WO2024011750A1 - 一种基于后台图像定位的无人机生姜种植巡检系统 - Google Patents

一种基于后台图像定位的无人机生姜种植巡检系统 Download PDF

Info

Publication number
WO2024011750A1
WO2024011750A1 PCT/CN2022/120039 CN2022120039W WO2024011750A1 WO 2024011750 A1 WO2024011750 A1 WO 2024011750A1 CN 2022120039 W CN2022120039 W CN 2022120039W WO 2024011750 A1 WO2024011750 A1 WO 2024011750A1
Authority
WO
WIPO (PCT)
Prior art keywords
module
ginger
angle
infrared laser
abnormal
Prior art date
Application number
PCT/CN2022/120039
Other languages
English (en)
French (fr)
Inventor
李洪雷
李佳林
Original Assignee
重庆文理学院
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 重庆文理学院 filed Critical 重庆文理学院
Publication of WO2024011750A1 publication Critical patent/WO2024011750A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
    • G01C11/02—Picture taking arrangements specially adapted for photogrammetry or photographic surveying, e.g. controlling overlapping of pictures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H04N7/183—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a single remote source
    • H04N7/185—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a single remote source from a mobile camera, e.g. for remote control

Definitions

  • the invention relates to the technical field of ginger planting, and in particular to a UAV ginger planting inspection system based on background image positioning.
  • ginger planting In the process of ginger planting, it takes 180 days from emergence to harvest, with a total of four stages.
  • the first two stages of ginger planting are the seedling stage. Its root system is weak and young, and it is easy to get sick. As a result, the ginger seedlings cannot grow normally, and even It affects surrounding seedlings, causing problems such as reduced yield in ginger planting fields, low ginger yield, and poor edible effect.
  • the image features have certain errors (due to the shooting equipment itself or the shooting environment, etc.), resulting in inaccurate positioning of abnormal ginger seedlings and low positioning accuracy. The location of abnormal seedlings cannot be effectively obtained, which affects the efficiency of subsequent processing.
  • the purpose of the present invention is to provide a drone ginger planting inspection system based on background image positioning.
  • the inspection system can efficiently monitor the growth of ginger seedlings, and locate abnormal seedlings accurately and with small errors, ensuring that Treat abnormal seedlings in a timely and effective manner to avoid abnormal seedlings from disturbing surrounding ginger seedlings, thereby ensuring high yield and quality of ginger.
  • a UAV ginger planting inspection system based on background image positioning characterized by: including a UAV module, a UAV camera module, anomaly detection module and anomaly positioning module;
  • a GPS positioning module is provided on the drone module to obtain the GPS positioning value of the shooting point;
  • the UAV camera module is installed on the UAV module and includes a camera, a vertical launch module and an angle launch module; the camera is fixedly installed directly below the UAV module and the lens is vertically downward; the vertical launch module and the angle launch module
  • the emission modules are respectively arranged on the left and right sides of the camera, and the vertical emission module is set parallel to the camera, that is, the emission line of the vertical emission module is vertically downward, and the angle emission module is set at an angle a with the camera, that is, the emission line of the angle emission module It is at an angle a with the camera's line of sight; the vertical emission module and the angle emission module both include a laser range finder and a visible infrared laser transmitter;
  • the abnormality detection module uses an offline-trained ginger abnormal growth image detection model to detect the captured images transmitted by the drone camera module, and outputs a ginger abnormality detection frame;
  • the abnormal positioning module includes a visible infrared laser point detection module and a correction module.
  • the UAV module also includes a UAV flight control module and a wireless transmission module.
  • the UAV flight control module is used to control the flight, landing, steering, etc. of the UAV module;
  • the wireless transmission module is used Remotely connect the drone camera module and the anomaly detection module, thereby transmitting the captured image, angle a and shooting point GPS positioning value of the drone camera module to the anomaly detection module.
  • the anomaly detection module composes a data set by screening annotated abnormal growth pictures of ginger, and uses a recognition framework for training, thereby obtaining an image detection model for abnormal ginger growth.
  • the recognition framework can adopt any one of the YOLOv3 model, YOLOv4 model, and YOLOv5 model.
  • the ginger abnormality detection frame is a rectangular frame
  • the output data of the ginger abnormal growth image detection model includes the coordinates of the upper left corner of the rectangular frame and the width and height of the rectangular frame.
  • the visible infrared laser point detection module is used to detect the infrared laser point emitted by the visible infrared laser transmitter in the photographed image transmitted by the drone camera module; the correction module is based on the GPS positioning value, angle a, and ginger of the photographed point. Abnormal detection frame and infrared laser point correct the GPS positioning value of the abnormal position of ginger in the image.
  • the specific steps of correcting the GPS positioning value of the abnormal location of ginger in the image include:
  • x represents the distance between the laser rangefinder of the vertical emission module and the ground
  • y represents the distance between the laser rangefinder of the angle emission module and the ground
  • d p is obtained by measuring the pixel value distance between the visible infrared laser emitter of the vertical emission module and the visible infrared laser emitter of the angle emission module in the captured image;
  • the counterclockwise angle ⁇ between the first connection line and the second connection line is measured by taking the image; where the first connection line is the connection between the center of the ginger anomaly detection frame and the infrared laser point of the vertical emission module, The second connection line is the connection line between the infrared laser point of the vertical emission module and the infrared laser point of the angle emission module;
  • ⁇ l is the counterclockwise angle between the transmitter plane and the due north direction when the drone takes pictures; among them, the visible infrared laser transmitter of the transmitter plane vertical transmit module and the visible infrared laser of the angle transmit module The plane formed by the emitter;
  • the inspection system also includes a storage module, a database and a comparison module.
  • the storage module is connected to the output end of the anomaly detection module and interacts with the database to classify captured images containing ginger anomaly detection frames.
  • Storage; the comparison module is connected to the storage module and the database respectively, so that when an abnormally growing ginger image occurs, it is compared with the images of the same area in different time periods stored in the database (comparison in the time dimension) to obtain More detailed analysis results.
  • This application uses a drone module to implement inspections in ginger planting fields. It has a high degree of automation and a wide range of inspections. With the drone camera module and anomaly detection module, it can efficiently, quickly and timely identify abnormally growing ginger. Seedlings, effectively saving manpower and material resources, and high recognition efficiency; through the cooperation of the visible infrared laser point detection module and the correction module, the location of abnormally growing ginger seedlings can be obtained through the background image.
  • the positioning is accurate, the error is small, and can be accurately obtained through the background image
  • the location of abnormally growing ginger seedlings in the front-end planting field reduces the hardware requirements and computing power requirements of the system front-end (i.e., the UAV module), thereby reducing the operational difficulty and operating costs of the UAV module, and improving the efficiency of inspection and identification (
  • the UAV module needs to lower its height or land on the ground to accurately obtain the actual location of abnormal ginger seedlings, which not only increases the operating cost of the UAV module, but also tests the operator's UAV operation level, and also lengthens the identification time. inspection time and reduce inspection efficiency).
  • the front-end of the UAV in this application is only used for photographing images and inspection flights.
  • the parts that require complex calculation judgment and logical processing are handed over to the back-end, which can effectively reduce the computing workload of the front-end UAV and avoid unmanned operations.
  • the machine module is overloaded (due to the effective setting space of the UAV module), resulting in short circuits, slow operation and other problems; in addition, through back-end training, identification and judgment, the accuracy is higher and the judgment is more timely, thus providing better conditions for the processing and maintenance of abnormal seedlings. Provide more time.
  • Figure 1 is a schematic structural diagram for obtaining the actual distance between two laser range finders on the ground in an embodiment of the present invention.
  • a UAV ginger planting inspection system based on background image positioning characterized by: including a UAV module, a UAV camera module, anomaly detection module and anomaly positioning module;
  • the UAV module is equipped with a GPS positioning module to obtain the GPS positioning value of the shooting point; the UAV module also includes a UAV flight control module and a wireless transmission module.
  • the UAV flight control module is used to control the UAV module. Flying, landing, turning, etc.; the wireless transmission module is used to remotely connect the drone camera module and the anomaly detection module, thereby transmitting the captured image, angle a and shooting point GPS positioning value of the drone camera module to the anomaly detection module.
  • the UAV module can also be equipped with a battery pack, battery charging and discharging circuit and solar panel, so as to use solar energy to charge and improve the endurance of the UAV module (the UAV module can adopt existing common agricultural production equipment). drone).
  • the UAV camera module is installed on the UAV module and includes a camera 100, a vertical launch module and an angle launch module; the camera 100 is fixedly installed directly below the UAV module (specifically, installed at the bottom of the UAV) and the lens is vertical downward (that is, the camera angle is vertically downward); the vertical emission module and the angle emission module are respectively arranged on the left and right sides of the camera 100, and the vertical emission module is arranged parallel to the camera 100, that is, the emission line of the vertical emission module is vertically downward.
  • the angle emission module and the camera 100 are set at an angle a, that is, the emission line of the angle emission module and the line of sight of the camera 100 are at an angle a (as shown in Figure 1); both the vertical emission module and the angle emission module include a laser rangefinder and visible infrared laser transmitter;
  • the anomaly detection module composes a data set (the data set includes a training set, a test set and a verification set) by screening annotated abnormal growth pictures of ginger (that is, filtering and annotating abnormal ginger growth pictures in the past), and uses the recognition framework for training (the training method is: The conventional image training method in this field is to use the training set to obtain the initial model, and then use the verification set to verify the initial model to obtain the abnormal ginger growth image detection model).
  • the recognition framework can use any of the YOLOv3 model, YOLOv4 model, and YOLOv5 model.
  • the output data of the abnormal ginger growth image detection model includes the coordinates of the upper left corner of the rectangular frame and the width and height of the rectangular frame.
  • the abnormal positioning module includes a visible infrared laser point detection module and a correction module;
  • the visible infrared laser point detection module is used to detect the infrared laser points emitted by the visible infrared laser transmitter in the captured images transmitted by the drone camera module; the correction module is based on the GPS positioning value of the shooting point, angle a, ginger anomaly detection frame, infrared laser point to correct the GPS positioning value of the abnormal location of ginger in the image.
  • the specific steps to correct the GPS positioning value of the abnormal location of ginger in the image include:
  • x represents the distance between the laser rangefinder of the vertical emission module and the ground
  • y represents the distance between the laser rangefinder of the angle emission module and the ground
  • d p is obtained by measuring the pixel value distance between the visible infrared laser emitter of the vertical emission module and the visible infrared laser emitter of the angle emission module in the captured image;
  • the counterclockwise angle ⁇ between the first connection line and the second connection line is measured by taking the image; where the first connection line is the connection between the center of the ginger anomaly detection frame and the infrared laser point of the vertical emission module, The second connection line is the connection line between the infrared laser point of the vertical emission module and the infrared laser point of the angle emission module;
  • ⁇ l is the counterclockwise angle between the transmitter plane and the due north direction when the drone takes pictures; among them, the visible infrared laser transmitter of the transmitter plane vertical transmit module and the visible infrared laser of the angle transmit module The plane formed by the emitter;
  • a classification module can also be set up in the anomaly detection module to output estimates of abnormal causes, such as water shortage, nutrient deficiency, insect damage, etc.; that is, the anomaly detection module obtains the ginger abnormal growth image detection model and inputs it into the pre-trained classification module Classify (the classification model can use any one of support vector machine or deep neural network, such as VGG, RexNet, etc., which can be understood by those skilled in the art and will not be discussed in detail in this application), thereby outputting results with abnormal causes.
  • abnormal causes such as water shortage, nutrient deficiency, insect damage, etc.
  • the inspection system also includes a storage module, a database and a comparison module.
  • the storage module is connected to the output end of the anomaly detection module and interacts with the database to classify and store the captured images containing ginger abnormality detection frames; the comparison module is respectively connected with the storage module. Module and database connection, so that when an abnormally growing ginger image occurs, it can be compared with images of the same area in different time periods stored in the database (comparison in the time dimension) to obtain more detailed analysis results.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Multimedia (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Guiding Agricultural Machines (AREA)

Abstract

一种基于后台图像定位的无人机生姜种植巡检系统，包括无人机模块、无人机摄像模块、异常检测模块及异常定位模块；无人机模块上设置GPS定位模块；无人机摄像模块安装在无人机模块上，包括摄像头（100）、垂直发射模块与角度发射模块，且垂直发射模块与角度发射模块均包括激光测距仪与可见红外激光发射器；异常定位模块包括可见光红外激光点检测模块与修正模块。巡检系统能够高效的对生姜苗的生长情况进行监测，且对异常苗的定位准确、误差小，确保对异常苗及时、有效的处理，避免异常苗影响周围姜苗，从而保证生姜的高产及品质。

Description

一种基于后台图像定位的无人机生姜种植巡检系统 技术领域
本发明涉及生姜种植技术领域，具体涉及一种基于后台图像定位的无人机生姜种植巡检系统。
背景技术
在生姜种植过程中，从出苗到收获需经历180天、共四个阶段；生姜种植的前两个阶段为幼苗期，其根系较弱、幼小，容易生病，从而导致生姜苗无法正常生长、甚至影响周围幼苗，造成生姜种植田减产、生姜成品率低、食用效果差等问题。目前，人们通常利用现场拍摄照片的方式获取农田内的种苗数据指标；然而，生姜苗矮小且姜苗与姜苗之间间隔小，此外，影像特征具有一定的误差(由拍摄设备本身或拍摄环境等导致)，造成对异常姜苗的定位不准确、定位精度低，无法有效获取异常苗的位置，影响后续处理效率。
发明内容
本发明的目的在于提供一种基于后台图像定位的无人机生姜种植巡检系统，该巡检系统能够高效的对生姜苗的生长情况进行监测，且对异常苗的定位准确、误差小，确保对异常苗及时、有效的处理，避免异常苗异响周围姜苗，从而保证生姜的高产及品质。
本发明目的通过如下技术方案实现：
一种基于后台图像定位的无人机生姜种植巡检系统，其特征在于：包括无人机模块、无人机摄像模块、异常检测模块及异常定位模块；
所述无人机模块上设置GPS定位模块，用于获取拍摄点GPS定位值；
所述无人机摄像模块安装在无人机模块上，包括摄像头、垂直发射模块与角度发射模块；所述摄像头固定安装在无人机模块的正下方且镜头垂直向下；垂直发射模块与角度发射模块分别设置在摄像头的左、右两侧，且垂直发射模块与摄像头平行设置、即垂直发射模块的发射线垂直向下，角度发射模块与摄像头呈角度a设置、即角度发射模块的发射线与摄像头视线呈角度a；所述垂直发射模块与角度发射模块均包括激光测距仪与可见红外激光发射器；
所述异常检测模块，采用线下训练完成生姜异常生长图像检测模型对无人机摄像模块传输的拍摄图像进行检测，输出生姜异常检测框；
所述异常定位模块包括可见光红外激光点检测模块与修正模块。
进一步，所述无人机模块还包括无人机飞控模块与无线传输模块，所述无人机飞控模块用于控制无人机模块的飞行、降落、转向等； 所述无线传输模块用于远程连接无人机摄像模块与异常检测模块，从而将无人机摄像模块的拍摄图像、角度a及拍摄点GPS定位值传输给异常检测模块。
进一步，所述异常检测模块通过筛选带标注的生姜异常生长图片组成数据集，采用识别框架进行训练，从而获得生姜异常生长图像检测模型。
优选的，所述识别框架可采用YOLOv3模型、YOLOv4模型、YOLOv5模型中的任一种。
进一步，所述生姜异常检测框为矩形框，生姜异常生长图像检测模型输出数据包括矩形框的左上角坐标及矩形框的宽、高。
进一步，所述可见红外激光点检测模块用于检测无人机摄像模块传输的拍摄图像中可见红外激光发射器发出的红外激光点；所述修正模块根据拍摄点的GPS定位值、角度a、生姜异常检测框、红外激光点，修正图像中生姜异常位置的GPS定位值。
进一步，所述修正图像中生姜异常位置的GPS定位值的具体步骤包括：
S01、通过垂直发射模块与角度发射模块的激光点物理距离获取两个激光测距仪、即两个激光点在地面的1/2间实际距离d，具体为：
当种植地为斜面时：
Figure PCTCN2022120039-appb-000001
式中，x表示垂直发射模块的激光测距仪与地面之间的距离；y表示角度发射模块的激光测距仪与地面之间的距离；
当种植地为平面时，即x＝y·cos a：
d＝y sinα＝x tanα；
S02、获取图像像素距离d p与实际距离d之间的映射关系β，具体为：
Figure PCTCN2022120039-appb-000002
式中，d p通过测量垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器在拍摄图像中的红外激光点之间的像素值距离获得；
S03、获取生姜异常检测框位置图像的像素距离与角度，具体为：
通过拍摄图像测量出生姜异常检测框中心与拍摄图像中垂直发射模块的红外激光点之间的像素距离d l；
通过拍摄图像测量出第一连线与第二连线之间的逆时针方向夹 角θ；其中，第一连线为生姜异常检测框中心与垂直发射模块的红外激光点之间的连线，第二连线为垂直发射模块的红外激光点与角度发射模块的红外激光点之间的连线；
S04、获取生姜异常检测框位置图像的实际距离d F与实际角度θ F，具体为：
d F＝d l·β；
θ F＝θ+θ l；
式中，θ l为无人机拍摄图片时，发射器平面与正北方向之间的逆时针夹角；其中，发射器平面垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器组成的平面；
S05、通过生姜异常检测框位置图像的实际距离d F与实际角度θ F、结合此时拍摄点GPS定位值，能够获取实际生姜异常生长苗的矩形框，从而实现异常苗的准确定位。
作进一步优化，所述巡检系统还包括存储模块、数据库及比对模块，所述存储模块与异常检测模块的输出端连接、且与数据库交互，从而将含有生姜异常检测框的拍摄图像进行分类储存；所述比对模块分别与存储模块、数据库连接，从而当发生异常生长的生姜图像时、与储存在数据库中的同一区域的不同时间段的图像进行对比(时间维度上进行比较)，获取更详细的分析结果。
本发明具有如下技术效果：
本申请通过无人机模块实现对生姜种植田内进行巡检，自动化程度高、巡检范围广，配合无人机摄像模块与异常检测模块，能够高效、快速、及时的识别出异常生长的生姜苗，有效节省人力物力，识别效率高；通过可见光红外激光点检测模块与修正模块的配合，通过后台图像即可获得异常生长生姜苗的位置定位，定位精确、误差小，能够精准通过后台图像获得前端种植田内的异常生长生姜苗的位置，降低系统前端(即无人机模块)的硬件需求及算力需求，从而降低无人机模块的操作难度、运行成本，提高巡检识别的效率(如无人机模块需降低高度或落地才能准确获得实际中异常生姜苗的位置，既提高了无人机模块的运行成本，又十分考验操作人员的无人机操作水平，同时还拉长了识别巡检的时间、降低巡检效率)。同时，本申请的无人机前端仅用于拍摄图像与巡检飞行，将需要复杂运算判断与逻辑处理的部分交给后端进行，能够有效降低前端无人机的运算工作量，避免无人机模块过载(由于无人机模块设置空间有效)而出现短路、运行缓慢等问题；此外，通过后端的训练、识别及判断，精准度更高、判断更及时，从而为异常苗的处理和维护提供更充分的时间。
附图说明
图1为本发明实施例中获取两个激光测距仪在地面的实际距离的结构示意图。
其中，100、摄像头。
具体实施方式
下面通过实施例对本发明进行具体的描述，有必要在此指出的是以下实施例只用于对本发明进行进一步说明，不能理解为对本发明保护范围的限制，基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
实施例1：
一种基于后台图像定位的无人机生姜种植巡检系统，其特征在于：包括无人机模块、无人机摄像模块、异常检测模块及异常定位模块；
无人机模块上设置GPS定位模块，用于获取拍摄点GPS定位值；无人机模块还包括无人机飞控模块与无线传输模块，无人机飞控模块用于控制无人机模块的飞行、降落、转向等；无线传输模块用于远程连接无人机摄像模块与异常检测模块，从而将无人机摄像模块的拍摄图像、角度a及拍摄点GPS定位值传输给异常检测模块。同时，无人机模块上还可以配置电池组、电池充放电电路及太阳能电池板，从而利用太阳能进行充电、提高无人机模块的续航能力(无人机模块可采用现有常见的农业生产用无人机)。
无人机摄像模块安装在无人机模块上，包括摄像头100、垂直发射模块与角度发射模块；摄像头100固定安装在无人机模块的正下方(具体为安装在无人机底部)且镜头垂直向下(即摄像角度垂直向下)；垂直发射模块与角度发射模块分别设置在摄像头100的左、右两侧，且垂直发射模块与摄像头100平行设置、即垂直发射模块的发射线垂直向下，角度发射模块与摄像头100呈角度a设置、即角度发射模块的发射线与摄像头100视线呈角度a(如图1所示)；垂直发射模块与角度发射模块均包括激光测距仪与可见红外激光发射器；
异常检测模块通过筛选带标注的生姜异常生长图片(即筛选以往生姜异常生长图片且进行标注)组成数据集(数据集包括训练集、测试集与验证集)，采用识别框架进行训练(训练方式采用本领域常规图像训练方式，即利用训练集得到初模型，再利用验证集对初模型进行验证，获得生姜异常生长图像检测模型)，识别框架可采用YOLOv3模型、YOLOv4模型、YOLOv5模型中的任一种(优选YOLOv5模型)，从而获得生姜异常生长图像检测模型；然后采用线下训练完成生姜异常生长图像检测模型对无人机摄像模块传输的拍摄图像进行检测，输出生姜异常检测框；生姜异常检测框为矩形框，则生姜异常生长图像 检测模型输出数据包括矩形框的左上角坐标及矩形框的宽、高。
异常定位模块包括可见光红外激光点检测模块与修正模块；
可见红外激光点检测模块用于检测无人机摄像模块传输的拍摄图像中可见红外激光发射器发出的红外激光点；修正模块根据拍摄点的GPS定位值、角度a、生姜异常检测框、红外激光点，修正图像中生姜异常位置的GPS定位值。
修正图像中生姜异常位置的GPS定位值的具体步骤包括：
S01、通过垂直发射模块与角度发射模块的激光点物理距离获取两个激光测距仪、即两个激光点在地面的1/2间实际距离d，如图1所示，具体为：
当种植地为斜面时：
Figure PCTCN2022120039-appb-000003
式中，x表示垂直发射模块的激光测距仪与地面之间的距离；y表示角度发射模块的激光测距仪与地面之间的距离；
当种植地为平面时，即x＝y·cos a：
d＝y sinα＝x tanα ；
S02、获取图像像素距离d p与实际距离d之间的映射关系β，具体为：
Figure PCTCN2022120039-appb-000004
式中，d p通过测量垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器在拍摄图像中的红外激光点之间的像素值距离获得；
S03、获取生姜异常检测框位置图像的像素距离与角度，具体为：
通过拍摄图像测量出生姜异常检测框中心与拍摄图像中垂直发射模块的红外激光点之间的像素距离d l；
通过拍摄图像测量出第一连线与第二连线之间的逆时针方向夹角θ；其中，第一连线为生姜异常检测框中心与垂直发射模块的红外激光点之间的连线，第二连线为垂直发射模块的红外激光点与角度发射模块的红外激光点之间的连线；
S04、获取生姜异常检测框位置图像的实际距离d F与实际角度θ F，具体为：
d F＝d l·β；
θ F＝θ+θ l；
式中，θ l为无人机拍摄图片时，发射器平面与正北方向之间的逆时针夹角；其中，发射器平面垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器组成的平面；
S05、通过生姜异常检测框位置图像的实际距离d F与实际角度θ F、结合此时拍摄点GPS定位值，能够获取实际生姜异常生长苗的矩形框，从而实现异常苗的准确定位。
实施例2：
异常检测模块中还可以设置分类模块，用于输出具有的异常原因估计，如缺水、缺营养、虫害等；即异常检测模块获得生姜异常生长图像检测模型后、输入到预先训练完成的分类模块中进行分类(分类模型可采用支撑向量机或深度神经网络中任一种，如VGG、RexNet等，本领域技术人员能够理解，本申请不做过多论述)，从而输出具有异常原因的结果。
实施例3：
巡检系统还包括存储模块、数据库及比对模块，存储模块与异常检测模块的输出端连接、且与数据库交互，从而将含有生姜异常检测框的拍摄图像进行分类储存；比对模块分别与存储模块、数据库连接，从而当发生异常生长的生姜图像时、与储存在数据库中的同一区域的不同时间段的图像进行对比(时间维度上进行比较)，获取更详细的分析结果。

Claims (5)

  1. 一种基于后台图像定位的无人机生姜种植巡检系统，其特征在于：包括无人机模块、无人机摄像模块、异常检测模块及异常定位模块；
    所述无人机模块上设置GPS定位模块，用于获取拍摄点GPS定位值；
    所述无人机摄像模块安装在无人机模块上，包括摄像头、垂直发射模块与角度发射模块；所述摄像头固定安装在无人机模块的正下方且镜头垂直向下；垂直发射模块与角度发射模块分别设置在摄像头的左、右两侧，且垂直发射模块与摄像头平行设置、即垂直发射模块的发射线垂直向下，角度发射模块与摄像头呈角度a设置、即角度发射模块的发射线与摄像头视线呈角度a；所述垂直发射模块与角度发射模块均包括激光测距仪与可见红外激光发射器；
    所述异常检测模块，采用线下训练完成生姜异常生长图像检测模型对无人机摄像模块传输的拍摄图像进行检测，输出生姜异常检测框；
    所述异常定位模块包括可见光红外激光点检测模块与修正模块。
  2. 根据权利要求1所述的无人机生姜种植巡检系统，其特征在于：所述无人机模块还包括无人机飞控模块与无线传输模块；所述无线传输模块用于远程连接无人机摄像模块与异常检测模块。
  3. 根据权利要求2所述的无人机生姜种植巡检系统，其特征在于：所述异常检测模块通过筛选带标注的生姜异常生长图片组成数据集，采用识别框架进行训练，从而获得生姜异常生长图像检测模型。
  4. 根据权利要求3所述的无人机生姜种植巡检系统，其特征在于：所述可见红外激光点检测模块用于检测无人机摄像模块传输的拍摄图像中可见红外激光发射器发出的红外激光点；所述修正模块根据拍摄点的GPS定位值、角度a、生姜异常检测框、红外激光点，修正图像中生姜异常位置的GPS定位值
  5. 根据权利要求4所述的无人机生姜种植巡检系统，其特征在于：所述修正图像中生姜异常位置的GPS定位值的具体步骤包括：
    S01、通过垂直发射模块与角度发射模块的激光点物理距离获取两个激光测距仪、即两个激光点在地面的1/2间实际距离d，具体为：
    当种植地为斜面时：
    Figure PCTCN2022120039-appb-100001
    式中，x表示垂直发射模块的激光测距仪与地面之间的距离；y表示角度发射模块的激光测距仪与地面之间的距离；
    当种植地为平面时，即x＝y·cos a：
    d＝y sinα＝x tanα；
    S02、获取图像像素距离d p与实际距离d之间的映射关系β，具体为：
    Figure PCTCN2022120039-appb-100002
    式中，d p通过测量垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器在拍摄图像中的红外激光点之间的像素值距离获得；
    S03、获取生姜异常检测框位置图像的像素距离与角度，具体为：
    通过拍摄图像测量出生姜异常检测框中心与拍摄图像中垂直发射模块的红外激光点之间的像素距离d l；
    通过拍摄图像测量出第一连线与第二连线之间的逆时针方向夹角θ；其中，第一连线为生姜异常检测框中心与垂直发射模块的红外激光点之间的连线，第二连线为垂直发射模块的红外激光点与角度发射模块的红外激光点之间的连线；
    S04、获取生姜异常检测框位置图像的实际距离d F与实际角度θ F，具体为：
    d F＝d l·β；
    θ F＝θ+θ l；
    式中，θ l为无人机拍摄图片时，发射器平面与正北方向之间的逆时针夹角；其中，发射器平面垂直发射模块的可见红外激光发射器与角度发射模块的可见红外激光发射器组成的平面；
    S05、通过生姜异常检测框位置图像的实际距离d F与实际角度θ F、结合此时拍摄点GPS定位值，能够获取实际生姜异常生长苗的矩形框，从而实现异常苗的准确定位。
PCT/CN2022/120039 2022-07-12 2022-09-20 一种基于后台图像定位的无人机生姜种植巡检系统 WO2024011750A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210813383.9 2022-07-12
CN202210813383.9A CN115355888B (zh) 2022-07-12 2022-07-12 一种基于后台图像定位的无人机生姜种植巡检系统

Publications (1)

Publication Number Publication Date
WO2024011750A1 true WO2024011750A1 (zh) 2024-01-18

Family

ID=84032071

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/120039 WO2024011750A1 (zh) 2022-07-12 2022-09-20 一种基于后台图像定位的无人机生姜种植巡检系统

Country Status (3)

Country Link
CN (1) CN115355888B (zh)
LU (1) LU506358B1 (zh)
WO (1) WO2024011750A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090284644A1 (en) * 2008-05-12 2009-11-19 Flir Systems, Inc. Optical Payload with Folded Telescope and Cryocooler
CN105698742A (zh) * 2016-02-29 2016-06-22 北方民族大学 一种基于无人机的快速土地面积测量装置及测量方法
CN108111807A (zh) * 2017-11-20 2018-06-01 宁德师范学院 一种电力巡线设备及故障诊断方法
CN109598215A (zh) * 2018-11-22 2019-04-09 仲恺农业工程学院 一种基于无人机定位拍摄的果园建模分析系统和方法
CN215767057U (zh) * 2021-12-27 2022-02-08 浙江公路水运工程咨询有限责任公司 一种提高无人机调查复杂边坡岩体精度的动态调整装置
JP2022043644A (ja) * 2020-09-04 2022-03-16 株式会社セベック 無人飛行体およびそれを用いた健康管理システム

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107765706A (zh) * 2017-10-17 2018-03-06 山东交通学院 船舶无人机舱火灾巡检用四旋翼飞行器及其控制方法
CN208110052U (zh) * 2018-04-27 2018-11-16 四川海讯电子开发集团有限公司 一种无人机目标定位系统
CN113916187B (zh) * 2020-07-07 2024-04-30 中国电信股份有限公司 基于无人机的基站天线下倾角测量方法、装置和系统
WO2022094854A1 (zh) * 2020-11-05 2022-05-12 深圳市大疆创新科技有限公司 农作物的生长监测方法、设备及存储介质
CN114265418A (zh) * 2021-09-03 2022-04-01 国家电投集团江苏新能源有限公司 一种用于光伏电站的无人机巡检与缺陷定位系统及方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090284644A1 (en) * 2008-05-12 2009-11-19 Flir Systems, Inc. Optical Payload with Folded Telescope and Cryocooler
CN105698742A (zh) * 2016-02-29 2016-06-22 北方民族大学 一种基于无人机的快速土地面积测量装置及测量方法
CN108111807A (zh) * 2017-11-20 2018-06-01 宁德师范学院 一种电力巡线设备及故障诊断方法
CN109598215A (zh) * 2018-11-22 2019-04-09 仲恺农业工程学院 一种基于无人机定位拍摄的果园建模分析系统和方法
JP2022043644A (ja) * 2020-09-04 2022-03-16 株式会社セベック 無人飛行体およびそれを用いた健康管理システム
CN215767057U (zh) * 2021-12-27 2022-02-08 浙江公路水运工程咨询有限责任公司 一种提高无人机调查复杂边坡岩体精度的动态调整装置

Also Published As

Publication number Publication date
CN115355888B (zh) 2024-03-15
LU506358B1 (en) 2024-04-08
CN115355888A (zh) 2022-11-18

Similar Documents

Publication Publication Date Title
US20210358106A1 (en) Crop yield prediction method and system based on low-altitude remote sensing information from unmanned aerial vehicle
CN108881825A (zh) 基于Jetson TK1的水稻杂草无人机监控系统及其监控方法
WO2022094854A1 (zh) 农作物的生长监测方法、设备及存储介质
JP5020444B2 (ja) 作物生育量測定装置、作物生育量測定方法、作物生育量測定プログラム及びその作物生育量測定プログラムを記録したコンピュータ読取可能な記録媒体
JP7074126B2 (ja) 画像処理装置、生育調査画像作成システム及びプログラム
CN109685858A (zh) 一种单目摄像头在线标定方法
CN112113542A (zh) 一种无人机航摄建设用地土地专项数据验收的方法
CN112489130A (zh) 一种输电线路与目标物的距离测量方法、装置及电子设备
CN106708075B (zh) 基于固定翼无人机的大范围油菜田spad值遥感系统及采集方法
CN111426309A (zh) 一种基于三维地形测绘数据的采集处理方法
CN114488099A (zh) 一种激光雷达系数标定方法、装置、电子设备及存储介质
CN108007437B (zh) 一种基于多旋翼飞行器测量农田边界与内部障碍的方法
WO2024011750A1 (zh) 一种基于后台图像定位的无人机生姜种植巡检系统
CN112837314B (zh) 基于2D-LiDAR和Kinect的果树冠层参数检测系统和方法
CN114442665A (zh) 基于无人机的风电叶片巡检线路规划方法
CN111598937A (zh) 一种基于标定区块对靶校正的农田测亩方法及系统
CN111814585A (zh) 无人机近地空作物苗情遥感监测方法、装置及存储介质
CN116577800A (zh) 基于系统噪声估计的光电吊舱自适应ekf目标定位方法
Li et al. Prediction of wheat gains with imagery from four-rotor UAV
CN116202489A (zh) 输电线路巡检机与杆塔协同定位方法及系统、存储介质
Manish et al. Agbug: Agricultural robotic platform for in-row and under canopy crop monitoring and assessment
WO2023019445A1 (zh) 图像处理方法、无人飞行器和存储介质
CN114830911A (zh) 智能除草方法、装置和存储介质
CN209820522U (zh) 一种基于小型无人机的作物干旱预警巡测系统
CN116543309B (zh) 一种作物异常信息获取方法、系统、电子设备及介质

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: 22950861

Country of ref document: EP

Kind code of ref document: A1