[go: up one dir, main page]

WO2022087841A1 - 用于农业无人飞行器仿地作业的方法和农业无人飞行器 - Google Patents

用于农业无人飞行器仿地作业的方法和农业无人飞行器 Download PDF

Info

Publication number
WO2022087841A1
WO2022087841A1 PCT/CN2020/124082 CN2020124082W WO2022087841A1 WO 2022087841 A1 WO2022087841 A1 WO 2022087841A1 CN 2020124082 W CN2020124082 W CN 2020124082W WO 2022087841 A1 WO2022087841 A1 WO 2022087841A1
Authority
WO
WIPO (PCT)
Prior art keywords
unmanned aerial
aerial vehicle
point cloud
agricultural unmanned
canopy
Prior art date
Application number
PCT/CN2020/124082
Other languages
English (en)
French (fr)
Inventor
祝煌剑
王璐
王春明
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 PCT/CN2020/124082 priority Critical patent/WO2022087841A1/zh
Priority to CN202080030164.2A priority patent/CN113966514A/zh
Publication of WO2022087841A1 publication Critical patent/WO2022087841A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining

Definitions

  • the present application relates to the technical field of agricultural unmanned aerial vehicles, and more particularly to a method for imitating the ground of agricultural unmanned aerial vehicles and agricultural unmanned aerial vehicles.
  • the existing agricultural unmanned aerial vehicle also has real-time perception of the operation scene based on machine vision, and realizes the autonomous ground imitation operation in the fruit tree scene.
  • the disadvantage of this technical solution is that the machine vision is easily affected by the environment, especially the liquid medicine, dust mist, etc. in the operating environment of the agricultural unmanned aerial vehicle.
  • the present application provides a method and an agricultural unmanned aerial vehicle for the ground imitation operation of an agricultural unmanned aerial vehicle, which can ensure the sensitive ground imitation operation of the agricultural unmanned aerial vehicle in the vegetation canopy, and improve the operation safety and operation efficiency.
  • an embodiment of the present application provides a method for an agricultural unmanned aerial vehicle (UAV) to simulate the ground.
  • the method includes: acquiring real-time point cloud data of a to-be-operated area in the flight direction of the agricultural UAV;
  • the point cloud data is fitted to obtain the canopy envelope of the vegetation in the to-be-operated area;
  • the terrain prediction is performed on the to-be-operated area based on the canopy envelope, and the terrain prediction result is used to control the Agricultural unmanned aerial vehicles perform ground imitation operations.
  • an embodiment of the present application provides an agricultural unmanned aerial vehicle, the agricultural unmanned aerial vehicle includes a processor and a memory, the memory is used for storing instructions, and the processor calls the instructions stored in the memory for Perform the following operations: obtain real-time point cloud data of the area to be operated in the flight direction of the agricultural unmanned aerial vehicle; fit the real-time point cloud data to obtain the canopy envelope of the vegetation in the area to be operated; The canopy envelope performs terrain prediction on the to-be-operated area, and controls the agricultural unmanned aerial vehicle to perform terrain imitation operations based on the terrain prediction result.
  • embodiments of the present application provide a computer-readable storage medium, where the computer-readable storage medium includes instructions that, when executed on a computer, cause the computer to execute the above-mentioned simulation for agricultural unmanned aerial vehicles. method of work.
  • the method for imitating the ground of an agricultural unmanned aerial vehicle and the agricultural unmanned aerial vehicle detect the surrounding environment in real time, intelligently perceive changes in vegetation and terrain height and drop, and ensure that the agricultural unmanned aerial vehicle operates sensitively and imitating the ground in the vegetation canopy , and it will not drop at will when encountering open space and empty space, thereby improving operation safety and operation efficiency.
  • FIG. 1 shows a schematic diagram of an example of a conventional agricultural unmanned aerial vehicle that simulates the ground.
  • FIG. 2 shows a schematic diagram of another example of the conventional agricultural unmanned aerial vehicle ground-mimicking operation.
  • FIG. 3 shows a schematic flow chart of a method for an agricultural unmanned aerial vehicle to simulate the ground according to an embodiment of the present application.
  • FIG. 4 shows an exemplary schematic diagram of noise removal in the method for imitating the ground of an agricultural unmanned aerial vehicle according to an embodiment of the present application.
  • FIG. 5 shows an exemplary effect diagram after noise removal in the method for imitating the ground of an agricultural unmanned aerial vehicle according to an embodiment of the present application.
  • FIG. 6 shows an exemplary schematic diagram of canopy envelope fitting in the method for agricultural unmanned aerial vehicle ground-mimicking operations according to an embodiment of the present application.
  • FIG. 7 shows a schematic block diagram of an agricultural unmanned aerial vehicle according to an embodiment of the present application.
  • the existing agricultural unmanned aerial vehicles have some problems when operating in the fruit tree scene. For example, for the scheme of using a surveying and mapping machine to generate the operation map in advance, it is not only costly but also time-consuming; for the scheme of real-time perception of the operation scene, there is a poor terrain following effect. , high risk of flight safety.
  • FIG. 1 shows a schematic diagram of an example of the existing agricultural unmanned aerial vehicle ground-mimicking operation.
  • the agricultural UAV performs ground imitation operations based on flat-panel radar.
  • the flat-panel radar perceives the height of the object directly under the flight of the agricultural UAV relative to the agricultural UAV in real time, and the object directly under the flight may be a canopy of vegetation (such as a tree), or it may be a canopy of non-vegetation, such as an open space or a vacant space;
  • the flat-panel radar perceives the height of the vegetation canopy relative to the agricultural UAV in real time, it is still possible to fly and operate safely at a certain distance above the canopy on this basis;
  • the height of the unmanned aerial vehicle is high, it will still fly at a certain distance above the empty space and the open space according to the original safe height distance standard, which will cause the agricultural unmanned aerial vehicle to descend blindly, which may lead to flight safety problems and affect the operation efficiency.
  • FIG. 2 shows a schematic diagram of an example of a conventional agricultural unmanned aerial vehicle that simulates the ground.
  • the agricultural UAV performs ground imitation operations based on rotating radar.
  • the rotating radar perceives the angle ( ⁇ as shown in Fig. 2) and distance (L as shown in Fig. 2) of the object under the flight of the agricultural UAV relative to the agricultural UAV in real time.
  • ⁇ as shown in Fig. 2
  • L distance
  • the rotating radar perceives the distance between the airspace and the airspace relative to the agricultural UAV, it will still fly at a certain distance above the airspace and the airspace according to the original safe height distance standard, which leads to the agricultural UAV. Blindly descending the ground may cause flight safety problems and affect operational efficiency.
  • the present application provides a solution for imitating the ground operation of agricultural unmanned aerial vehicles. A detailed description will be given below with reference to FIGS. 3 to 7 .
  • FIG. 3 shows a schematic flowchart of a method 300 for an agricultural unmanned aerial vehicle to simulate the ground according to an embodiment of the present application.
  • a method 300 for an agricultural unmanned aerial vehicle to simulate the ground according to an embodiment of the present application may include:
  • S310 Acquire real-time point cloud data of the area to be operated in the flight direction of the agricultural unmanned aerial vehicle.
  • S320 Fitting the real-time point cloud data to obtain the canopy envelope of the vegetation in the to-be-operated area.
  • the canopy envelope of the vegetation (such as fruit trees) in the to-be-operated area is generated based on the fitting of point cloud data collected in real time on the to-be-operated area in the flight direction of the agricultural unmanned aerial vehicle (such as a plant protection aircraft), wherein
  • the canopy envelope is a curve or surface that simulates the change in the height of the top layer of vegetation, which reflects the relationship between the horizontal distance of the top layer of vegetation relative to the agricultural UAV and the vertical distance relative to the agricultural UAV.
  • terrain prediction is performed based on the obtained canopy envelope to control the agricultural UAV's ground imitation operation, which makes both the flight control and the ground imitation operation of the agricultural UAV be performed above the vegetation canopy, without Blindly descending on the ground when encountering an empty space or open space between vegetation can ensure that the agricultural unmanned aerial vehicle can operate sensitively in the vegetation canopy, improve the sensitivity of the ground imitation, and improve the safety and efficiency of operation.
  • real-time point cloud data may be collected for the to-be-operated area in the flight direction of the agricultural unmanned aerial vehicle based on sensors installed on the agricultural unmanned aerial vehicle.
  • sensors mounted on agricultural UAVs may include rotating millimeter wave radars, flat panel radars, vision sensors, and the like. Based on the real-time point cloud data of the to-be-operated area in the flight direction of the agricultural UAV collected by the sensor, the canopy envelope of the vegetation in the to-be-operated area can be calculated.
  • the acquired real-time point cloud data of the to-be-operated area in the flight direction of the agricultural unmanned aerial vehicle it can be preprocessed first, and then the preprocessed real-time point cloud data can be fitted to obtain the real-time point cloud data. Generate the canopy envelope of the vegetation in the area to be operated to improve the accuracy of the fitting results.
  • the preprocessing performed on the acquired real-time point cloud data may include coordinate system transformation.
  • the acquired real-time point cloud data can be converted from a sensor coordinate system (for example, a radar coordinate system as shown in FIG. 2, that is, a polar coordinate system) to a Cartesian coordinate system to obtain coordinate-transformed data, the coordinates
  • the coordinates of each point cloud point in the transformed data indicate the horizontal and vertical distances of the point cloud point relative to the agricultural UAV.
  • the coordinate transformation can be done with the following formula:
  • X i and Y i represent the point cloud data after coordinate transformation, where X i represents the horizontal distance of the reflective target relative to the sensor carrier (agricultural unmanned aerial vehicle), and Y i represents the reflective target relative to the sensor carrier ( L and ⁇ represent the point cloud data before coordinate transformation, where L represents the radial distance of the reflective target relative to the sensor, and ⁇ represents the azimuth angle of the reflective target relative to the sensor.
  • the coordinate transformation of the acquired real-time point cloud data can directly reflect the horizontal and vertical distances of the reflective target relative to the agricultural UAV.
  • the fitted canopy envelope is easier to treat the position of interest in the operation area. Make height prediction and/or slope prediction, so that it is easy to realize sensitive control of the ground-flight operation of agricultural unmanned aerial vehicle.
  • the acquired real-time point cloud data can also be culled to eliminate the error data (for example, the point cloud data reflected from the utility poles existing in the vegetation area and the like that affect the vegetation canopy package). Then, the real-time point cloud data after noise removal is fitted to generate the canopy envelope of the vegetation in the to-be-operated area, so as to improve the accuracy of the fitting results.
  • noise can be removed from the data after coordinate transformation to obtain valid data for fitting.
  • noise removal may be performed based on a method similar to morphological filtering or other suitable methods.
  • the noise removal may include taking each point cloud point in the coordinate-transformed data as a point to be detected and performing the following operations: acquiring a predetermined number of point cloud points closest to the point to be detected in the horizontal direction ; Calculate the mean or median value of the vertical distance of the predetermined number of point cloud points relative to the agricultural unmanned aerial vehicle as a reference value; Calculate the vertical distance of the point to be detected relative to the agricultural unmanned aerial vehicle and the reference value.
  • the predetermined number of point cloud points closest to the point to be detected in the horizontal direction may include: a predetermined number of point cloud points on one side of the point to be detected in the horizontal direction (generally applicable to the point to be detected) There are no point cloud points on the other side, or the number of point cloud points on the other side does not reach the predetermined number); or point cloud points on both sides of the point to be detected in the horizontal direction, the number of point cloud points on each side is half of the predetermined number. It will be described below with reference to FIG. 4 .
  • FIG. 4 shows an exemplary schematic diagram of noise removal in the method for imitating the ground of an agricultural unmanned aerial vehicle according to an embodiment of the present application.
  • each point in the coordinate system is a point cloud point.
  • the four (example) points closest to P1 in the horizontal direction are obtained. point), then you can calculate the average or median of the vertical distances of these four points relative to the agricultural UAV (that is, the ordinate of these four points) as a reference value, and then calculate the current point P1 to be detected relative to the agricultural unmanned aerial vehicle.
  • the difference between the vertical distance of the human aircraft (that is, the ordinate of P1) and the reference value if the difference is large (greater than or equal to a certain threshold), it is considered that the current point P1 to be detected is a noise point and will be eliminated. If it is larger (less than a certain threshold), it is considered that the current point P1 to be detected is a valid point to be reserved.
  • the obtained point cloud data is used for fitting the vegetation canopy envelope.
  • the fitting of the vegetation canopy envelope may be performed based on any one of the following fitting methods: fitting based on linear interpolation, polynomial fitting, fitting based on Gaussian regression, fitting based on b Spline curve (BSPLINE) fitting, support vector machine (SVM) based fitting.
  • BSPLINE b Spline curve
  • SVM support vector machine
  • (x 0 , y 0 ), (x 1 , y 1 ), (x 2 , y 2 ), ..., (x n-1 , y n-1 ), (x n , y n ) refers to the actually collected point cloud data (such as the data after the aforementioned coordinate transformation and noise removal), which are points obtained by interpolation, and each interpolation point can be calculated from two collected points.
  • the resulting canopy envelope is continuous data, as shown in the exemplary schematic diagram of canopy envelope fitting shown in FIG. 6 .
  • the manner of using linear interpolation for fitting is only exemplary, and other suitable manners may also be used for fitting.
  • the fitting method of linear interpolation is simple and computationally simple, and the fitting accuracy is affected by the distribution of the original point cloud points; it can also be fitted by other methods such as polynomial fitting, BSPLINE, Gauss regression, SVM, etc. , to improve the fitting accuracy.
  • terrain prediction can be performed on the to-be-operated area based on the canopy envelope, and the agricultural unmanned aerial vehicle can be controlled to perform ground imitation operations based on the terrain prediction result.
  • performing terrain prediction on the to-be-operated area based on the canopy envelope may include: obtaining, based on the canopy envelope, a height prediction result and/or a slope prediction result of the position of interest in the to-be-operated area .
  • acquiring the height prediction result of the position of interest in the to-be-operated area based on the canopy envelope may include: acquiring the horizontal distance of the position of interest relative to the agricultural UAV ; Substitute the horizontal distance into the canopy envelope to obtain the height prediction result of the position of interest.
  • the obtained canopy envelope of vegetation is continuous data, so the horizontal distance of any position from the agricultural UAV is substituted into the canopy envelope, and the vertical position relative to the agricultural UAV can be obtained. distance (altitude condition).
  • obtaining the slope prediction result of the position of interest in the to-be-operated area based on the canopy envelope may include: obtaining the horizontal distance of the position of interest relative to the agricultural unmanned aerial vehicle ; Calculate the derivative of the canopy envelope at the horizontal distance to obtain the slope prediction result of the position of interest. Since the canopy envelope of the vegetation in the to-be-operated area in the flight direction of the agricultural UAV has been obtained, in order to obtain the slope at a certain position in the to-be-operated area, the canopy envelope can be derived. For example, the horizontal distance at this position relative to the agricultural UAV is denoted as xi , and the slope prediction result at this position can be obtained by calculating the derivative of the canopy envelope at xi .
  • obtaining the slope prediction result of the position of interest in the to-be-operated area based on the canopy envelope may include: obtaining a relative value of the position of interest relative to the agricultural unmanned aerial vehicle. horizontal distance; obtain a predetermined number of point cloud points closest to the position of interest in the horizontal direction; perform fitting on the predetermined number of point cloud points to obtain a fitting function; calculate the fitting function in the horizontal The derivative at the distance is obtained to obtain the slope prediction result of the position of interest.
  • the horizontal distance relative to the agricultural unmanned aerial vehicle at a certain position is expressed as xi
  • the horizontal distance of the points in the vicinity of the position relative to the agricultural unmanned aerial vehicle can be obtained, and then a plurality of sampling points relative to the agricultural unmanned aerial vehicle can be obtained.
  • the set of horizontal distances ⁇ x in ,...,x i-1 ,x i ,x i+1 ,...,x i+n ⁇ substitute this set into the canopy envelope, and get the relative Sets ⁇ y in ,...,y i-1 ,y i ,y i+1 ,...,y i+n ⁇ of the vertical distances of agricultural UAVs, and fit based on these two sets, A new fitting equation is obtained, and the derivative of the new fitting equation at x i is obtained, that is, the slope prediction result at the position is obtained.
  • the vertical distances of multiple positions in the vicinity of the location of interest relative to the agricultural UAV are obtained based on the obtained canopy envelope, and then the horizontal distances relative to the agricultural UAV at these positions in the area and Refitting the vertical distance to a more accurate canopy envelope for this region and then derivation yields a more accurate slope prediction at the location of interest.
  • the position of interest is within the observation range of the sensor of the agricultural unmanned aerial vehicle.
  • the height/slope prediction result of the position of interest is obtained mainly based on the height interpolation of the adjacent points of the position of interest.
  • a certain algorithm can be used to extrapolate to obtain the height/slope prediction result of the position of interest.
  • the speed of the agricultural UAV in the vertical direction can be adjusted, so as to achieve more sensitive ground imitation operations while ensuring flight safety.
  • the method for imitating the ground of agricultural unmanned aerial vehicle and the agricultural unmanned aerial vehicle detect the surrounding environment in real time, intelligently sense vegetation and changes in terrain height and drop, and ensure that the agricultural unmanned aerial vehicle is in the vegetation crown It can operate on the ground flexibly, and it will not drop at will when encountering open space and empty space, thereby improving the safety and efficiency of operation.
  • costs are reduced because there is no need to map the work environment in advance with a mapping machine.
  • FIG. 7 shows a schematic block diagram of an agricultural unmanned aerial vehicle 700 according to an embodiment of the present application.
  • the agricultural UAV 700 includes a memory 710 and a processor 720 .
  • the memory 710 stores program instructions for implementing the corresponding steps in the method for agricultural unmanned aerial vehicle ground-mimicking operations according to the embodiments of the present application.
  • the processor 720 is configured to run the program stored in the memory 710 to execute the corresponding steps of the method for an agricultural unmanned aerial vehicle (UAV) simulation operation according to the embodiment of the present application.
  • UAV agricultural unmanned aerial vehicle
  • the processor 720 invokes the instructions stored in the memory 710 to perform the following operations: acquire real-time point cloud data of the area to be operated in the flight direction of the agricultural unmanned aerial vehicle 700; perform simulation on the real-time point cloud data to obtain the canopy envelope of the vegetation in the to-be-operated area; perform terrain prediction for the to-be-operated area based on the canopy envelope, and control the agricultural unmanned aerial vehicle 700 to simulate the terrain based on the result of the terrain prediction ground work.
  • the processor 720 further performs the following operations: after acquiring the real-time point cloud data, preprocessing the real-time point cloud data to obtain preprocessed data, and the fitting is based on performed on the preprocessed data.
  • the preprocessing performed by the processor 720 includes converting the real-time point cloud data from a sensor coordinate system to a Cartesian coordinate system to obtain coordinate-transformed data, the coordinate-transformed data
  • the coordinates of each point cloud point in indicate the horizontal and vertical distances of the point cloud point relative to the agricultural UAV 700 .
  • the preprocessing performed by the processor 720 further includes removing noise on the coordinate-transformed data to obtain valid data for performing the fitting.
  • the noise removal performed by the processor 720 is performed based on morphological filtering.
  • the noise removal performed by the processor 720 includes taking each point cloud point in the coordinate-transformed data as a point to be detected and performing the following operations: obtaining the distance in the horizontal direction from the A predetermined number of point cloud points that are closest to the point to be detected; calculate the average or median value of the vertical distance of the predetermined number of point cloud points relative to the agricultural UAV 700 as a reference value; The difference between the vertical distance of the UAV 700 and the reference value; when the difference is greater than or equal to a predetermined threshold, the to-be-detected point is removed as a noise point; when the difference is less than the predetermined threshold When the threshold is set, the to-be-detected point is reserved as a valid point.
  • the predetermined number of point cloud points closest to the point to be detected in the horizontal direction include: a predetermined number of point cloud points on one side of the point to be detected in the horizontal direction; Or point cloud points on both sides of the point to be detected in the horizontal direction, the number of point cloud points on each side is half of the predetermined number.
  • the fitting includes any one of the following ways: fitting based on linear interpolation, polynomial fitting, fitting based on Gaussian regression, fitting based on b-spline curve, and support based on Fitting of a vector machine.
  • the performing terrain prediction on the to-be-operated area based on the canopy envelope performed by the processor 720 includes: acquiring, based on the canopy envelope, the area of interest in the to-be-operated area The altitude prediction and/or the slope prediction for the location.
  • the obtaining the height prediction result of the position of interest in the to-be-operated area based on the canopy envelope performed by the processor 720 includes: obtaining the position of interest relative to the agricultural unmanned area The horizontal distance of the aircraft 700; the horizontal distance is substituted into the canopy envelope to obtain the height prediction result of the position of interest.
  • the obtaining the slope prediction result of the position of interest in the area to be operated based on the canopy envelope performed by the processor 720 includes: obtaining the position of interest relative to the agricultural unmanned area The horizontal distance of the aircraft 700; the derivative of the canopy envelope at the horizontal distance is calculated to obtain the slope prediction result of the position of interest.
  • the obtaining the slope prediction result of the position of interest in the area to be operated based on the canopy envelope performed by the processor 720 includes: obtaining the position of interest relative to the agricultural unmanned area the horizontal distance of the aircraft 700; obtain a predetermined number of point cloud points closest to the position of interest in the horizontal direction; fit the predetermined number of point cloud points to obtain a fitting function; calculate the fitting function in The derivative at the horizontal distance is used to obtain the slope prediction result of the position of interest.
  • the agricultural unmanned aerial vehicle 700 may further include a sensor (not shown) for collecting point cloud data in real time for the to-be-operated area in the flight direction of the agricultural unmanned aerial vehicle 700.
  • the agricultural unmanned aerial vehicle 700 may further include a power device (not shown) for controlling the ground-simulating flight of the agricultural unmanned aerial vehicle 700 based on the instructions issued by the processor 720 .
  • the agricultural UAV 700 may further include a working device (not shown) for performing plant protection operations based on instructions issued by the processor 720 .
  • a computer-readable storage medium is also provided, where program instructions are stored on the computer-readable storage medium, and the program instructions are used to execute the present application when the program instructions are run by a computer or a processor Corresponding steps of the method for agricultural unmanned aerial vehicle ground simulation of the embodiment.
  • the computer-readable storage medium may include, for example, a memory card of a smartphone, a storage component of a tablet computer, a hard disk of a personal computer, a read only memory (ROM), an erasable programmable read only memory (EPROM), a portable compact disk Read only memory (CD-ROM), USB memory, or any combination of the above computer readable storage media.
  • the computer-readable storage medium can be any combination of one or more computer-readable storage media.
  • the method for imitating the ground of agricultural unmanned aerial vehicle and the agricultural unmanned aerial vehicle detect the surrounding environment in real time, intelligently sense vegetation and changes in terrain height and drop, and ensure that the agricultural unmanned aerial vehicle is in the vegetation crown It can operate on the ground flexibly, and it will not drop at will when encountering open space and empty space, thereby improving the safety and efficiency of operation.
  • the disclosed apparatus and method may be implemented in other manners.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components may be combined or May be integrated into another device, or some features may be omitted, or not implemented.
  • Various component embodiments of the present application may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof.
  • a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all functions of some modules according to the embodiments of the present application.
  • DSP digital signal processor
  • the present application can also be implemented as a program of apparatus (eg, computer programs and computer program products) for performing part or all of the methods described herein.
  • Such a program implementing the present application may be stored on a computer-readable storage medium, or may be in the form of one or more signals. Such signals may be downloaded from Internet sites, or provided on carrier signals, or in any other form.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Databases & Information Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Health & Medical Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mining & Mineral Resources (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Animal Husbandry (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Primary Health Care (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Agronomy & Crop Science (AREA)
  • Remote Sensing (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Navigation (AREA)

Abstract

一种用于农业无人飞行器仿地作业的方法，包括：获取农业无人飞行器飞行方向上待作业区域的实时点云数据；对实时点云数据进行拟合，以得到待作业区域的植被的冠层包络；基于冠层包络对待作业区域进行地形预测，并基于地形预测的结果控制农业无人飞行器进行仿地作业。该方法对周围环境实时探测，智能感知植被及地形高低落差变化，保障农业无人飞行器在植被冠层灵敏仿地作业，且遇到空地和空当不会随意下降，从而提高作业安全性及作业效率。还包括一种农业无人飞行器和一种计算机可读存储介质。

Description

用于农业无人飞行器仿地作业的方法和农业无人飞行器
说明书
技术领域
本申请涉及农业无人飞行器技术领域，更具体地涉及一种用于农业无人飞行器仿地作业的方法和农业无人飞行器。
背景技术
现有农业无人飞行器在果树场景作业时大多需要测绘机预先对作业场景进行测绘，然后生成作业地图，导入农业无人飞行器中进行。该技术方案的缺点在于：需要另外购置测绘机，成本高；另外，农业无人飞行器作业前，需要对作业场景进行测绘，耗时多。
现有农业无人飞行器也有基于机器视觉对作业场景实时感知，实现在果树场景下的自主仿地作业。该技术方案的缺点在于：机器视觉易受环境影响，特别是农业无人飞行器作业环境的药液、尘雾等。
现有农业无人飞行器也有基于平板雷达的仿地作业方案。该方案的缺点在于，雷达实时感知飞机正下方的地面相对飞机的高度，这就将导致遇到空地和大的空当，会盲目仿地下降。
发明内容
本申请提供一种用于农业无人飞行器仿地作业的方法和农业无人飞行器，可以保障农业无人飞行器在植被冠层灵敏仿地作业，提高了作业安全性及作业效率。
第一方面，本申请实施例提供了一种用于农业无人飞行器仿地作业的方法，所述方法包括：获取农业无人飞行器飞行方向上待作业区域的实时点云数据；对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络；基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制所述农业无人飞行器进行仿地作业。
第二方面，本申请实施例提供了一种农业无人飞行器，所述农业无人 飞行器包括处理器和存储器，所述存储器用于存储指令，所述处理器调用所述存储器存储的指令用于执行以下操作：获取农业无人飞行器飞行方向上待作业区域的实时点云数据；对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络；基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制所述农业无人飞行器进行仿地作业。
第三方面，本申请实施例提供了一种计算机可读存储介质，所述计算机可读存储介质包括指令，当其在计算机上运行时，使得所述计算机执行上述用于农业无人飞行器仿地作业的方法。
根据本申请实施例的用于农业无人飞行器仿地作业的方法和农业无人飞行器对周围环境实时探测，智能感知植被及地形高低落差变化，保障农业无人飞行器在植被冠层灵敏仿地作业，且遇到空地和空当不会随意下降，从而提高作业安全性及作业效率。
附图说明
图1示出现有农业无人飞行器仿地作业的一个示例的示意图。
图2示出现有农业无人飞行器仿地作业的另一个示例的示意图。
图3示出根据本申请实施例的用于农业无人飞行器仿地作业的方法的示意性流程图。
图4示出根据本申请实施例的用于农业无人飞行器仿地作业的方法中杂点剔除的示例性示意图。
图5示出根据本申请实施例的用于农业无人飞行器仿地作业的方法中杂点剔除之后的示例性效果图。
图6示出根据本申请实施例的用于农业无人飞行器仿地作业的方法中冠层包络拟合的示例性示意图。
图7示出根据本申请实施例的农业无人飞行器的示意性框图。
具体实施方式
下面将参照附图详细描述根据本申请的示例实施例。
应当理解的是，本申请能够以不同形式实施，而不应当解释为局限于 这里提出的实施例。相反地，提供这些实施例将使公开彻底和完全，并且将本申请的范围完全地传递给本领域技术人员。
在此使用的术语的目的仅在于描述具体实施例并且不作为本申请的限制。在此使用时，单数形式的“一”、“一个”和“所述/该”也意图包括复数形式，除非上下文清楚指出另外的方式。还应明白术语“组成”和/或“包括”，当在该说明书中使用时，确定所述特征、整数、步骤、操作、元件和/或部件的存在，但不排除一个或更多其它的特征、整数、步骤、操作、元件、部件和/或组的存在或添加。在此使用时，术语“和/或”包括相关所列项目的任何及所有组合。
现有农业无人飞行器在果树场景作业时存在一些问题，例如，对于采用测绘机预先生成作业地图的方案，不仅成本高，还耗时多；对于实时感知作业场景的方案，存在地形跟随效果差、飞行安全风险高的问题。
图1示出了现有农业无人飞行器仿地作业的一个示例的示意图。如图1所示，该农业无人飞行器基于平板雷达进行仿地作业。平板雷达实时感知农业无人飞行器飞行正下方的对象相对于农业无人飞行器的高度，而飞行正下方的对象可能是植被(例如果树)冠层，也可能是非植被冠层，诸如空地或空当；当平板雷达实时感知到植被冠层相对于农业无人飞行器的高度时，尚可以在此基础上在冠层上方一定距离处安全飞行和作业；但当平板雷达实时感知到空当和空地相对于农业无人飞行器的高度时，会仍然按照原有安全高度距离标准在空当和空地上方一定距离处飞行，这导致农业无人飞行器盲目仿地下降，可能导致飞行安全问题，并影响作业效率。
图2示出了现有农业无人飞行器仿地作业的一个示例的示意图。如图2所示，该农业无人飞行器基于旋转雷达进行仿地作业。旋转雷达实时感知农业无人飞行器飞行下方的对象相对于农业无人飞行器的角度(如图2中所示的θ)和距离(如图2中所示的L)。与图1所示类似的，旋转雷达感知到空当和空地相对于农业无人飞行器的距离时，会仍然按照原有安全高度距离标准在空当和空地上方一定距离处飞行，这导致农业无人飞行器盲目仿地下降，可能导致飞行安全问题，并影响作业效率。
针对上述问题，本申请提供了一种用于农业无人飞行器仿地作业的方案。下面结合图3到图7进行详细描述。
图3示出了根据本申请实施例的用于农业无人飞行器仿地作业的方法300的示意性流程图。如图3所示，根据本申请实施例的用于农业无人飞行器仿地作业的方法300可以包括：
S310，获取农业无人飞行器飞行方向上待作业区域的实时点云数据。
S320，对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络。
S330，基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制所述农业无人飞行器进行仿地作业。
在本申请的实施例中，基于对农业无人飞行器(诸如植保机)飞行方向上待作业区域实时采集的点云数据拟合生成待作业区域的植被(诸如果树)的冠层包络，其中该冠层包络是模拟植被顶层高度变化情况的曲线或曲面，其反映了植被顶层相对于农业无人飞行器的水平距离与相对于农业无人飞行器的垂直距离之间的关系。在本申请的实施例中，基于所得冠层包络进行地形预测以控制农业无人飞行器仿地作业，这使得农业无人飞行器的飞行控制和仿地作业均在植被的冠层上方进行，不会出现在遇到植被之间的空当或空地时盲目仿地下降，从而保证农业无人飞行器在植被冠层灵敏仿地作业，提高仿地灵敏度，提高作业安全性和作业效率。
在本申请的实施例中，可以基于安装在农业无人飞行器上的传感器来针对农业无人飞行器飞行方向上的待作业区域采集实时点云数据。示例性地，安装在农业无人飞行器上的传感器可以包括旋转毫米波雷达、平板雷达、视觉传感器等等。基于传感器采集的农业无人飞行器飞行方向上的待作业区域的实时点云数据，可计算待作业区域的植被的冠层包络。
在本申请的实施例中，针对获取的农业无人飞行器飞行方向上的待作业区域的实时点云数据，可先对其进行预处理，再对预处理后的实时点云数据进行拟合而生成待作业区域的植被的冠层包络，以提高拟合结果的准确度。
示例性地，对所获取的实时点云数据所进行的预处理可以包括坐标系转换。例如，可以将所获取的实时点云数据从传感器坐标系(例如，如图2所示的雷达坐标系，即极坐标系)转换为笛卡尔坐标系，得到坐标转换后的数据，所述坐标转换后的数据中每个点云点的坐标指示所述点云点相 对于所述农业无人飞行器的水平距离和垂直距离。在该示例中，坐标转换可以通过如下公式来完成：
X i＝L×sin(θ)
Y i＝L×cos(θ)
在上式中，X i和Y i表示坐标转换后的点云数据，其中，X i表示反射目标相对于传感器载体(农业无人飞行器)的水平距离，Y i表示反射目标相对于传感器载体(农业无人飞行器)的垂直距离；L和θ表示坐标转换前的点云数据，其中，L表示反射目标相对于传感器的径向距离，θ表示反射目标相对于传感器的方位角。对所获取的实时点云数据进行坐标转换可以获得直接反映反射目标相对于农业无人飞行器的水平和垂直距离，在此基础上拟合得到的冠层包络更易于对待作业区域中感兴趣位置进行高度预测和/或坡度预测，从而易于实现对农业无人飞行器仿地飞行作业的灵敏控制。
在本申请的实施例中，还可以对所获取的实时点云数据进行杂点剔除，以剔除误差数据(例如植被区域中存在的电线杆反射得到的点云数据之类的影响植被冠层包络拟合精准度的数据等等)，之后再对杂点剔除后的实时点云数据进行拟合而生成待作业区域的植被的冠层包络，以提高拟合结果的准确度。优选地，可以对前述坐标转换后的数据进行杂点剔除，得到用于进行拟合的有效数据。
在本申请的实施例中，可以基于类似形态学滤波的方式或其他合适的方式进行杂点剔除。在一个示例中，杂点剔除可以包括将坐标转换后的数据中的每个点云点作为待检测点执行以下操作：获取在水平方向上距离所述待检测点最近的预定数目的点云点；计算所述预定数目的点云点相对于农业无人飞行器的垂直距离的均值或中值以作为参考值；计算待检测点相对于农业无人飞行器的垂直距离与所述参考值之间的差值；当所述差值大于或等于预定阈值时，将所述待检测点作为杂点剔除；当所述差值小于所述预定阈值时，将所述待检测点作为有效点保留。其中，在水平方向上距离所述待检测点最近的预定数目的点云点可以包括：在水平方向上在所述待检测点一侧的预定数目的点云点(一般可以适用于待检测点另一侧没有点云点，或者另一侧点云点数目没有达到预定数目的场景)；或者在水平方向上在所述待检测点两侧的点云点，每侧的点云点的数目为所述预定数目 的一半。下面结合图4来描述。
图4示出根据本申请实施例的用于农业无人飞行器仿地作业的方法中杂点剔除的示例性示意图。如图4所示，坐标系中的每个点即为一个点云点，对于当前待检测点P1，获取在水平方向上距离P1最近的四个(示例)点(如图4所示的邻近点)，则可以计算这四个点相对于农业无人飞行器的垂直距离(即这四个点的纵坐标)的均值或中值以作为参考值，接着计算当前待检测点P1相对于农业无人飞行器的垂直距离(即P1的纵坐标)与参考值的差值，如果差值较大(大于或等于某阈值)，则认为当前待检测点P1为杂点予以剔除，反之如果差值较大(小于某阈值)，则认为当前待检测点P1为有效点予以保留。
在图4所示的示例中，很明显，可以看出当前检测点P1相对于其四个邻近点相比，纵坐标之间的差值较大，即该点应该是作为杂点被剔除，而诸如P2、P3这样的待检测点与其水平方向上邻近点相比，明显它们的纵坐标之间的差值较小，因为它们与周围点的同是属于植被冠层上的点，这样的点会被保留下来，用于精确地计算植被的冠层包络。杂点剔除之后，可以得到植被冠层的反射点，如图5示出的杂点剔除之后的示例性效果图所示的。
经过了前述处理，所得到的点云数据用于进行植被冠层包络的拟合。在本申请的实施例中，可以基于以下拟合方式中的任一种来进行植被冠层包络的拟合：基于线性插值的拟合、多项式拟合、基于高斯回归的拟合、基于b样条曲线(BSPLINE)的拟合、基于支持向量机(SVM)的拟合。下面以线性插值为例来描述，线性插值过程可以如下式所示的：
Figure PCTCN2020124082-appb-000001
Figure PCTCN2020124082-appb-000002
...
Figure PCTCN2020124082-appb-000003
在上面的式子中，(x 0,y 0)、(x 1,y 1)、(x 2,y 2)、……、(x n-1,y n-1)、(x n,y n)是指实际采集的点云数据(诸如经过前述坐标转换和杂点剔除后的数据)，是 通过插值得到的点，每个插值的点可以根据两个采集点计算得到。例如，基于(x 0,y 0)和(x 1,y 1)计算(x 0,y 0)和(x 1,y 1)之间的插值点的坐标，基于(x 1,y 1)和(x 2,y 2)计算(x 1,y 1)和(x 2,y 2)之间的插值点的坐标，诸如此类等等。插值后得到更多的点，然后输入一个预设模型，比如多项式，再把所有点代入预设模型得到方程组，即可以计算多项式的系数，最终得到的即为表示植被的冠层包络的方程。相对于之前实时采集的离散点云数据，得到的冠层包络是连续数据，如图6示出的冠层包络拟合的示例性示意图所示的。
应理解，采用线性插值进行拟合的方式仅是示例性的，还可以采用其他合适的方式进行拟合。相对来说，线性插值的拟合方式计算简单、计算量小，拟合精度受原始点云点的分布情况影响；还可以通过其他诸如多项式拟合、BSPLINE、高斯回归、SVM等方法进行拟合，以提高拟合精度。
在本申请的实施例中，在得到植被的冠层包络后，可以基于所述冠层包络对待作业区域进行地形预测，并基于地形预测的结果控制农业无人飞行器进行仿地作业。在本申请的实施例中，基于所述冠层包络对待作业区域进行地形预测，可以包括：基于所述冠层包络获取待作业区域中感兴趣位置的高度预测结果和/或坡度预测结果。基于所得到的冠层包络，可以预测农业无人飞行器飞行方向上待作业区域的高度情况和/或高度变化情况(坡度情况)，以基于其调整农业无人飞行器在垂直方向上的速度，在确保飞行安全的情况下实现更为灵敏的仿地作业。
在本申请的实施例中，基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果，可以包括：获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；将所述水平距离代入所述冠层包络得到所述感兴趣位置的高度预测结果。如前所述，所得到的植被的冠层包络是连续数据，因此将任一位置距离农业无人飞行器的水平距离代入冠层包络，即可得到该位置相对于农业无人飞行器的垂直距离(高度情况)。
在本申请的实施例中，基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，可以包括：获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；计算所述冠层包络在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。由于已经得到农业无人飞行器飞行方向上待作业区域的植被的冠层包络，因此要想获得该待作业区域某位置处 的坡度，对冠层包络求导即可。例如，该位置处相对于农业无人飞行器的水平距离表示为x i,则计算冠层包络在x i处的导数即可得到该位置处的坡度预测结果。
在本申请的另一个实施例中，基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，可以包括：获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；获取在水平方向上距离所述感兴趣位置最近的预定数目的点云点；对所述预定数目的点云点进行拟合得到拟合函数；计算所述拟合函数在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。例如，某位置处相对于农业无人飞行器的水平距离表示为x i，可以获取该位置处附近区域的点相对于农业无人飞行器的水平距离，于是得到多个采样点相对于农业无人飞行器的水平距离的集合{x i-n,...,x i-1,x i,x i+1,...,x i+n}，将该集合代入冠层包络，得到各采样点相对于农业无人飞行器的垂直距离的集合{y i-n,...,y i-1,y i,y i+1,...,y i+n}，基于这两个集合进行拟合，得到新的拟合方程，求该新的拟合方程在x i处的导数，即得到该位置处的坡度预测结果。在该实施例中，基于已得冠层包络获取感兴趣位置附近区域多个位置相对于农业无人飞行器的垂直距离，再根据该区域内这些位置处相对于农业无人飞行器的水平距离和垂直距离重新拟合此区域更为精确的冠层包络，然后再求导，可以得到感兴趣位置处更为准确的坡度预测结果。
以上是感兴趣位置处于农业无人飞行器的传感器的观测范围之内的情况，在该情况下主要基于感兴趣位置的邻近点的高度内推得到感兴趣位置的高度/坡度预测结果。当感兴趣位置处于农业无人飞行器的传感器的观测范围之外时，可以使用一定的算法进行外推而得到感兴趣位置的高度/坡度预测结果。总之，根据所获得的感兴趣位置的高度预测结果和/或坡度预测结果，可以调整农业无人飞行器在垂直方向上的速度，在确保飞行安全的情况下实现更为灵敏的仿地作业。
基于上面的描述，根据本申请实施例的用于农业无人飞行器仿地作业的方法和农业无人飞行器对周围环境实时探测，智能感知植被及地形高低落差变化，保障农业无人飞行器在植被冠层灵敏仿地作业，且遇到空地和空当不会随意下降，从而提高作业安全性及作业效率。此外，由于无需事 先使用测绘机对作业环境进行测绘，还降低了成本。
以上示例性地描述了根据本申请实施例的用于农业无人飞行器仿地作业的方法。下面结合图7描述根据本申请另一方面提供的农业无人飞行器。
图7示出根据本申请实施例的农业无人飞行器700的示意性框图。如图7所示，农业无人飞行器700包括存储器710以及处理器720。其中，存储器710存储用于实现根据本申请实施例的用于农业无人飞行器仿地作业的方法中的相应步骤的程序指令。处理器720用于运行存储器710中存储的程序，以执行根据本申请实施例的用于农业无人飞行器仿地作业的方法的相应步骤。本领域技术人员可以结合前文描述理解处理器720执行的操作，为了简洁，此处仅描述一些主要操作，具体细节不再赘述。
在本申请的实施例中，处理器720调用存储器710存储的指令用于执行以下操作：获取农业无人飞行器700飞行方向上待作业区域的实时点云数据；对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络；基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制农业无人飞行器700进行仿地作业。
在本申请的实施例中，处理器720还执行以下操作：在获取所述实时点云数据之后，对所述实时点云数据进行预处理得到预处理后的数据，并且所述拟合是基于所述预处理后的数据进行的。
在本申请的实施例中，处理器720执行的所述预处理包括将所述实时点云数据从传感器坐标系转换为笛卡尔坐标系，得到坐标转换后的数据，所述坐标转换后的数据中每个点云点的坐标指示所述点云点相对于农业无人飞行器700的水平距离和垂直距离。
在本申请的实施例中，处理器720执行的所述预处理还包括对所述坐标转换后的数据进行杂点剔除，得到用于进行所述拟合的有效数据。
在本申请的实施例中，处理器720执行的所述杂点剔除是基于形态学滤波的方式进行的。
在本申请的实施例中，处理器720执行的所述杂点剔除包括将所述坐标转换后的数据中的每个点云点作为待检测点执行以下操作：获取在水平方向上距离所述待检测点最近的预定数目的点云点；计算所述预定数目的 点云点相对于农业无人飞行器700的垂直距离的均值或中值以作为参考值；计算所述待检测点相对于农业无人飞行器700的垂直距离与所述参考值之间的差值；当所述差值大于或等于预定阈值时，将所述待检测点作为杂点剔除；当所述差值小于所述预定阈值时，将所述待检测点作为有效点保留。
在本申请的实施例中，所述在水平方向上距离所述待检测点最近的预定数目的点云点包括：在水平方向上在所述待检测点一侧的预定数目的点云点；或者在水平方向上在所述待检测点两侧的点云点，每侧的点云点的数目为所述预定数目的一半。
在本申请的实施例中，所述拟合包括以下方式中的任一种：基于线性插值的拟合、多项式拟合、基于高斯回归的拟合、基于b样条曲线的拟合、基于支持向量机的拟合。
在本申请的实施例中，处理器720执行的所述基于所述冠层包络对所述待作业区域进行地形预测，包括：基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果和/或坡度预测结果。
在本申请的实施例中，处理器720执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果，包括：获取所述感兴趣位置相对于农业无人飞行器700的水平距离；将所述水平距离代入所述冠层包络得到所述感兴趣位置的高度预测结果。
在本申请的实施例中，处理器720执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：获取所述感兴趣位置相对于农业无人飞行器700的水平距离；计算所述冠层包络在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
在本申请的实施例中，处理器720执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：获取所述感兴趣位置相对于农业无人飞行器700的水平距离；获取在水平方向上距离所述感兴趣位置最近的预定数目的点云点；对所述预定数目的点云点进行拟合得到拟合函数；计算所述拟合函数在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
在本申请的实施例中，农业无人飞行器700还可以包括传感器(未示 出)，用于针对农业无人飞行器700飞行方向上待作业区域实时采集点云数据。此外，农业无人飞行器700还可以包括动力装置(未示出)，用于基于处理器720发出的指令控制农业无人飞行器700的仿地飞行。此外，农业无人飞行器700还可以包括作业装置(未示出)，用于基于处理器720发出的指令执行植保作业。
此外，根据本申请实施例，还提供了一种计算机可读存储介质，在所述计算机可读存储介质上存储了程序指令，在所述程序指令被计算机或处理器运行时用于执行本申请实施例的用于农业无人飞行器仿地作业的方法的相应步骤。所述计算机可读存储介质例如可以包括智能电话的存储卡、平板电脑的存储部件、个人计算机的硬盘、只读存储器(ROM)、可擦除可编程只读存储器(EPROM)、便携式紧致盘只读存储器(CD-ROM)、USB存储器、或者上述计算机可读存储介质的任意组合。所述计算机可读存储介质可以是一个或多个计算机可读存储介质的任意组合。
基于上面的描述，根据本申请实施例的用于农业无人飞行器仿地作业的方法和农业无人飞行器对周围环境实时探测，智能感知植被及地形高低落差变化，保障农业无人飞行器在植被冠层灵敏仿地作业，且遇到空地和空当不会随意下降，从而提高作业安全性及作业效率。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的设备和方法，可以通过其它的方式实现。例如，以上所描述的设备实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个设备，或一些特征可以忽略，或不执行。
在此处所提供的说明书中，说明了大量具体细节。然而，能够理解，本申请的实施例可以在没有这些具体细节的情况下实践。在一些实例中，并未详细示出公知的方法、结构和技术，以便不模糊对本说明书的理解。
类似地，应当理解，为了精简本申请并帮助理解各个发明方面中的一个或多个，在对本申请的示例性实施例的描述中，本申请的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而，并不应将该本申请的方法解释成反映如下意图：即所要求保护的本申请要求比在权利要求中所明确记载的特征更多的特征。更确切地说，如相应的权利要求书所反映的那样，其发明点在于可以用少于某个公开的单个实施例的所有特征的特征来解决相应的技术问题。因此，遵循具体实施方式的权利要求书由此明确地并入该具体实施方式，其中权利要求本身都作为本申请的单独实施例。
本领域的技术人员可以理解，除了特征之间相互排斥之外，可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述，本说明书(包括伴随的权利要求、摘要和附图)中公开的特征可以由提供相同、等同或相似目的的替代特征来代替。
此外，本领域的技术人员能够理解，尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征，但是不同实施例的特征的组合意味着处于本申请的范围之内并且形成不同的实施例。例如，在权利要求书中，所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本申请的各个部件实施例可以以硬件实现，或者以在一个或者多个处理器上运行的软件模块实现，或者以它们的组合实现。本领域的技术人员应当理解，可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本申请实施例的一些模块的一些或者全部功能。本申请还可以实现为用于执行这里所描述的方法的一部分或者全部的装置程序(例如，计算机程序和计算机程序产品)。这样的实现本申请的程序可以存储在计算机可读存储介质上，或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到，或者在载体信号上提供，或者以任何其他形式提供。
应该注意的是上述实施例对本申请进行说明而不是对本申请进行限制，并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出 替换实施例。在权利要求中，不应将位于括号之间的任何参考符号构造成对权利要求的限制。本申请可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中，这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。
以上所述，仅为本申请的具体实施方式或对具体实施方式的说明，本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。本申请的保护范围应以权利要求的保护范围为准。

Claims (25)

  1. 一种用于农业无人飞行器仿地作业的方法，其特征在于，所述方法包括：
    获取农业无人飞行器飞行方向上待作业区域的实时点云数据；
    对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络；
    基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制所述农业无人飞行器进行仿地作业。
  2. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    在获取所述实时点云数据之后，对所述实时点云数据进行预处理得到预处理后的数据，并且所述拟合是基于所述预处理后的数据进行的。
  3. 根据权利要求2所述的方法，其特征在于，所述预处理包括将所述实时点云数据从传感器坐标系转换为笛卡尔坐标系，得到坐标转换后的数据，所述坐标转换后的数据中每个点云点的坐标指示所述点云点相对于所述农业无人飞行器的水平距离和垂直距离。
  4. 根据权利要求3所述的方法，其特征在于，所述预处理还包括对所述坐标转换后的数据进行杂点剔除，得到用于进行所述拟合的有效数据。
  5. 根据权利要求4所述的方法，其特征在于，所述杂点剔除是基于形态学滤波的方式进行的。
  6. 根据权利要求4所述的方法，其特征在于，所述杂点剔除包括将所述坐标转换后的数据中的每个点云点作为待检测点执行以下操作：
    获取在水平方向上距离所述待检测点最近的预定数目的点云点；
    计算所述预定数目的点云点相对于所述农业无人飞行器的垂直距离的均值或中值以作为参考值；
    计算所述待检测点相对于所述农业无人飞行器的垂直距离与所述参考值之间的差值；
    当所述差值大于或等于预定阈值时，将所述待检测点作为杂点剔除；
    当所述差值小于所述预定阈值时，将所述待检测点作为有效点保留。
  7. 根据权利要求6所述的方法，其特征在于，所述在水平方向上距离所述待检测点最近的预定数目的点云点包括：
    在水平方向上在所述待检测点一侧的预定数目的点云点；或者
    在水平方向上在所述待检测点两侧的点云点，每侧的点云点的数目为所述预定数目的一半。
  8. 根据权利要求1-7中的任一项所述的方法，其特征在于，所述拟合包括以下方式中的任一种：基于线性插值的拟合、多项式拟合、基于高斯回归的拟合、基于b样条曲线的拟合、基于支持向量机的拟合。
  9. 根据权利要求1-7中的任一项所述的方法，其特征在于，所述基于所述冠层包络对所述待作业区域进行地形预测，包括：
    基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果和/或坡度预测结果。
  10. 根据权利要求9所述的方法，其特征在于，所述基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    将所述水平距离代入所述冠层包络得到所述感兴趣位置的高度预测结果。
  11. 根据权利要求9所述的方法，其特征在于，所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    计算所述冠层包络在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
  12. 根据权利要求9所述的方法，其特征在于，所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    获取在水平方向上距离所述感兴趣位置最近的预定数目的点云点；
    对所述预定数目的点云点进行拟合得到拟合函数；
    计算所述拟合函数在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
  13. 一种农业无人飞行器，其特征在于，所述农业无人飞行器包括处理器和存储器，所述存储器用于存储指令，所述处理器调用所述存储器存储的指令用于执行以下操作：
    获取农业无人飞行器飞行方向上待作业区域的实时点云数据；
    对所述实时点云数据进行拟合，以得到所述待作业区域的植被的冠层包络；
    基于所述冠层包络对所述待作业区域进行地形预测，并基于所述地形预测的结果控制所述农业无人飞行器进行仿地作业。
  14. 根据权利要求13所述的农业无人飞行器，其特征在于，所述处理器还执行以下操作：
    在获取所述实时点云数据之后，对所述实时点云数据进行预处理得到预处理后的数据，并且所述拟合是基于所述预处理后的数据进行的。
  15. 根据权利要求14所述的农业无人飞行器，其特征在于，所述处理器执行的所述预处理包括将所述实时点云数据从传感器坐标系转换为笛卡尔坐标系，得到坐标转换后的数据，所述坐标转换后的数据中每个点云点的坐标指示所述点云点相对于所述农业无人飞行器的水平距离和垂直距离。
  16. 根据权利要求15所述的农业无人飞行器，其特征在于，所述处理器执行的所述预处理还包括对所述坐标转换后的数据进行杂点剔除，得到用于进行所述拟合的有效数据。
  17. 根据权利要求16所述的农业无人飞行器，其特征在于，所述处理器执行的所述杂点剔除是基于形态学滤波的方式进行的。
  18. 根据权利要求16所述的农业无人飞行器，其特征在于，所述处理器执行的所述杂点剔除包括将所述坐标转换后的数据中的每个点云点作为待检测点执行以下操作：
    获取在水平方向上距离所述待检测点最近的预定数目的点云点；
    计算所述预定数目的点云点相对于所述农业无人飞行器的垂直距离的均值或中值以作为参考值；
    计算所述待检测点相对于所述农业无人飞行器的垂直距离与所述参考值之间的差值；
    当所述差值大于或等于预定阈值时，将所述待检测点作为杂点剔除；
    当所述差值小于所述预定阈值时，将所述待检测点作为有效点保留。
  19. 根据权利要求18所述的农业无人飞行器，其特征在于，所述在 水平方向上距离所述待检测点最近的预定数目的点云点包括：
    在水平方向上在所述待检测点一侧的预定数目的点云点；或者
    在水平方向上在所述待检测点两侧的点云点，每侧的点云点的数目为所述预定数目的一半。
  20. 根据权利要求13-19中的任一项所述的农业无人飞行器，其特征在于，所述拟合包括以下方式中的任一种：基于线性插值的拟合、多项式拟合、基于高斯回归的拟合、基于b样条曲线的拟合、基于支持向量机的拟合。
  21. 根据权利要求13-19中的任一项所述的农业无人飞行器，其特征在于，所述处理器执行的所述基于所述冠层包络对所述待作业区域进行地形预测，包括：
    基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果和/或坡度预测结果。
  22. 根据权利要求21所述的农业无人飞行器，其特征在于，所述处理器执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的高度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    将所述水平距离代入所述冠层包络得到所述感兴趣位置的高度预测结果。
  23. 根据权利要求21所述的农业无人飞行器，其特征在于，所述处理器执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    计算所述冠层包络在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
  24. 根据权利要求21所述的农业无人飞行器，其特征在于，所述处理器执行的所述基于所述冠层包络获取所述待作业区域中感兴趣位置的坡度预测结果，包括：
    获取所述感兴趣位置相对于所述农业无人飞行器的水平距离；
    获取在水平方向上距离所述感兴趣位置最近的预定数目的点云点；
    对所述预定数目的点云点进行拟合得到拟合函数；
    计算所述拟合函数在所述水平距离处的导数，得到所述感兴趣位置的坡度预测结果。
  25. 一种计算机可读存储介质，其特征在于，包括指令，当其在计算机上运行时，使得所述计算机执行如权利要求1-12中的任一项所述的用于农业无人飞行器仿地作业的方法。
PCT/CN2020/124082 2020-10-27 2020-10-27 用于农业无人飞行器仿地作业的方法和农业无人飞行器 WO2022087841A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2020/124082 WO2022087841A1 (zh) 2020-10-27 2020-10-27 用于农业无人飞行器仿地作业的方法和农业无人飞行器
CN202080030164.2A CN113966514A (zh) 2020-10-27 2020-10-27 用于农业无人飞行器仿地作业的方法和农业无人飞行器

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/124082 WO2022087841A1 (zh) 2020-10-27 2020-10-27 用于农业无人飞行器仿地作业的方法和农业无人飞行器

Publications (1)

Publication Number Publication Date
WO2022087841A1 true WO2022087841A1 (zh) 2022-05-05

Family

ID=79460360

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/124082 WO2022087841A1 (zh) 2020-10-27 2020-10-27 用于农业无人飞行器仿地作业的方法和农业无人飞行器

Country Status (2)

Country Link
CN (1) CN113966514A (zh)
WO (1) WO2022087841A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114868527A (zh) * 2022-05-16 2022-08-09 云南省林业和草原科学院 基于无人机的深纹核桃果实采收方法
CN115629619A (zh) * 2022-11-04 2023-01-20 广东电网有限责任公司 无人机输电线路巡检系统及方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103885454A (zh) * 2014-03-07 2014-06-25 华南农业大学 农用飞行器跟随冠层特征参数飞行的作业方法及装置
CN106199627A (zh) * 2016-09-14 2016-12-07 中国农业科学院农业资源与农业区划研究所 草地植被参数获取方法
CN107148633A (zh) * 2014-08-22 2017-09-08 克莱米特公司 用于使用无人机系统进行农艺和农业监测的方法
US20180373259A1 (en) * 2017-06-22 2018-12-27 360 Yield Center, Llc Enhanced automated steering system for a vehicle
CN110045748A (zh) * 2019-04-03 2019-07-23 深圳高速工程检测有限公司 飞行器控制方法、装置、计算机设备和存储介质
CN110832425A (zh) * 2018-10-31 2020-02-21 深圳市大疆创新科技有限公司 控制方法、设备、测绘无人机和喷洒无人机
CN111506097A (zh) * 2020-04-30 2020-08-07 苏州恒井泰信息技术有限公司 一种无人机遥感技术在精准农业中的应用系统及方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109032157A (zh) * 2018-07-23 2018-12-18 拓攻（南京）机器人有限公司 无人机仿地作业方法、装置、设备及存储介质
CN109144097B (zh) * 2018-08-15 2021-04-06 广州极飞科技有限公司 障碍物或地面识别及飞行控制方法、装置、设备及介质

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103885454A (zh) * 2014-03-07 2014-06-25 华南农业大学 农用飞行器跟随冠层特征参数飞行的作业方法及装置
CN107148633A (zh) * 2014-08-22 2017-09-08 克莱米特公司 用于使用无人机系统进行农艺和农业监测的方法
CN106199627A (zh) * 2016-09-14 2016-12-07 中国农业科学院农业资源与农业区划研究所 草地植被参数获取方法
US20180373259A1 (en) * 2017-06-22 2018-12-27 360 Yield Center, Llc Enhanced automated steering system for a vehicle
CN110832425A (zh) * 2018-10-31 2020-02-21 深圳市大疆创新科技有限公司 控制方法、设备、测绘无人机和喷洒无人机
CN110045748A (zh) * 2019-04-03 2019-07-23 深圳高速工程检测有限公司 飞行器控制方法、装置、计算机设备和存储介质
CN111506097A (zh) * 2020-04-30 2020-08-07 苏州恒井泰信息技术有限公司 一种无人机遥感技术在精准农业中的应用系统及方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114868527A (zh) * 2022-05-16 2022-08-09 云南省林业和草原科学院 基于无人机的深纹核桃果实采收方法
CN115629619A (zh) * 2022-11-04 2023-01-20 广东电网有限责任公司 无人机输电线路巡检系统及方法

Also Published As

Publication number Publication date
CN113966514A (zh) 2022-01-21

Similar Documents

Publication Publication Date Title
RU2768997C1 (ru) Способ, устройство и оборудование для распознавания препятствий или земли и управления полетом, и носитель данных
JP6700482B2 (ja) プロペラの羽根に統合された撮像装置を用いたステレオ距離情報の判定
Yu et al. Automatic detection of harvested trees and determination of forest growth using airborne laser scanning
CN105159297B (zh) 输电线路无人机巡检避障系统与方法
Rosell et al. Obtaining the three-dimensional structure of tree orchards from remote 2D terrestrial LIDAR scanning
Palleja et al. Sensitivity of tree volume measurement to trajectory errors from a terrestrial LIDAR scanner
CN110673107B (zh) 基于多线激光雷达的路沿检测方法及装置
CA2781604C (en) Method and apparatus for analyzing tree canopies with lidar data
US20110254924A1 (en) Optronic system and method dedicated to identification for formulating three-dimensional images
WO2022087841A1 (zh) 用于农业无人飞行器仿地作业的方法和农业无人飞行器
US10139493B1 (en) Rotor safety system
US8682037B2 (en) Method and system for thinning a point cloud
CN111615677A (zh) 一种无人机的安全降落方法、装置、无人机及介质
Vauhkonen et al. Deriving airborne laser scanning based computational canopy volume for forest biomass and allometry studies
Grznárová et al. The crown diameter estimation from fixed wing type of UAV imagery
WO2022141221A1 (zh) 喷洒作业控制方法、装置、农业无人飞行器及存储介质
CN109407112A (zh) 基于LiDAR点云数据的果树树冠体积测算方法及系统
JP2015200615A (ja) レーザー計測結果解析システム
Pádua et al. UAS-based imagery and photogrammetric processing for tree height and crown diameter extraction
JP7672834B2 (ja) 森林状態推定システム、森林状態推定方法およびコンピュータプログラム
Xu et al. On-the-fly extraction of polyhedral buildings from airborne LiDAR data
CN106240454A (zh) 提供车辆碰撞预警的系统及车载设备
Gangadharan et al. A comparison of approaches for citrus canopy profile generation using ultrasonic and Leddar® sensors
JP7374977B2 (ja) 樹木情報推定システム、樹木情報推定方法およびコンピュータプログラム
Liu et al. Comparison of Sick and Hokuyo UTM-30LX laser sensors in canopy detection for variable-rate sprayer

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: 20959005

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: 20959005

Country of ref document: EP

Kind code of ref document: A1