执行巡检任务的飞行器中。所述设施飞行器包括如下结构。
获取模块501，用于当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；确定模块502，用于从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；处理模块503，用于根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；控制模块504，用于根据所述飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
进一步可选地，所述飞行规则包括飞行轨迹，所述处理模块503，具体用于获取与所述检测对象关联的巡检策略；根据所述检测对象在所述环境图像中的图像位置和所述检测位置，生成满足所述巡检策略的飞行轨迹。
进一步可选地，得到的所述检测对象包括多个，每一个检测对象均关联了巡检策略，所述处理模块503，具体用于获取每一个检测对象的巡检策略；根据各检测对象在所述环境图像中的图像位置和所述检测位置，生成飞行规则；其中，所述飞行规则中包括满足所有巡检策略的飞行轨迹，或者包括多段飞行轨迹，每一段飞行轨迹满足部分巡检策略。
进一步可选地，生成的飞行规则还满足预设的限制参数；所述限制参数包括：飞行距离参数、飞行时长参数、飞行安全参数、能量损耗参数中的任意一种或多种。
进一步可选地，所述装置还可以包括：生成模块505，用于生成检测参数，所述检测参数用于在控制所述飞行器飞行的过程中指示所述飞行器对检测对象进行巡检的感测参数，所述感测参数包括：用于对检测对象进行检测的传感器的拍摄角度参数、用于对检测对象进行拍摄的拍摄机的拍摄参数。
进一步可选地，所述处理模块503，还用于在根据所述飞行规则控制所述 飞行器飞行的过程中，获取检测得到的检测图像；根据检测图像中检测对象的位置和为检测对象设置的巡检策略，更新所述飞行规则；根据更新后的飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
进一步可选地，所述处理模块503，具体用于根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的初始飞行规则；获取所述飞行器的剩余能量值；根据所述剩余能量值对所述初始飞行规则进行调整，将调整后得到的规则作为关于所述检测对象飞行规则。
进一步可选地，所述确定模块502，具体用于获取为所述目标设施预设的对象模型；按照对象模型对图像区域进行图像分割，得到与所述对象模型之间的形状相似度满足相似度条件的检测对象。
进一步可选地，所述对象模型配置有模型标识，根据模型标识获取与检测对象关联的巡检策略。
进一步可选地，所述装置还可以包括：设置模块506，用于在显示地图的交互界面上配置一个或者多个待检测的设施位置点；将被选中的一个或者多个设施位置点所对应的设施确定为目标设施；根据选中的设施位置点控制飞行器向目标设施飞行，以便于飞行到针对目标设施的检测位置。
进一步可选地，所述装置还可以包括：接收模块507，用于接收所述飞行器返回的位置信息，所述位置信息包括：由所述飞行器生成的所述飞行器相对于目标对象的距离信息和方向信息，或由所述飞行器返回的所述飞行器的位置坐标信息。
进一步可选地，所述控制模块504，还用于在飞行器向目标设施飞行过程中，控制飞行器按照指定规则向目标设施飞行，所述指定规则用于指示所述飞行器飞行到用于获取深度图的至少两个能够以不同角度拍摄的拍摄位置；基于至少两个拍摄位置来获取所述飞行器在行进方向上的深度图；根据获取到的深度图进行飞行避障处理。
进一步可选地，所述控制模块504，还用于在完成对所述目标设施的检测后，根据预设的返航轨迹控制所述飞行器返回；所述预设的返航轨迹包括：记录的在所述飞行器飞行至所述检测位置之前的飞行轨迹。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个 或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。并且还能够自动返航和避障，进一步地满足了巡检的自动化、智能化需求，提高了巡检的安全性。
再请参见图6，是本发明实施例的一种控制设备的结构示意图，本发明实施例的所述控制设备包括供电电路，该控制设备可以由一块单独的电池供电，也可以通过一个供电接口由无人机等飞行器的电池供电。所述控制设备还可包括处理器601、数据接口602以及存储器603。
所述数据接口602主要用于与飞行器交互数据，或者进一步地，所述数据接口602还可以与地面的用于接收并显示由飞行器检测到的图像等数据的监控设备之间交互数据。
所述存储器603可以包括易失性存储器(volatile memory)，例如随机存取存储器(random-access memory，RAM)；存储器603也可以包括非易失性存储器(non-volatile memory)，例如快闪存储器(flash memory)，硬盘(hard disk drive，HDD)或固态硬盘(solid-state drive，SSD)；存储器603还可以包括上述种类的存储器的组合。
所述处理器601可以是中央处理器(central processing unit，CPU)。所述处理器601还可以进一步包括硬件芯片。上述硬件芯片可以是专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。上述PLD可以是复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)或其任意组合。
可选地，所述存储器603还用于存储程序指令。所述处理器601可以调用所述程序指令，实现如本申请图1，2，3和4所对应实施例中所示的设施检测方法。
在一个实施例中，所述处理器601，用于当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；根据所述飞行规则生成控制指令，并 通过所述数据接口602发送给所述飞行器以控制所述飞行器飞行，以便于完成对所述目标设施的检测。
可选地，所述飞行规则包括飞行轨迹，所述处理器601，用于获取与所述检测对象关联的巡检策略；根据所述检测对象在所述环境图像中的图像位置和所述检测位置，生成满足所述巡检策略的飞行轨迹。
可选地，得到的所述检测对象包括多个，每一个检测对象均关联了巡检策略，所述处理器601，用于获取每一个检测对象的巡检策略；根据各检测对象在所述环境图像中的图像位置和所述检测位置，生成飞行规则；其中，所述飞行规则中包括满足所有巡检策略的飞行轨迹，或者包括多段飞行轨迹，每一段飞行轨迹满足部分巡检策略。
可选地，生成的飞行规则还满足预设的限制参数；所述限制参数包括：飞行距离参数、飞行时长参数、飞行安全参数、能量损耗参数中的任意一种或多种。
可选地，所述处理器601，还用于生成检测参数，并通过所述数据接口602将所述检测参数发送给飞行器，所述检测参数用于在控制所述飞行器飞行的过程中指示所述飞行器对检测对象进行巡检的感测参数，所述感测参数包括：用于对检测对象进行检测的传感器的拍摄角度参数、用于对检测对象进行拍摄的拍摄机的拍摄参数。
可选地，所述处理器601，还用于在根据所述飞行规则控制所述飞行器飞行的过程中，获取检测得到的检测图像；根据检测图像中检测对象的位置和为检测对象设置的巡检策略，更新所述飞行规则；根据更新后的飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
可选地，所述处理器601，用于根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的初始飞行规则；获取所述飞行器的剩余能量值；根据所述剩余能量值对所述初始飞行规则进行调整，将调整后得到的规则作为关于所述检测对象飞行规则。
可选地，所述处理器601，用于获取为所述目标设施预设的对象模型；按照对象模型对图像区域进行图像分割，得到与所述对象模型之间的形状相似度满足相似度条件的检测对象。
可选地，所述对象模型配置有模型标识，根据模型标识获取与检测对象关 联的巡检策略。
可选地，所述处理器601，还用于在显示地图的交互界面上配置一个或者多个待检测的设施位置点；将被选中的一个或者多个设施位置点所对应的设施确定为目标设施；根据选中的设施位置点控制飞行器向目标设施飞行，以便于飞行到针对目标设施的检测位置。
可选地，所述处理器601，还用于接收所述飞行器返回的位置信息，所述位置信息包括：由所述飞行器生成的所述飞行器相对于目标对象的距离信息和方向信息，或由所述飞行器返回的所述飞行器的位置坐标信息。
可选地，所述处理器601，还用于在飞行器向目标设施飞行过程中，控制飞行器按照指定规则向目标设施飞行，所述指定规则用于指示所述飞行器飞行到用于获取深度图的至少两个能够以不同角度拍摄的拍摄位置；基于至少两个拍摄位置来获取所述飞行器在行进方向上的深度图；根据获取到的深度图进行飞行避障处理。
可选地，所述处理器601，还用于在完成对所述目标设施的检测后，根据预设的返航轨迹控制所述飞行器返回；所述预设的返航轨迹包括：记录的在所述飞行器飞行至所述检测位置之前的飞行轨迹。
本发明实施例在对某些设施进行巡检时，特别是对较高设施或者处于不易到达的地区的设施，能够基于图像识别和自动控制飞行的方式对设施中的一个或者多个需要检测的对象进行巡检，降低了巡检的人力成本以及安全性，提高了巡检的效率。并且还能够自动返航和避障，进一步地满足了巡检的自动化、智能化需求，提高了巡检的安全性。
以上所揭露的仅为本发明部分实施例而已，当然不能以此来限定本发明之权利范围，因此依本发明权利要求所作的等同变化，仍属本发明所涵盖的范围。

Claims (26)

  1. 一种基于飞行器的设施检测方法，其特征在于，包括：
    当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；
    从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；
    根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；
    根据所述飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
  2. 如权利要求1所述的方法，其特征在于，所述飞行规则包括飞行轨迹，所述根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则，包括：
    获取与所述检测对象关联的巡检策略；
    根据所述检测对象在所述环境图像中的图像位置和所述检测位置，生成满足所述巡检策略的飞行轨迹。
  3. 如权利要求1所述的方法，其特征在于，得到的所述检测对象包括多个，每一个检测对象均关联了巡检策略，所述根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则，包括：
    获取每一个检测对象的巡检策略；
    根据各检测对象在所述环境图像中的图像位置和所述检测位置，生成飞行规则；
    其中，所述飞行规则中包括满足所有巡检策略的飞行轨迹，或者包括多段飞行轨迹，每一段飞行轨迹满足所有巡检策略中的部分巡检策略。
  4. 如权利要求2或3所述的方法，其特征在于，生成的飞行规则还满足 预设的限制参数；所述限制参数包括：飞行距离参数、飞行时长参数、飞行安全参数、能量损耗参数中的任意一种或多种。
  5. 如权利要求1所述的方法，其特征在于，还包括：
    生成检测参数，所述检测参数包括用于在所述飞行器飞行的过程中指示所述飞行器对检测对象进行巡检的感测参数，所述感测参数包括：用于对检测对象进行检测的传感器的拍摄角度参数、用于对检测对象进行拍摄的拍摄机的拍摄参数。
  6. 如权利要求1-5任一项所述的方法，其特征在于，还包括：
    在根据所述飞行规则控制所述飞行器飞行的过程中，获取检测得到的检测图像；
    根据检测图像中检测对象的位置和为检测对象设置的巡检策略，更新所述飞行规则；
    根据更新后的飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
  7. 如权利要求1所述的方法，其特征在于，所述根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则，包括：
    根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的初始飞行规则；
    获取所述飞行器的剩余能量值；
    根据所述剩余能量值对所述初始飞行规则进行调整，将调整后得到的规则作为关于所述检测对象的飞行规则。
  8. 如权利要求1-7任一项所述的方法，其特征在于，所述对该图像区域进行图像分割，得到关于所述目标设施的检测对象，包括：
    获取为所述目标设施预设的对象模型；
    按照对象模型对图像区域进行图像分割，得到与所述对象模型之间的相似 度满足相似度条件的检测对象。
  9. 如权利要求8所述的方法，其特征在于，所述对象模型配置有模型标识，根据模型标识获取对应检测对象关联的巡检策略。
  10. 如权利要求1-9任一项所述的方法，其特征在于，还包括：
    在显示地图的交互界面上配置一个或者多个待检测的设施位置点；
    将被选中的一个或者多个设施位置点所对应的设施确定为目标设施；
    根据选中的设施位置点控制飞行器向目标设施飞行，以便于飞行到针对目标设施的检测位置。
  11. 如权利要求10所述的方法，其特征在于，还包括：
    接收所述飞行器返回的位置信息，所述位置信息包括：由所述飞行器生成的所述飞行器相对于目标对象的距离信息和方向信息，或由所述飞行器返回的所述飞行器的位置坐标信息。
  12. 如权利要求10所述的方法，其特征在于，还包括：
    在飞行器向目标设施飞行过程中，控制飞行器按照指定规则向目标设施飞行，所述指定规则用于指示所述飞行器飞行到用于获取深度图的至少两个能够以不同角度拍摄的拍摄位置；
    基于至少两个拍摄位置来获取所述飞行器在行进方向上的深度图；
    根据获取到的深度图进行飞行避障处理。
  13. 如权利要求1-12任一项所述的方法，其特征在于，还包括：
    在完成对所述目标设施的检测后，根据预设的返航轨迹控制所述飞行器返回；
    所述预设的返航轨迹包括：记录的在所述飞行器飞行至所述检测位置之前的飞行轨迹。
  14. 一种控制设备，其特征在于，包括：处理器和数据接口；
    所述数据接口，用于与飞行器交互数据；
    所述处理器，用于当飞行器位于针对目标设施的检测位置时，获取包括所述目标设施的环境图像；从所述环境图像中确定出所述目标设施所属的图像区域，并对该图像区域进行图像分割，得到关于所述目标设施的检测对象；根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的飞行规则；根据所述飞行规则生成控制指令，并通过所述数据接口发送给所述飞行器以控制所述飞行器飞行，以便于完成对所述目标设施的检测。
  15. 如权利要求14所述的装置，其特征在于，所述飞行规则包括飞行轨迹，所述处理器，用于获取与所述检测对象关联的巡检策略；根据所述检测对象在所述环境图像中的图像位置和所述检测位置，生成满足所述巡检策略的飞行轨迹。
  16. 如权利要求14所述的装置，其特征在于，得到的所述检测对象包括多个，每一个检测对象均关联了巡检策略，所述处理器，用于获取每一个检测对象的巡检策略；根据各检测对象在所述环境图像中的图像位置和所述检测位置，生成飞行规则；其中，所述飞行规则中包括满足所有巡检策略的飞行轨迹，或者包括多段飞行轨迹，每一段飞行轨迹满足所有巡检策略中的部分巡检策略。
  17. 如权利要求15或16所述的装置，其特征在于，生成的飞行规则还满足预设的限制参数；所述限制参数包括：飞行距离参数、飞行时长参数、飞行安全参数、能量损耗参数中的任意一种或多种。
  18. 如权利要求14所述的装置，其特征在于，
    所述处理器，还用于生成检测参数，并通过所述数据接口将所述检测参数发送给飞行器，所述检测参数用于在控制所述飞行器飞行的过程中指示所述飞行器对检测对象进行巡检的感测参数，所述感测参数包括：用于对检测对象进行检测的传感器的拍摄角度参数、用于对检测对象进行拍摄的拍摄机的拍摄参数。
  19. 如权利要求14-18任一项所述的装置，其特征在于，
    所述处理器，还用于在根据所述飞行规则控制所述飞行器飞行的过程中，获取检测得到的检测图像；根据检测图像中检测对象的位置和为检测对象设置的巡检策略，更新所述飞行规则；根据更新后的飞行规则控制所述飞行器飞行，以便于完成对所述目标设施的检测。
  20. 如权利要求14所述的装置，其特征在于，
    所述处理器，用于根据检测对象在所述环境图像中的图像位置和所述检测位置，获取关于所述检测对象的初始飞行规则；获取所述飞行器的剩余能量值；根据所述剩余能量值对所述初始飞行规则进行调整，将调整后得到的规则作为关于所述检测对象的飞行规则。
  21. 如权利要求14-20任一项所述的装置，其特征在于，
    所述处理器，用于获取为所述目标设施预设的对象模型；按照对象模型对图像区域进行图像分割，得到与所述对象模型之间的形状相似度满足相似度条件的检测对象。
  22. 如权利要求21所述的装置，其特征在于，所述对象模型配置有模型标识，根据模型标识获取对应检测对象关联的巡检策略。
  23. 如权利要求14-22任一项所述的装置，其特征在于，
    所述处理器，还用于在显示地图的交互界面上配置一个或者多个待检测的设施位置点；将被选中的一个或者多个设施位置点所对应的设施确定为目标设施；根据选中的设施位置点控制飞行器向目标设施飞行，以便于飞行到针对目标设施的检测位置。
  24. 如权利要求23所述的装置，其特征在于，
    所述处理器，还用于接收所述飞行器返回的位置信息，所述位置信息包括：由所述飞行器生成的所述飞行器相对于目标对象的距离信息和方向信息，或由 所述飞行器返回的所述飞行器的位置坐标信息。
  25. 如权利要求23所述的方法，其特征在于，
    所述处理器，还用于在飞行器向目标设施飞行过程中，控制飞行器按照指定规则向目标设施飞行，所述指定规则用于指示所述飞行器飞行到用于获取深度图的至少两个能够以不同角度拍摄的拍摄位置；基于至少两个拍摄位置来获取所述飞行器在行进方向上的深度图；
    根据获取到的深度图进行飞行避障处理。
  26. 如权利要求14-25任一项所述的装置，其特征在于，
    所述处理器，还用于在完成对所述目标设施的检测后，根据预设的返航轨迹控制所述飞行器返回；所述预设的返航轨迹包括：记录的在所述飞行器飞行至所述检测位置之前的飞行轨迹。
PCT/CN2017/082501 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备 WO2018195955A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN202111071109.0A CN113791641A (zh) 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备
CN201780004504.2A CN108496129B (zh) 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备
PCT/CN2017/082501 WO2018195955A1 (zh) 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/082501 WO2018195955A1 (zh) 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备

Publications (1)

Publication Number Publication Date
WO2018195955A1 true WO2018195955A1 (zh) 2018-11-01

Family

ID=63344773

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/082501 WO2018195955A1 (zh) 2017-04-28 2017-04-28 一种基于飞行器的设施检测方法及控制设备

Country Status (2)

Country Link
CN (2) CN108496129B (zh)
WO (1) WO2018195955A1 (zh)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110307837A (zh) * 2019-07-22 2019-10-08 张琦 一种基于图像识别的无人机导航系统及方法
CN112180955A (zh) * 2020-08-26 2021-01-05 国网安徽省电力有限公司淮南供电公司 一种自动巡检无人机的基于视觉反馈的二次复查方法和系统
CN112229845A (zh) * 2020-10-12 2021-01-15 国网河南省电力公司濮阳供电公司 基于视觉导航技术的无人机高精度饶塔智能巡检方法
CN112236727A (zh) * 2019-07-30 2021-01-15 深圳市大疆创新科技有限公司 无人机仿真飞行方法及装置、记录介质
CN112262319A (zh) * 2019-08-30 2021-01-22 深圳市大疆创新科技有限公司 电力线路的检测方法、毫米波雷达、系统和存储介质
CN112327920A (zh) * 2020-11-16 2021-02-05 国网新疆电力有限公司检修公司 一种无人机自主避障巡检路径规划方法及装置
CN112788292A (zh) * 2020-12-28 2021-05-11 深圳市朗驰欣创科技股份有限公司 巡检观测点的确定方法、装置、巡检机器人和存储介质
CN112823323A (zh) * 2020-05-06 2021-05-18 深圳市大疆创新科技有限公司 巡检方法、无人机、地面控制平台、系统及存储介质
CN113938609A (zh) * 2021-11-04 2022-01-14 中国联合网络通信集团有限公司 区域监控方法、装置及设备
CN114636424A (zh) * 2019-02-21 2022-06-17 国网浙江平湖市供电有限公司 基于可穿戴设备的变电站巡检路径规划方法
CN115052133A (zh) * 2022-07-06 2022-09-13 国网江苏省电力有限公司南通市通州区供电分公司 一种基于无人机的配电台架验收方法
CN115275870A (zh) * 2022-09-28 2022-11-01 合肥优晟电力科技有限公司 一种基于高空线路维护的巡检系统
EP4145238A1 (de) * 2021-09-07 2023-03-08 Spleenlab GmbH Verfahren zum steuern eines unbemannten fluggeräts für einen inspektionsflug zum inspizieren eines objekts und unbemanntes inspektionsfluggerät
CN116225062A (zh) * 2023-03-14 2023-06-06 广州天勤数字科技有限公司 一种应用于桥梁巡检的无人机导航方法及无人机
CN116878518A (zh) * 2023-09-06 2023-10-13 滨州市华亿电器设备有限公司 用于城市输电线路维护的无人机巡检路径规划方法
CN116909318A (zh) * 2023-09-14 2023-10-20 众芯汉创(江苏)科技有限公司 一种基于高精度三维点云的无人机自主巡检航线规划系统
CN117389293A (zh) * 2023-10-31 2024-01-12 广州天海翔航空科技有限公司 巡检无人机飞行控制管理方法及系统

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109471447A (zh) * 2018-12-14 2019-03-15 国网冀北电力有限公司检修分公司 无人机导航方法、装置、无人机和数据可读存储装置
CN109358650B (zh) * 2018-12-14 2022-11-18 国网冀北电力有限公司检修分公司 巡检路径规划方法、装置、无人机和计算机可读存储介质
CN109885083A (zh) * 2019-03-06 2019-06-14 国网陕西省电力公司检修公司 基于激光雷达的输电线路精细化巡检飞行平台及巡检方法
CN109885098B (zh) * 2019-04-11 2022-02-11 株洲时代电子技术有限公司 一种桥梁边栏巡检航线规划方法
WO2021056144A1 (zh) * 2019-09-23 2021-04-01 深圳市大疆创新科技有限公司 可移动平台的返航控制方法、装置及可移动平台
CN111401146A (zh) * 2020-02-26 2020-07-10 长江大学 一种无人机电力巡检方法、设备及存储介质
CN111582117A (zh) * 2020-04-29 2020-08-25 长江大学 一种无人机违章建筑物巡检方法、设备及存储介质
CN113741413B (zh) * 2020-05-29 2022-11-08 广州极飞科技股份有限公司 一种无人设备的作业方法、无人设备及存储介质
CN112068591A (zh) * 2020-08-25 2020-12-11 中国南方电网有限责任公司超高压输电公司天生桥局 输电线路自动巡检无人机、控制方法、装置和存储介质
CN112233270A (zh) * 2020-10-30 2021-01-15 国家电网有限公司 一种无人机自主智能绕塔巡检系统
CN113014904A (zh) * 2021-02-24 2021-06-22 苏州臻迪智能科技有限公司 一种无人机巡检图像处理的方法、装置、系统和存储介质
CN113112098A (zh) * 2021-05-12 2021-07-13 上海野戈智能科技发展有限公司 建筑物缺陷的检测方法及装置
CN113625730B (zh) * 2021-06-30 2023-07-14 南京邮电大学 一种基于超扭滑模的四旋翼自适应容错控制方法
CN113872680B (zh) * 2021-12-03 2022-04-26 特金智能科技(上海)有限公司 Tdoa辅助rid信号接收的控制方法、装置、系统
CN114326794A (zh) * 2021-12-13 2022-04-12 广东省建设工程质量安全检测总站有限公司 幕墙缺陷的识别方法、控制终端、服务器及可读存储介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103196430A (zh) * 2013-04-27 2013-07-10 清华大学 基于无人机的飞行轨迹与视觉信息的映射导航方法及系统
CN104035446A (zh) * 2014-05-30 2014-09-10 深圳市大疆创新科技有限公司 无人机的航向生成方法和系统
CN105023014A (zh) * 2015-08-21 2015-11-04 马鞍山市安工大工业技术研究院有限公司 一种无人机巡检输电线路图像内的杆塔目标提取方法
US20150331711A1 (en) * 2014-05-19 2015-11-19 Qualcomm Incorporated Systems and methods for context-aware application control
CN105551032A (zh) * 2015-12-09 2016-05-04 国网山东省电力公司电力科学研究院 一种基于视觉伺服的杆塔图像采集系统及其方法
CN105678289A (zh) * 2016-03-07 2016-06-15 谭圆圆 一种无人飞行器的控制方法及装置
CN105955308A (zh) * 2016-05-20 2016-09-21 腾讯科技（深圳）有限公司 一种飞行器的控制方法和装置
CN106054924A (zh) * 2016-07-06 2016-10-26 北京北方猎天科技有限公司 一种无人机伴飞方法、伴飞装置和伴飞系统

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101477169B (zh) * 2009-01-16 2011-07-13 华北电力大学 巡检飞行机器人对电力线路的检测方法
CN102589524B (zh) * 2011-01-13 2014-01-08 国家电网公司 一种电力线路巡检方法
KR20130127822A (ko) * 2012-05-15 2013-11-25 한국전자통신연구원 도로상 물체 분류 및 위치검출을 위한 이종 센서 융합처리 장치 및 방법
CN102941920A (zh) * 2012-12-05 2013-02-27 南京理工大学 一种基于多旋翼飞行器的高压输电线路巡检机器人及其方法
CN104049641B (zh) * 2014-05-29 2017-08-25 深圳市大疆创新科技有限公司 一种自动降落方法、装置及飞行器
CN104298248B (zh) * 2014-10-08 2018-02-13 南京航空航天大学 旋翼无人机精确视觉定位定向方法
CN106468918B (zh) * 2015-08-18 2020-03-20 航天图景（北京）科技有限公司 一种线路巡检的标准化数据采集方法及系统
CN205920057U (zh) * 2015-09-29 2017-02-01 柳州欧维姆机械股份有限公司 检测结构件表面裂缝的多旋翼无人机检测平台系统
CN105787447A (zh) * 2016-02-26 2016-07-20 深圳市道通智能航空技术有限公司 一种无人机基于双目视觉的全方位避障的方法及系统
CN106127788B (zh) * 2016-07-04 2019-10-25 触景无限科技（北京）有限公司 一种视觉避障方法和装置
CN106054931B (zh) * 2016-07-29 2019-11-05 北方工业大学 一种基于视觉定位的无人机定点飞行控制系统
CN106228862A (zh) * 2016-09-28 2016-12-14 国家电网公司 一种输电线路无人机巡检仿真培训方法
CN106354156A (zh) * 2016-09-29 2017-01-25 腾讯科技（深圳）有限公司 一种跟踪目标对象的方法、装置及飞行器
CN106504362A (zh) * 2016-10-18 2017-03-15 国网湖北省电力公司检修公司 基于无人机的输变电系统巡检方法
CN106595631B (zh) * 2016-10-25 2019-08-23 纳恩博（北京）科技有限公司 一种躲避障碍物的方法及电子设备
CN106477038B (zh) * 2016-12-20 2018-12-25 北京小米移动软件有限公司 图像拍摄方法及装置、无人机

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103196430A (zh) * 2013-04-27 2013-07-10 清华大学 基于无人机的飞行轨迹与视觉信息的映射导航方法及系统
US20150331711A1 (en) * 2014-05-19 2015-11-19 Qualcomm Incorporated Systems and methods for context-aware application control
CN104035446A (zh) * 2014-05-30 2014-09-10 深圳市大疆创新科技有限公司 无人机的航向生成方法和系统
CN105023014A (zh) * 2015-08-21 2015-11-04 马鞍山市安工大工业技术研究院有限公司 一种无人机巡检输电线路图像内的杆塔目标提取方法
CN105551032A (zh) * 2015-12-09 2016-05-04 国网山东省电力公司电力科学研究院 一种基于视觉伺服的杆塔图像采集系统及其方法
CN105678289A (zh) * 2016-03-07 2016-06-15 谭圆圆 一种无人飞行器的控制方法及装置
CN105955308A (zh) * 2016-05-20 2016-09-21 腾讯科技（深圳）有限公司 一种飞行器的控制方法和装置
CN106054924A (zh) * 2016-07-06 2016-10-26 北京北方猎天科技有限公司 一种无人机伴飞方法、伴飞装置和伴飞系统

Cited By (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114636424B (zh) * 2019-02-21 2024-04-19 国网浙江省电力有限公司平湖市供电公司 基于可穿戴设备的变电站巡检路径规划方法
CN114636424A (zh) * 2019-02-21 2022-06-17 国网浙江平湖市供电有限公司 基于可穿戴设备的变电站巡检路径规划方法
CN110307837A (zh) * 2019-07-22 2019-10-08 张琦 一种基于图像识别的无人机导航系统及方法
CN112236727A (zh) * 2019-07-30 2021-01-15 深圳市大疆创新科技有限公司 无人机仿真飞行方法及装置、记录介质
CN112262319A (zh) * 2019-08-30 2021-01-22 深圳市大疆创新科技有限公司 电力线路的检测方法、毫米波雷达、系统和存储介质
CN112823323A (zh) * 2020-05-06 2021-05-18 深圳市大疆创新科技有限公司 巡检方法、无人机、地面控制平台、系统及存储介质
CN112180955A (zh) * 2020-08-26 2021-01-05 国网安徽省电力有限公司淮南供电公司 一种自动巡检无人机的基于视觉反馈的二次复查方法和系统
CN112180955B (zh) * 2020-08-26 2024-02-20 国网安徽省电力有限公司淮南供电公司 自动巡检无人机的基于视觉反馈的二次复查方法和系统
CN112229845A (zh) * 2020-10-12 2021-01-15 国网河南省电力公司濮阳供电公司 基于视觉导航技术的无人机高精度饶塔智能巡检方法
CN112327920A (zh) * 2020-11-16 2021-02-05 国网新疆电力有限公司检修公司 一种无人机自主避障巡检路径规划方法及装置
CN112327920B (zh) * 2020-11-16 2023-07-14 国网新疆电力有限公司检修公司 一种无人机自主避障巡检路径规划方法及装置
CN112788292B (zh) * 2020-12-28 2023-05-26 深圳市朗驰欣创科技股份有限公司 巡检观测点的确定方法、装置、巡检机器人和存储介质
CN112788292A (zh) * 2020-12-28 2021-05-11 深圳市朗驰欣创科技股份有限公司 巡检观测点的确定方法、装置、巡检机器人和存储介质
EP4145238A1 (de) * 2021-09-07 2023-03-08 Spleenlab GmbH Verfahren zum steuern eines unbemannten fluggeräts für einen inspektionsflug zum inspizieren eines objekts und unbemanntes inspektionsfluggerät
CN113938609B (zh) * 2021-11-04 2023-08-22 中国联合网络通信集团有限公司 区域监控方法、装置及设备
CN113938609A (zh) * 2021-11-04 2022-01-14 中国联合网络通信集团有限公司 区域监控方法、装置及设备
CN115052133A (zh) * 2022-07-06 2022-09-13 国网江苏省电力有限公司南通市通州区供电分公司 一种基于无人机的配电台架验收方法
CN115052133B (zh) * 2022-07-06 2023-09-12 国网江苏省电力有限公司南通市通州区供电分公司 一种基于无人机的配电台架验收方法
CN115275870B (zh) * 2022-09-28 2022-12-06 合肥优晟电力科技有限公司 一种基于高空线路维护的巡检系统
CN115275870A (zh) * 2022-09-28 2022-11-01 合肥优晟电力科技有限公司 一种基于高空线路维护的巡检系统
CN116225062B (zh) * 2023-03-14 2024-01-16 广州天勤数字科技有限公司 一种应用于桥梁巡检的无人机导航方法及无人机
CN116225062A (zh) * 2023-03-14 2023-06-06 广州天勤数字科技有限公司 一种应用于桥梁巡检的无人机导航方法及无人机
CN116878518B (zh) * 2023-09-06 2023-11-21 滨州市华亿电器设备有限公司 用于城市输电线路维护的无人机巡检路径规划方法
CN116878518A (zh) * 2023-09-06 2023-10-13 滨州市华亿电器设备有限公司 用于城市输电线路维护的无人机巡检路径规划方法
CN116909318B (zh) * 2023-09-14 2023-11-24 众芯汉创(江苏)科技有限公司 一种基于高精度三维点云的无人机自主巡检航线规划系统
CN116909318A (zh) * 2023-09-14 2023-10-20 众芯汉创(江苏)科技有限公司 一种基于高精度三维点云的无人机自主巡检航线规划系统
CN117389293A (zh) * 2023-10-31 2024-01-12 广州天海翔航空科技有限公司 巡检无人机飞行控制管理方法及系统
CN117389293B (zh) * 2023-10-31 2024-05-24 广州天海翔航空科技有限公司 巡检无人机飞行控制管理方法及系统

Also Published As

Publication number Publication date
CN108496129A (zh) 2018-09-04
CN113791641A (zh) 2021-12-14
CN108496129B (zh) 2021-10-01

Similar Documents

Publication Publication Date Title
WO2018195955A1 (zh) 一种基于飞行器的设施检测方法及控制设备
CN111512256B (zh) 自动和自适应的三维机器人现场勘测
JP7030431B2 (ja) 点検支援システム及び点検支援制御プログラム
JP7263630B2 (ja) 無人航空機による３次元再構成の実行
EP3818337B1 (en) Defect detection system using a camera equipped uav for building facades on complex asset geometry with optimal automatic obstacle deconflicted flightpath
US11880118B2 (en) Method of determining a path along an object, system and method for automatically inspecting an object
CN110142785A (zh) 一种基于目标检测的巡检机器人视觉伺服方法
CN108508916B (zh) 一种无人机编队的控制方法、装置、设备及存储介质
CN109373975A (zh) 检测车辆控制设备、控制方法和计算机程序
JP2006027448A (ja) 無人飛行体を利用した空撮方法及びその装置
JP6039050B1 (ja) 無人機を用いた構造物等の検査方法
JP2016177640A (ja) 映像監視システム
CN105526916A (zh) 动态图像遮蔽系统和方法
Savva et al. ICARUS: Automatic autonomous power infrastructure inspection with UAVs
US20220221857A1 (en) Information processing apparatus, information processing method, program, and information processing system
JP6482856B2 (ja) 監視システム
KR101118926B1 (ko) 주, 야간 겸용 이동물체 관측시스템
CN205721829U (zh) 一种无人飞行器
JP7437930B2 (ja) 移動体及び撮像システム
JP7314466B2 (ja) 無人航空機、画像撮像システム、及びプログラム
CN109144098A (zh) 一种无人机楼梯自动巡检方法
CN115158606B (zh) 一种水面救生方法、系统和介质
Ali Application of Quad-Copter Target Tracking Using Mask Based Edge Detection for Feasibility of Wind Turbine Blade Inspection During Uninterrupted Operation
JP2021086268A (ja) 移動体、情報処理装置、及び撮像システム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17907143

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17907143

Country of ref document: EP

Kind code of ref document: A1