[go: up one dir, main page]

CN111096107A - Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy - Google Patents

Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy Download PDF

Info

Publication number
CN111096107A
CN111096107A CN201911369918.2A CN201911369918A CN111096107A CN 111096107 A CN111096107 A CN 111096107A CN 201911369918 A CN201911369918 A CN 201911369918A CN 111096107 A CN111096107 A CN 111096107A
Authority
CN
China
Prior art keywords
orchard
weeding
mode
tree
weeder
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN201911369918.2A
Other languages
Chinese (zh)
Other versions
CN111096107B (en
Inventor
李佳
吕程序
王辉
苑严伟
韩娜娜
张帅扬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chinese Academy of Agricultural Mechanization Sciences
Original Assignee
Chinese Academy of Agricultural Mechanization Sciences
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 Chinese Academy of Agricultural Mechanization Sciences filed Critical Chinese Academy of Agricultural Mechanization Sciences
Priority to CN201911369918.2A priority Critical patent/CN111096107B/en
Publication of CN111096107A publication Critical patent/CN111096107A/en
Application granted granted Critical
Publication of CN111096107B publication Critical patent/CN111096107B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
    • A01B39/00—Other machines specially adapted for working soil on which crops are growing
    • A01B39/12—Other machines specially adapted for working soil on which crops are growing for special purposes, e.g. for special culture
    • A01B39/16—Other machines specially adapted for working soil on which crops are growing for special purposes, e.g. for special culture for working in vineyards, orchards, or the like ; Arrangements for preventing damage to vines
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
    • A01B39/00—Other machines specially adapted for working soil on which crops are growing
    • A01B39/12—Other machines specially adapted for working soil on which crops are growing for special purposes, e.g. for special culture
    • A01B39/18—Other machines specially adapted for working soil on which crops are growing for special purposes, e.g. for special culture for weeding
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
    • A01B39/00—Other machines specially adapted for working soil on which crops are growing
    • A01B39/20—Tools; Details
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
    • A01B39/00—Other machines specially adapted for working soil on which crops are growing
    • A01B39/20—Tools; Details
    • A01B39/26—Arrangements for protecting plants, e.g. fenders
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes

Landscapes

  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Environmental Sciences (AREA)
  • Soil Sciences (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Guiding Agricultural Machines (AREA)

Abstract

An orchard weeding machine comprises a dual-mode weeding device below a crown, and weed removal is carried out by the dual-mode weeding method below the crown. The weeding method comprises a common working mode and a crown working mode, the orchard weeding machine switches the common working mode and the crown working mode based on set conditions, the orchard weeding machine executes the common working mode in a non-crown lower area, and the orchard weeding machine starts the crown working mode in the crown lower area, and comprises the following steps: starting up the system to initialize and start a common working mode; according to the tree trunk detection result and the Hall sensor monitoring result, starting a tree crown working mode when the orchard weeding machine enters the position below the tree crown; if not, the common working mode is continuously executed; judging that the orchard weeding machine leaves the area under the trees and starts the common working mode after the weeds under the trees are cleared; if not, the crown working mode is continuously executed.

Description

Orchard weeding machine, and double-mode weeding method and weeding device below tree crown of orchard weeding machine
Technical Field
The invention relates to an agriculture and forestry planting maintenance and guarantee technology, in particular to an orchard weeding machine, a dual-mode weeding method and a weeding device below a crown of the orchard weeding machine.
Background
In the orchard production process, weeds and crops compete for light, water, fertilizer and gas, and the crop viability is reduced. In the selection of the weeding mode, not only the economic benefit problem but also the product quality and safety problem are considered, and the agroecological (environmental and biological chain) problem is considered. The weeding methods are various, and mechanical weeding is a common weeding mode in agricultural production processes (such as weeding in orchards, furrows, ridges, channels, the periphery of houses, mountain forests and the like).
The foreign weeding technology is perfect, the performance is reliable, and a whole set of mature protective tillage orchard weeding machine is formed through years of research and improvement, so that the production, application and operation requirements can be well met. The orchard weeding machine is late in research and development start in China, and has certain differences in technology, manufacturing means, process and the like compared with developed countries in Europe and America. With the development of agricultural modernization, a multipurpose composite type working machine appears. The novel orchard weeding machine is reasonable in speed parameter and breadth, high in intelligent degree, capable of achieving multiple purposes and meeting different farming purposes and weeding requirements. At present, the orchard weeding machine in China has a lot of improvements in product variety, matching property, product quality and technical level. Practical experience shows that the adoption of agricultural machinery with high intellectualization and automation degree is an effective way for developing high-efficiency cost-saving agriculture, and the orchard weeding machine is no exception. With the comprehensive application of computer technology, information technology and control technology to orchard weeding machines, the research on products organically combining field automatic navigation technology, visual system of orchard weeding machine, GPS positioning technology, high-precision machine, electricity, liquid and gas integration and orchard weeding machines is more and more emphasized, and the products are more and more important for the development of orchard weeding machines in the future.
In an orchard environment, weeds below the crown of a fruit tree compete with crop nutrients extremely violently, and the illumination condition is poor. In the orchard weeding process, the orchard weeding machine walks between fruit tree rows and around fruit trees, and certain influence is caused on the robustness of the weeding orchard weeding machine due to the switching of working positions and the change of working sub-scenes. The common orchard weeding machine based on vision in the prior art easily causes misjudgment under the trunk to damage the trunk, and weeds around the trunk are easily removed insufficiently thoroughly. Meanwhile, the growth vigor of weeds in the orchard is higher, the weeds are different from weeds on a common lawn, and the orchard weeding machine based on the radar or other sensors is easily interfered by the weeds with higher plant heights.
Disclosure of Invention
The invention aims to solve the technical problems in the prior art and provides an orchard weeding machine, a dual-mode weeding method below a crown and a weeding device of the orchard weeding machine.
In order to achieve the above object, the present invention provides a dual-mode weeding method under a crown of an orchard weeding machine, wherein the orchard weeding machine includes a normal operation mode and a crown operation mode, the orchard weeding machine switches the normal operation mode and the crown operation mode based on a set condition, the orchard weeding machine executes the normal operation mode in a non-crown-below area, the orchard weeding machine starts the crown operation mode in the crown-below area, and the dual-mode weeding method under the crown includes the steps of:
s100, starting the orchard weeding machine to initialize a system and starting a common working mode;
s200, the common working mode comprises the following steps: acquiring a color image and a depth image of an orchard in real time, detecting a trunk by using the color image, determining the position of a central point of the trunk, determining the relative distance between the central point of the trunk and an orchard weeding machine by combining the corresponding depth image, and enabling the orchard weeding machine to move towards the trunk with the shortest distance;
s300, judging whether the orchard weeding machine enters the position below a tree crown or not according to a trunk detection result and a Hall sensor monitoring result, and starting a tree crown working mode if the orchard weeding machine enters the position below the tree crown; if not, continuing to execute the common working mode;
s400, the crown working mode comprises the following steps: acquiring the position of a collision point of the trunk and the orchard weeding machine, and calculating the horizontal distance between the trunk and the orchard weeding machine; judging the relative position of the trunk and the orchard weeding machine according to the position of the collision point, and if the trunk is on the right side of the orchard weeding machine, controlling the orchard weeding machine to horizontally avoid the trunk and then cleaning weeds clockwise; if the trunk is on the left side or right ahead of the orchard weeder, controlling the orchard weeder to horizontally avoid the trunk and then cleaning weeds anticlockwise; and
s500, judging whether the weeds under the trees are completely cleared or not, if so, enabling the orchard weeding machine to leave the trees and starting the common working mode; if not, the crown working mode is continuously executed.
The dual-mode weeding method under the crown of the orchard weeding machine further comprises the following steps of:
s501, recording an initial point of the orchard weeding machine entering the lower part of the crown, and extracting and recording the characteristics of the image at the initial point; and
s502, acquiring a real-time color image, extracting features, comparing the feature with the image of the starting point, judging that the orchard weeding machine returns to the starting point if the similarity is greater than a threshold value SI, finishing under-tree cleaning, and starting the common working mode; and if the similarity is smaller than or equal to the threshold value SI, judging that the tree is not completely cleaned, and continuously executing the crown working mode.
The double-mode weeding method under the crown of the orchard weeding machine further comprises the following steps:
s600, judging whether the operation is stopped or not, if so, exiting the common working mode and shutting down the orchard weeding machine; if not, step S200 is executed.
The dual-mode weeding method under the crown of the orchard weeding machine comprises the following steps of:
the setting conditions of the common working mode are as follows: the fruit tree position can be detected, and the depth distance is greater than a threshold value S; the information returned by the pressure sensor is that the orchard weeding machine has no collision; and the information returned by the Hall sensor is that the orchard weeding machine normally advances;
the set conditions of the crown working mode are as follows: the fruit tree position cannot be detected, and the depth distance is less than or equal to a threshold value S; the information returned by the pressure sensor is that the front part of the orchard weeding machine is collided; and the information returned by the Hall sensor is that the orchard weeding machine is abnormally stopped;
and in the movement process of the orchard weeding machine, if all set conditions of the common working mode or the crown working mode are met at the same time, the orchard weeding machine is immediately switched into a corresponding mode, otherwise, the orchard weeding machine keeps the current state until the mode is switched.
The dual-mode weeding method under the tree crown of the orchard weeding machine comprises the following steps of:
judgment of
Figure BDA0002339398780000031
Determines the location of the collision point:
when f is larger than 0, the trunk is arranged on the right side of the orchard weeding machine, the orchard weeding machine horizontally moves leftwards for a distance of D-D to avoid the trunk, and a clockwise route is adopted for removing weeds;
when f is less than 0, the tree trunk is arranged on the left side of the orchard weeding machine, the orchard weeding machine horizontally moves to the right for a distance of D-D, and a counterclockwise route is adopted for removing weeds;
when f is equal to 0, the trunk is right in front of the orchard weeding machine, the orchard weeding machine horizontally moves to the right for a distance D, and a counterclockwise route is adopted for removing weeds;
wherein f is a collision point discrimination identifier, P is a unique collision point index value, N is a discrete number of pressure sensors, D is a width of a weeding member, and D is a horizontal distance between the orchard weeding machine and a trunk.
In the dual-mode weeding method under the crown of the orchard weeding machine, the pressure sensors uniformly distributed at the front end of the weeding mechanism are uniformly divided into N induction blocks from left to right, the returned data of the pressure sensors are expressed as 1 xN-dimensional vectors, and if K is a sensor receiving vector and K is a sensor receiving vectoriThe returned value of the ith sensing block is i ═ 1, 2., N; and T is a threshold value, when the return value of the induction block is greater than or equal to T, the induction block is the collision point, and a sensor measurement result R is calculated by adopting the following formula:
Figure BDA0002339398780000041
get RiThe median of all positions 1 is the collision point, and the unique collision point index value p is:
Figure BDA0002339398780000042
wherein q isjIs RiSense Block index value of 1, m is all R i1 induction blockAnd the total number of the index values, and the horizontal distance d between the orchard weeding machine and the trunk is calculated according to the position of the collision point p:
Figure BDA0002339398780000043
the dual-mode weeding method under the crown of the orchard weeding machine further comprises the following steps of:
to input an image IuAfter ORB feature extraction, the feature is expressed as an n-dimensional descriptor set Iu→{d1...dn}，diIs the characteristic point of ORB descriptor and the visual word in visual dictionary
Figure BDA0002339398780000044
Corresponding; by comparing the input image with the starting point image I when entering the crownvPerforming similarity calculation to obtain image IuAnd IvSimilarity of (A) to (B) S (I)u,Iv) Can be calculated by the following formula:
Figure BDA0002339398780000045
wherein, UiAs an image IuIs a vector of (a) represents an element, viAs an image IvThe vector of (a) represents an element,Nfis the number of all features in the visual dictionary.
In order to better achieve the above object, the present invention further provides an under-crown dual-mode weeding apparatus for an orchard weeding machine, comprising:
the controller is arranged on the orchard weeding machine and is connected with the driving mechanism of the orchard weeding machine;
the image acquisition device is used for acquiring an orchard image to detect a trunk and determine the position of a central point of the trunk, the image acquisition device is installed right ahead of the orchard weeding machine and connected with the controller, and a camera of the image acquisition device horizontally faces right ahead;
the Hall element pair is connected with the controller and comprises a first Hall sensor and a second Hall sensor, the first Hall sensor is installed at the tail of the orchard weeding machine, and the second Hall sensor is installed on a rear bearing of the orchard weeding machine and is arranged corresponding to the first Hall sensor; and
the pressure sensing area for real-time detection returns the orchard weeder with the pressure value at the collision position of trunk converts to the voltage value, the pressure sensing area includes a plurality of pressure sensor, a plurality of pressure sensor evenly set up the weeding mechanism front end of orchard weeder, and respectively with the controller is connected.
The double-mode weeding device below the crown of the orchard weeding machine is characterized in that the image acquisition device is a color camera, and the color camera is connected with the controller through a USB interface.
In order to better achieve the above object, the present invention further provides an orchard weeding machine, wherein the orchard weeding machine comprises the above under-crown dual-mode weeding device, and weed removal is performed by adopting the above under-crown dual-mode weeding method.
The invention has the technical effects that:
aiming at the conditions such as illumination constraint, weed distribution constraint and the like below the crown of the fruit tree, a dual-mode weeding mode below the crown of the fruit tree is established by using sensors such as vision, pressure and the like, dual modes are switched under different conditions so as to intelligently weed, and color image information and distance information in a view field are acquired by using a color depth image; returning surrounding obstacle detection information by using a peripheral pressure sensor of the machine body; processing image information obtained by the color camera to detect the trunk and determine the relative position of the trunk and the orchard weeding machine; the detection signals captured by the pressure sensors are filtered and analyzed, the position information of the trunk is obtained in a physical touch mode, under the premise that the trunk is not damaged, the weeds under the tree crown are comprehensively cleared, the weeding route is guided, and the technical support is provided for effective operation and reasonable obstacle avoidance of the orchard weeding machine.
The invention is described in detail below with reference to the drawings and specific examples, but the invention is not limited thereto.
Drawings
Fig. 1 is a schematic structural view of an orchard weeding machine according to an embodiment of the invention;
fig. 2 is a schematic view illustrating a trunk detection of a fruit tree according to an embodiment of the present invention;
FIG. 3 is a schematic view of a pressure sensing strip distribution according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating a pressure signal analysis according to an embodiment of the present invention;
FIG. 5 is a schematic view of a dual mode under crown weeding method according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a weeding path of a trunk on the right side of an orchard weeding machine according to an embodiment of the invention;
fig. 7 is a schematic diagram of a weeding path of a trunk on the left side of an orchard weeding machine according to an embodiment of the invention.
Wherein the reference numerals
1 orchard weeding machine
11 weeding mechanism
12 weeding knife
2 wheel
3 first Hall sensor
4 second Hall sensor
5 image acquisition device
6 pressure sensing belt
7 trunk
71 rectangular frame
72 Cross mark
73 collision point
Detailed Description
The invention will be described in detail with reference to the following drawings, which are provided for illustration purposes and the like:
referring to fig. 1, fig. 1 is a schematic structural view of an orchard weeding machine according to an embodiment of the invention. The damage of the trunk 7 (also can be called as a tree body) should be avoided when weeds below the crown are cleared away, the intelligent orchard weeding machine 1 based on pure vision in the prior art is easily influenced by light, the content of a view field is reduced when the intelligent orchard weeding machine is close to a fruit tree, weeds and the trunk 7 are difficult to distinguish, and the trunk 7 is easily damaged when the weeds are cleared away blindly. The orchard weeding machine 1 based on the sensor group is extremely easy to be interfered in an orchard with dense weed distribution and high growth vigor, and the orchard weeding machine 1 is poor in working robustness in a complex environment. The orchard weeding machine 1 comprises a dual-mode weeding device below a tree crown, and weed removal is carried out by adopting a dual-mode weeding method below the tree crown. Based on the visual technology, the data acquisition technology, the multi-data fusion technology and the like of the orchard weeding machine, the image information in the field of view is acquired by the camera, the trunk 7 is identified, and the relative position of the orchard weeding machine 1 and the trunk 7 is determined. The position of the collision point 73 can be acquired by the pressure sensor, and the positional relationship of the trunk 7 with respect to the orchard mower 1 is estimated. And finally, judging the advancing state of the orchard weeding machine 1 in real time by using methods such as loop detection, a Hall sensor and the like, establishing a dual-mode working mode, and removing weeds around the trunk 7. This orchard weeder 1 includes ordinary mode and crown mode, can be referred to as ordinary mode and crown mode for short, orchard weeder 1 carries out based on the settlement condition ordinary mode and crown mode's switching, orchard weeder 1 carries out ordinary mode when not crown below region, orchard weeder 1 opens crown mode when crown below region. The composition, structure, mutual position relationship, connection relationship, working principle and the like of other parts of the orchard weeding machine 1 are mature prior art, so that no further description is provided herein, and only the below-tree-crown dual-mode weeding device and the below-tree-crown dual-mode weeding method of the invention are described in detail below.
Referring to fig. 1, fig. 2 and fig. 3, fig. 2 is a schematic diagram of fruit tree trunk detection according to an embodiment of the present invention, and fig. 3 is a schematic diagram of pressure sensor strip distribution according to an embodiment of the present invention. The invention discloses a dual-mode weeding device below a tree crown, which comprises: a controller (not shown) installed on the orchard mower 1 and connected to a driving mechanism of the orchard mower 1; the image acquisition device 5 is used for acquiring an orchard image to detect a trunk 7 and determine the position of a center point of the trunk 7, the image acquisition device 5 is installed right ahead of the orchard weeding machine 1 and is connected with the controller, and a camera of the image acquisition device 5 horizontally faces right ahead; the Hall element pair is connected with the controller and comprises a first Hall sensor 3 and a second Hall sensor 4, the first Hall sensor 3 is installed at the tail of the orchard weeding machine 1, and the second Hall sensor 4 is installed on a rear bearing of a wheel 2 of the orchard weeding machine 1 and is arranged corresponding to the first Hall sensor 3 to form the Hall element pair; and the pressure sensing belt 6 is used for detecting and returning in real time the orchard weeding machine 1 and the pressure value of the collision position of the trunk 7 and converting the pressure value into a voltage value, the pressure sensing belt 6 comprises a plurality of pressure sensors, the pressure sensors are uniformly arranged at the front end of a weeding mechanism 11 of the orchard weeding machine 1, preferably distributed in the middle part of the weeding mechanism 11 in the vertical direction, and the pressure sensors are respectively connected with the controller. The weeding mechanism 11 comprises a plurality of evenly distributed weeding knives 12. The image acquisition device 5 is preferably a color camera which is fixed right in front of the orchard weeding machine 1 and is about 50cm away from the ground, and the color camera is connected with the controller through a USB interface.
The working states of the orchard weeding machine 1 in the area below the non-crown are all in a common working mode. In a common working mode, a color depth camera is mainly used for obtaining a front view field image of the orchard weeding machine 1, and a fruit tree trunk 7 in an RGB image is detected through an image processing algorithm. The orchard weeding machine 1 acquires depth and color images in a field of view in real time in the advancing process, positions the fruit tree trunks 7 in the color images by using a pre-established recognition algorithm, and determines the distance between the central point of each fruit tree trunk 7 and the orchard weeding machine 1, as shown in fig. 2, a rectangular frame 71 is a detection result, and a cross mark 72 is the central position of the rectangular frame 71. And selecting the trunk 7 closest to the orchard weeding machine 1 as the advancing direction to approach the fruit tree. When the orchard weeding machine 1 drives into the area below the tree crown, the visualization condition is deteriorated, and the image view field content is reduced. When weeding mechanism 11 collides with trunk 7, hall sensor detects orchard weeder 1 and advances and is hindered, when unable trunk 7 of detection in the field of view simultaneously, just gets into crown mode.
Referring to fig. 3 and 4, fig. 3 is a schematic diagram of a distribution of pressure sensor strips according to an embodiment of the invention, and fig. 4 is a schematic diagram of a pressure signal analysis according to an embodiment of the invention. In order to estimate the relative position of the trunk 7 and the orchard weeding machine 1, the front end of the weeding part is surrounded by a pressure sensing belt 6, and as shown in fig. 3, the dotted line is the pressure sensing belt 6. The sensors of the pressure sensor strip 6 can return the pressure value of the collision part in real time and convert the pressure value into corresponding voltage for analysis. When analyzing the pressure sensor data, the pressure sensor is equally divided into N sensing blocks from left to right, i.e., the sensor return data representation may be a 1 × N dimensional vector, as shown in fig. 4. Let K be the sensor receive vector, KiThe returned value of the ith sensing block is i ═ 1, 2. T is a threshold value, and when the return value of the sensing block is equal to or greater than T, the sensing block is the collision point 73. According to the principle, the induction zone result R can be calculated, and the calculation method is as follows;
Figure BDA0002339398780000081
where the vector R is the sensor measurement. To determine the unique collision point 73 of the trunk 7, R is takeniThe median of all positions at 1 is the collision point 73, i.e.:
Figure BDA0002339398780000082
in the formula qjIs RiSense Block index value of 1, m is all RiP is the unique collision point index value, which is the total number of sense blocks of 1. Knowing that the width of the weeding member is D, the horizontal distance D between the orchard weeding machine 1 and the trunk 7 can be calculated according to the position of the collision point 73, and the calculation method is as follows:
Figure BDA0002339398780000091
referring to fig. 5, fig. 5 is a schematic view of a dual-mode weeding method under a tree crown according to an embodiment of the present invention. The dual-mode weeding method under the tree crown comprises the following steps:
step S100, starting the orchard weeding machine 1 to initialize a system and start a common working mode;
step S200, the normal operation mode includes: acquiring a color image and a depth image of an orchard in real time, detecting a trunk 7 by using the color image, determining the position of a central point of the trunk 7, determining the relative distance between the central point of the trunk 7 and the orchard weeding machine 1 by combining the corresponding depth image, and enabling the orchard weeding machine 1 to move towards the trunk 7 with the shortest distance;
step S300, judging whether the orchard weeding machine 1 enters the lower part of a tree crown or not according to the detection result of the trunk 7 and the monitoring result of the Hall sensor, and starting a tree crown working mode if the orchard weeding machine 1 enters the lower part of the tree crown; if not, continuing to execute the common working mode;
step S400, the tree crown working mode comprises the following steps: acquiring the position of a collision point 73 between the trunk 7 and the orchard weeding machine 1, and calculating the horizontal distance between the trunk 7 and the orchard weeding machine 1; judging the relative position of the trunk 7 and the orchard weeding machine 1 according to the position of the collision point 73, and if the trunk 7 is on the right side of the orchard weeding machine 1, controlling the orchard weeding machine 1 to horizontally avoid the trunk 7 and then cleaning weeds clockwise; if the trunk 7 is on the left side or right in front of the orchard weeding machine 1, controlling the orchard weeding machine 1 to horizontally avoid the trunk 7 and then cleaning weeds anticlockwise; and
s500, judging whether the weeds under the trees are completely cleared, if so, leaving the orchard weeding machine 1 under the trees and starting the common working mode; if not, the crown working mode is continuously executed.
Wherein, the step of judging whether the under-tree weeds are completely cleared further comprises the following steps:
s501, recording an initial point of the orchard weeding machine 1 entering the lower part of the crown, and extracting and recording the characteristics of an image at the initial point; and
step S502, acquiring a real-time color image, extracting features, comparing the feature with the image of the starting point, judging that the orchard weeding machine 1 returns to the starting point if the similarity is greater than a threshold value, finishing under-tree cleaning, and starting the common working mode; and if the similarity is smaller than or equal to the threshold, judging that the tree is not completely cleaned, and continuously executing the crown working mode.
The embodiment may further perform loop detection, that is, performing similarity comparison with the image of the starting point further includes:
to input an image IuAfter ORB feature extraction, the feature is expressed as an n-dimensional descriptor set Iu→{d1...dn}，diIs the characteristic point of ORB descriptor and the visual word in visual dictionary
Figure BDA0002339398780000101
Corresponding; the visual dictionary constructs similar descriptor clusters through a BoW modeling method, and can be expressed as
Figure BDA0002339398780000102
IuBy different weights wiThe words and phrases of
Figure BDA0002339398780000103
Composition, weight wiIs the frequency of occurrence of each word in the entire image set, calculated by:
Figure BDA0002339398780000104
in the formula NfIs the number of all features in the visual dictionary, niIs diContains the number of features, m is the number of all features in the input image, miIs a characteristic diNumber of occurrences in the input image. If the visual dictionary contains NfA different vocabulary, then the vector of the image is represented as
Figure BDA0002339398780000105
uiCan be calculated by the following formula:
Figure BDA0002339398780000106
by comparing the input image with the starting point image I when entering the crownvPerforming similarity calculation to obtain image IuAnd IvSimilarity of (A) to (B) S (I)u,Iv) Can be calculated by the following formula:
Figure BDA0002339398780000107
wherein, UiAs an image IuIs a vector of (a) represents an element, viAs an image IvThe vector of (2) represents an element, NfIs the number of all features in the visual dictionary.
And determining whether the lower part of the crown is completely cleaned or not by comparing the similarity between the current input image and the crown starting point image. If the similarity is greater than the threshold value SI, the under-tree cleaning is determined to be finished, and the under-tree cleaning can be left and enter a common working mode; if the distance is less than or equal to the threshold value SI, the tree is not cleaned completely, and the tree crown working mode is kept until the cleaning is completed.
In this embodiment, the method may further include the following steps:
step S600, judging whether to stop operation, if so, exiting the common working mode and shutting down the orchard weeding machine 1; if not, step S200 is executed.
When the orchard weeding machine 1 moves among the fruit tree rows, the fruit tree detection can be carried out in real time based on a pre-established image recognition method, the pressure sensor in front of weeding does not have large collision, the Hall sensor judges that the orchard weeding machine 1 does not stop moving, and the orchard weeding machine 1 is in a common working mode at the moment; when the orchard weeding machine 1 moves to the lower portion of the crown of the fruit tree, the fruit tree cannot be detected due to the fact that illumination is obviously reduced and the orchard weeding machine 1 is close to the crown of the fruit tree, meanwhile, the pressure sensor returns to have a collision, the orchard weeding machine 1 stops moving due to the obstacle, and at the moment, the orchard weeding machine 1 enters a crown working mode. Wherein, the setting conditions of the common working mode and the crown working mode specifically include:
the setting conditions of the common working mode are as follows: the fruit tree position can be detected, and the depth distance is greater than a threshold value S; the information returned by the pressure sensor is that the orchard weeding machine 1 has no collision; and the information returned by the Hall sensor is that the orchard weeding machine 1 normally advances; the set conditions of the crown working mode are as follows: the fruit tree position cannot be detected, and the depth distance is less than or equal to a threshold value S; the information returned by the pressure sensor is that the front part of the orchard weeding machine 1 is collided; and the hall sensor returns information that the orchard weeding machine 1 is abnormally stopped; in the movement process of the orchard weeding machine 1, if all set conditions of the common working mode or the crown working mode are met at the same time, the orchard weeding machine 1 is immediately switched into a corresponding mode, otherwise, the orchard weeding machine 1 keeps the current state until the mode is switched.
Referring to fig. 6 and 7, fig. 6 is a schematic diagram illustrating a weeding path of a trunk on the right side of an orchard weeding machine according to an embodiment of the present invention, and fig. 7 is a schematic diagram illustrating a weeding path of a trunk on the left side of an orchard weeding machine according to an embodiment of the present invention. As can be seen from fig. 6 and 7, the clearing route of the orchard weeding machine 1 can be determined by two situations: the trunk 7 is on the left side of the orchard mower 1 and the trunk 7 is on the right side of the orchard mower 1. The collision condition can be judged
Figure BDA0002339398780000111
The sign is determined and different cleaning routes are used according to different collision states. The determining the relative position of the trunk 7 and the orchard herbicide 1 according to the position of the collision point 73 of the embodiment further includes:
judgment of
Figure BDA0002339398780000112
Determines the position of the collision point 73:
when f is larger than 0, the trunk 7 is arranged on the right side of the orchard weeder 1, the orchard weeder 1 horizontally moves leftwards for a distance of D-D to avoid the trunk 7, and weed removal is carried out by adopting a clockwise route;
when f is less than 0, the trunk 7 is arranged on the left side of the orchard weeder 1, the orchard weeder 1 horizontally moves to the right for a distance of D-D, and weeds are removed by adopting an anticlockwise route;
when f is equal to 0, the trunk 7 is right in front of the orchard weeding machine 1, the orchard weeding machine 1 horizontally moves to the right for a distance D, and weeds are removed by adopting an anticlockwise route;
wherein f is a collision point discrimination identifier, P is a unique collision point index value, N is a discrete number of pressure sensors, D is a width of a weeding member, and D is a horizontal distance between the orchard weeding machine 1 and the trunk 7.
The pressure sensors uniformly distributed at the front end of the weeding mechanism 11 are equally divided into N sensing blocks from left to right, the returned data of the pressure sensors are expressed as 1 xN-dimensional vectors, and if K is a sensor receiving vector, K isiThe returned value of the ith sensing block is i ═ 1, 2., N; t is a threshold, when the return value of the sensing block is greater than or equal to T, the sensing block is the collision point 73, and the sensor measurement result R is calculated by using the following formula:
Figure BDA0002339398780000121
get RiThe median of all positions at 1 is collision point 73, and the unique collision point index value p is:
Figure BDA0002339398780000122
wherein q isjIs RiSense Block index value of 1, m is all RiCalculating the horizontal distance d between the orchard weeding machine 1 and the trunk 7 according to the position of the collision point 73 as the total number of induction block index values of 1:
Figure BDA0002339398780000123
referring to fig. 5, the working process of the present invention specifically includes: starting up: starting the color depth camera, the controller and all the sensing devices to carry out system initialization; starting a common working mode; and (3) real-time trunk 7 detection: acquiring a color image and a depth image from a camera in real time; detecting the trunk 7 by taking the color image as input, determining the position of the center point of the trunk 7, and determining the relative distance between the center point of the trunk 7 and the orchard weeding machine 1 by combining the corresponding depth image; selecting the trunk 7 closest to the target, and controlling the orchard weeding machine 1 to move towards the trunk 7; judging the position of the orchard weeding machine 1: judging whether the orchard weeding machine 1 enters a region below the tree crown or not according to the detection result of the trunk 7 and the monitoring result of the Hall sensor, and starting a tree crown working mode if the orchard weeding machine 1 enters the region below the tree crown; if not, detecting the trunk 7 again; recording start point: in order to judge whether the orchard weeder 1 finishes cleaning weeds under trees, recording the starting point of entering the crown of a tree, and extracting and recording the characteristics of the image; acquisition of collision point 73: according to the detection result of the pressure sensing belt 6 in front of the weeding part, the position of the collision point 73 between the trunk 7 and the orchard weeding machine 1 is obtained, and the horizontal distance between the trunk 7 and the orchard weeding machine is calculated; and (3) judging a weeding strategy: judging whether the trunk 7 is on the right side of the orchard weeding machine 1 or not according to the position of the collision point 73, and if so, controlling the orchard weeding machine 1 to horizontally avoid the trunk 7 and then cleaning weeds clockwise; if the weed is left or in the middle, the orchard weeding machine 1 is controlled to horizontally avoid the trunk 7 and then weeds are cleaned anticlockwise; loop detection: acquiring a camera color image, extracting features, calculating the similarity with an initial point image, if the similarity is higher than a threshold value, considering that the initial point is returned, cleaning under the tree is finished, and leaving the orchard weeding machine 1 under the tree and starting a common working mode; if not, the tree is not cleared, and the tree crown operation mode is continued. Judging whether to stop the operation, if so, quitting and shutting down; if the operation is not stopped, the trunk 7 is detected again.
The invention establishes a dual-mode weeding mode aiming at the conditions of illumination constraint, weed distribution constraint and the like below a crown. On the premise of not damaging the trunk, the weeds under the tree crown can be completely removed. And returning surrounding obstacle detection information by using a peripheral pressure sensor of the machine body by using color image information and distance information in a view field acquired by using the color depth image. And processing image information obtained by the color camera to detect the trunk and determine the relative position of the trunk and the orchard weeding machine. And filtering and analyzing the detection signal captured by the pressure sensor, and acquiring the position information of the trunk in a physical touch mode. Aiming at the environmental constraints such as weed distribution constraint and illumination and the like below the crown of the fruit tree, a dual-mode weeding mode below the crown of the fruit tree is established, so that the weeds can be removed in an all-round manner and a weeding route can be guided under the condition that a trunk is not damaged.
The present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof, and it should be understood that various changes and modifications can be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (10)

1.一种果园除草机的树冠下方双模式除草方法，其特征在于，果园除草机包括普通工作模式和树冠工作模式，所述果园除草机基于设定条件进行所述普通工作模式和树冠工作模式的切换，所述果园除草机在非树冠下方区域时执行普通工作模式，所述果园除草机在树冠下方区域时开启树冠工作模式，所述树冠下方双模式除草方法包括如下步骤：1. a dual-mode weeding method under the canopy of an orchard weeder, is characterized in that, the orchard weeder comprises a common work pattern and a tree canopy work pattern, and the orchard weeder carries out the common work pattern and the tree canopy work pattern based on a set condition Switching, the orchard weeding machine performs the normal working mode when it is not in the area below the canopy, the orchard weeding machine starts the canopy working mode when the orchard weeding machine is in the area below the canopy, and the dual-mode weeding method under the canopy includes the following steps: S100、所述果园除草机开机进行系统初始化并启动普通工作模式；S100, the orchard weeder is powered on to initialize the system and start the normal working mode; S200、所述普通工作模式包括：实时获取果园的彩色图像和深度图像，利用所述彩色图像进行树干检测并确定所述树干的中心点位置，结合对应的所述深度图像确定所述树干的中心点与所述果园除草机的相对距离，所述果园除草机向距离最近的树干运动；S200. The common working mode includes: acquiring a color image and a depth image of the orchard in real time, using the color image to detect a tree trunk and determining the position of the center point of the tree trunk, and determining the center of the tree trunk in combination with the corresponding depth image The relative distance between the point and the orchard weeder, the orchard weeder moves to the nearest tree trunk; S300、根据树干检测结果和霍尔传感器监测结果，判断所述果园除草机是否进入树冠下方，若是则启动树冠工作模式；若否则继续执行所述普通工作模式；S300, according to the trunk detection result and the Hall sensor monitoring result, determine whether the orchard weeder enters under the tree crown, if so, start the tree crown working mode; otherwise, continue to execute the ordinary working mode; S400、所述树冠工作模式包括：获取所述树干与果园除草机碰撞点的位置，并计算所述树干与果园除草机的水平距离；根据所述碰撞点的位置判断所述树干与所述果园除草机的相对位置，若所述树干在所述果园除草机的右侧，则控制所述果园除草机水平避开所述树干后顺时针清理杂草；若所述树干在所述果园除草机的左侧或正前方，则控制所述果园除草机水平避开所述树干后逆时针清理杂草；以及S400. The canopy working mode includes: acquiring the position of the collision point between the trunk and the orchard weeder, and calculating the horizontal distance between the trunk and the orchard weeder; judging the trunk and the orchard according to the position of the collision point The relative position of the weeder, if the trunk is on the right side of the orchard weeder, control the orchard weeder horizontally to avoid the tree trunk and then clear the weeds clockwise; if the tree trunk is on the orchard weeder to the left or directly in front of the orchard, control the orchard weeder horizontally to avoid the tree trunk and then clear the weeds counterclockwise; and S500、判断树下杂草是否清理完毕，若是，则所述果园除草机离开树下并启动所述普通工作模式；若否，则继续执行所述树冠工作模式。S500, judging whether the weeds under the tree are cleaned up, if so, the orchard weeder leaves the tree and starts the common working mode; if not, continues to execute the canopy working mode. 2.如权利要求1所述的果园除草机的树冠下方双模式除草方法，其特征在于，判断树下杂草是否清理完毕步骤进一步包括：2. The double-mode weeding method under the tree crown of the orchard weeder as claimed in claim 1, wherein the step of judging whether the weeds under the tree are cleaned up further comprises: S501、记录所述果园除草机进入所述树冠下方的起始点，并对所述起始点处的图像进行特征提取并记录；以及S501, record the starting point at which the orchard weeder enters under the canopy, and perform feature extraction and record on the image at the starting point; and S502、获取实时彩色图像并进行特征提取，与所述起始点的图像进行相似度比较，若相似度大于阈值SI则判断所述果园除草机回到起始点，树下清理完毕，并启动所述普通工作模式；若相似度小于或等于所述阈值SI，则判断树下没有清理完毕，继续执行所述树冠工作模式。S502: Acquire a real-time color image and perform feature extraction, compare the similarity with the image at the starting point, and if the similarity is greater than a threshold SI, determine that the orchard weeder has returned to the starting point, and the tree has been cleaned up, and start the Normal working mode; if the similarity is less than or equal to the threshold SI, it is determined that the under-tree has not been cleaned up, and the canopy working mode is continued. 3.如权利要求1或2所述的果园除草机的树冠下方双模式除草方法，其特征在于，还包括如下步骤：3. the double-mode weeding method under the tree crown of the orchard weeding machine as claimed in claim 1 or 2, is characterized in that, also comprises the steps: S600、判断是否停止作业，若是，则所述果园除草机退出所述普通工作模式并关机；若否，则执行步骤S200。S600, judging whether to stop the operation, and if so, the orchard weeder exits the ordinary working mode and shuts down; if not, executes step S200. 4.如权利要求1或2所述的果园除草机的树冠下方双模式除草方法，其特征在于，所述普通工作模式和树冠工作模式的设定条件包括：4. The double-mode weeding method under the canopy of the orchard weeder as claimed in claim 1 or 2, wherein the setting conditions of the common working mode and the canopy working mode include: 所述普通工作模式的设定条件为：能够检测到果树位置且深度距离大于阈值S；压力传感器返回信息为所述果园除草机无碰撞；且霍尔传感器返回信息为所述果园除草机正常行进；The setting conditions of the ordinary working mode are: the position of the fruit tree can be detected and the depth distance is greater than the threshold value S; the return information of the pressure sensor is that the orchard weeder has no collision; and the return information of the hall sensor is that the orchard weeder is traveling normally ; 所述树冠工作模式的设定条件为：无法检测到果树位置且深度距离小于或等于阈值S；所述压力传感器返回信息为所述果园除草机前方发生碰撞；且所述霍尔传感器返回信息为所述果园除草机异常停止；The setting conditions of the canopy working mode are: the position of the fruit tree cannot be detected and the depth distance is less than or equal to the threshold S; the return information of the pressure sensor is that a collision occurs in front of the orchard weeder; and the return information of the Hall sensor is The orchard weeder stops abnormally; 所述果园除草机运动过程中，若同时满足所述普通工作模式或所述树冠工作模式的所有设定条件，所述果园除草机立即切换进入相应模式，否则所述果园除草机保持目前状态直至模式切换。During the movement of the orchard weeding machine, if all the set conditions of the ordinary working mode or the canopy working mode are met at the same time, the orchard weeding machine will immediately switch to the corresponding mode, otherwise the orchard weeding machine will remain in the current state until Mode switch. 5.如权利要求1或2所述的果园除草机的树冠下方双模式除草方法，其特征在于，根据所述碰撞点的位置判断所述树干与所述果园除草机的相对位置进一步包括：5. The dual-mode weeding method under the tree canopy of the orchard weeding machine according to claim 1 or 2, wherein judging the relative position of the tree trunk and the orchard weeding machine according to the position of the collision point further comprises: 判断
Figure FDA0002339398770000021
的符号确定所述碰撞点的位置：judge
Figure FDA0002339398770000021
The sign of determines the location of the collision point: 当f＞0时，所述树干在所述果园除草机右侧，所述果园除草机水平向左运动D-d距离以避开树干，并采用顺时针路线进行杂草清除；When f>0, the tree trunk is on the right side of the orchard weeder, and the orchard weeder moves horizontally to the left for a distance of D-d to avoid the tree trunk, and uses a clockwise route to remove weeds; 当f＜0时，所述树干在所述果园除草机左侧，所述果园除草机水平向右运动D-d距离，并采用逆时针路线进行杂草清除；When f<0, the tree trunk is on the left side of the orchard weeder, and the orchard weeder moves horizontally to the right for a distance of D-d, and uses a counterclockwise route to remove weeds; 当f＝0时，所述树干在所述果园除草机正前方，所述果园除草机水平向右运动D距离，并采用逆时针路线进行杂草清除；When f=0, the trunk is directly in front of the orchard weeder, and the orchard weeder moves horizontally to the right for a distance of D, and uses a counterclockwise route to remove weeds; 其中，f为碰撞点判别标识符，P为唯一碰撞点索引值，N为压力传感器离散数量，D为除草部件的宽度，d为所述果园除草机与树干的水平距离。Among them, f is the identification identifier of the collision point, P is the index value of the unique collision point, N is the discrete quantity of the pressure sensor, D is the width of the weeding part, and d is the horizontal distance between the orchard weeder and the tree trunk. 6.如权利要求5所述的果园除草机的树冠下方双模式除草方法，其特征在于，将均布于所述除草机构前端的多个压力传感器从左至右均分分割为N个感应块，所述压力传感器返回数据表示为1×N维向量，假设K为传感器接收向量，Ki为第i个感应块的返回值，i＝1,2,...,N；T为阈值，当所述感应块的返回值大于等于T时，则所述感应块为所述碰撞点，采用如下公式计算传感器测量结果R：6. The dual-mode weeding method under the canopy of the orchard weeding machine according to claim 5, wherein the plurality of pressure sensors evenly distributed at the front end of the weeding mechanism are equally divided into N sensing blocks from left to right , the return data of the pressure sensor is represented as a 1×N-dimensional vector, assuming K is the sensor receiving vector, K i is the return value of the ith sensing block, i=1, 2,...,N; T is the threshold, When the return value of the sensing block is greater than or equal to T, the sensing block is the collision point, and the following formula is used to calculate the sensor measurement result R:
Figure FDA0002339398770000031
Figure FDA0002339398770000031
 取Ri＝1处所有位置的中位值为碰撞点，唯一碰撞点索引值p为：Taking the median value of all positions at Ri = 1 as the collision point, the index value p of the unique collision point is:
Figure FDA0002339398770000032
Figure FDA0002339398770000032
 其中，qj为Ri＝1的感应块索引值，m为所有Ri＝1的感应块索引值的总数，根据碰撞点p的位置计算所述果园除草机与树干的水平距离d为：Among them, q j is the index value of the induction block with R i =1, m is the total number of index values of all the induction blocks with R i =1, and the horizontal distance d between the orchard weeder and the tree trunk is calculated according to the position of the collision point p:
Figure FDA0002339398770000033
Figure FDA0002339398770000033
 7.如权利要求6所述的果园除草机的树冠下方双模式除草方法，其特征在于，与所述起始点的图像进行相似度比较进一步包括：7. The dual-mode weeding method under the tree canopy of an orchard weeder as claimed in claim 6, wherein comparing the similarity with the image of the starting point further comprises: 将输入图像Iu进行ORB特征提取后，表示为n维描述符集合Iu→{d1...dn}，di为ORB描述符特征点，与视觉字典中的视觉单词
Figure FDA0002339398770000034
对应；通过将输入图像与进入树冠时的起始点图像Iv进行相似度计算，图像Iu与Iv的相似度S(Iu,Iv)可通过下式计算：After the ORB feature extraction is performed on the input image I u , it is expressed as an n-dimensional descriptor set I u →{d 1 ...d n }, where d i is the ORB descriptor feature point, which is related to the visual word in the visual dictionary.
Figure FDA0002339398770000034
Corresponding; by calculating the similarity between the input image and the starting point image I v when entering the canopy, the similarity S (I u , I v ) of the image I u and I v can be calculated by the following formula:
Figure FDA0002339398770000035
Figure FDA0002339398770000035
 其中，Ui为图像Iu的矢量表示元素，vi为图像Iv的矢量表示元素，Nf为视觉字典中所有特征数量。Among them, U i is the vector representation element of the image I u , v i is the vector representation element of the image I v , and N f is the number of all features in the visual dictionary. 8.一种果园除草机的树冠下方双模式除草装置，其特征在于，包括：8. A dual-mode weeding device under the canopy of an orchard weeder, characterized in that, comprising: 控制器，安装在果园除草机上，并与所述果园除草机的驱动机构连接；a controller, installed on the orchard weeder, and connected with the drive mechanism of the orchard weeder; 图像采集装置，用于采集果园图像进行树干检测并确定所述树干的中心点位置，所述图像采集装置安装在所述果园除草机的正前方并与所述控制器连接，所述图像采集装置的摄像头水平朝向正前方；an image acquisition device, used for collecting images of the orchard for tree trunk detection and determining the position of the center point of the tree trunk, the image acquisition device is installed in front of the orchard weeder and connected to the controller, the image acquisition device The camera faces straight ahead horizontally; 霍尔元件对，与所述控制器连接，所述霍尔元件对包括第一霍尔传感器和第二霍尔传感器，所述第一霍尔传感器安装在所述果园除草机的尾部，所述第二霍尔传感器安装在所述果园除草机的后轴承上并与所述第一霍尔传感器对应设置；以及a pair of hall elements, connected with the controller, the pair of hall elements includes a first hall sensor and a second hall sensor, the first hall sensor is installed at the tail of the orchard weeder, the A second Hall sensor is mounted on the rear bearing of the orchard weeder and is arranged corresponding to the first Hall sensor; and 压力传感带，用于实时检测并返回所述果园除草机与所述树干的碰撞部位的压力值并转化为电压值，所述压力传感带包括多个压力传感器，所述多个压力传感器均匀设置在所述果园除草机的除草机构前端，并分别与所述控制器连接。A pressure sensing strip, used for real-time detection and return of the pressure value of the collision site between the orchard weeder and the tree trunk, and converting it into a voltage value, the pressure sensing strip includes a plurality of pressure sensors, and the plurality of pressure sensors They are evenly arranged at the front end of the weeding mechanism of the orchard weeding machine, and are respectively connected with the controller. 9.如权利要求8所述的果园除草机的树冠下方双模式除草装置，其特征在于，所述图像采集装置为彩色摄像头，所述彩色摄像头通过USB接口与所述控制器连接。9 . The dual-mode weeding device under the canopy of an orchard weeder according to claim 8 , wherein the image acquisition device is a color camera, and the color camera is connected to the controller through a USB interface. 10 . 10.一种果园除草机，其特征在于，包括上述权利要求8或9所述的树冠下方双模式除草装置，且采用上述权利要求1-7中任意一项所述的树冠下方双模式除草方法进行杂草清除。10. An orchard weeding machine, characterized in that it comprises the dual-mode weeding device under the canopy according to claim 8 or 9, and adopts the dual-mode weeding method under the canopy according to any one of the above claims 1-7 Carry out weed removal.
CN201911369918.2A 2019-12-26 2019-12-26 Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy Active CN111096107B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911369918.2A CN111096107B (en) 2019-12-26 2019-12-26 Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911369918.2A CN111096107B (en) 2019-12-26 2019-12-26 Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy

Publications (2)

Publication Number Publication Date
CN111096107A true CN111096107A (en) 2020-05-05
CN111096107B CN111096107B (en) 2021-03-19

Family

ID=70424483

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911369918.2A Active CN111096107B (en) 2019-12-26 2019-12-26 Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy

Country Status (1)

Country Link
CN (1) CN111096107B (en)

Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2662045A1 (en) * 1990-05-15 1991-11-22 Furlani Pierino Machine for eradicating weeds, particularly between fruit trees
CN101707992A (en) * 2009-10-15 2010-05-19 南京林业大学 high-efficiency weeding robot
CN102282922A (en) * 2011-06-09 2011-12-21 中国农业大学 Inter-seedling weeding device and method
AU2012233051A1 (en) * 2011-10-04 2013-04-18 Lance Dear Mowing Apparatus
CN103329653A (en) * 2013-07-06 2013-10-02 西北农林科技大学 Intelligent all-dimensional mechanical weeding machine of orchards
CN104521936A (en) * 2015-01-15 2015-04-22 无锡北斗星通信息科技有限公司 Automatic weed cleaning system
WO2015064780A1 (en) * 2013-10-30 2015-05-07 주식회사 드림씨엔지 Intelligent unmanned robotic weeder
CN106772631A (en) * 2016-12-01 2017-05-31 成都市宏德永兴养殖有限公司 Fruit tree for unmanned vehicle operation in orchard recognizes preventing collision method
CN107295818A (en) * 2017-07-20 2017-10-27 华南农业大学 Paddy field weed-killer machine automatic seedling avoiding system and method
CN207443361U (en) * 2017-10-31 2018-06-05 江西伊禾农产品科技发展有限公司 A kind of novel traction type woods orchard floating weeder
CN208113359U (en) * 2018-04-14 2018-11-20 宋国权 A kind of automatic weeder of agricultural
CN208387294U (en) * 2018-04-18 2019-01-18 石河子大学 Weeder between a kind of strain of orchard
CN208490686U (en) * 2018-06-04 2019-02-15 南京农业大学 A kind of improved orchard weeder
CN208908534U (en) * 2018-05-18 2019-05-31 宁县海升现代农业有限公司 A kind of weeder allows tree device automatically
CN110268817A (en) * 2019-08-01 2019-09-24 吉林大学 A kind of orchard ring-weeding is mechanical
CN110298914A (en) * 2019-05-29 2019-10-01 江苏大学 A kind of method of fruit tree canopy characteristic map in orchard establishing
CN110447374A (en) * 2019-09-11 2019-11-15 宁夏农林科学院枸杞工程技术研究所 Grass trimmer between a kind of fructus lycii special electronic induction plant

Patent Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2662045A1 (en) * 1990-05-15 1991-11-22 Furlani Pierino Machine for eradicating weeds, particularly between fruit trees
CN101707992A (en) * 2009-10-15 2010-05-19 南京林业大学 high-efficiency weeding robot
CN102282922A (en) * 2011-06-09 2011-12-21 中国农业大学 Inter-seedling weeding device and method
AU2012233051A1 (en) * 2011-10-04 2013-04-18 Lance Dear Mowing Apparatus
CN103329653A (en) * 2013-07-06 2013-10-02 西北农林科技大学 Intelligent all-dimensional mechanical weeding machine of orchards
WO2015064780A1 (en) * 2013-10-30 2015-05-07 주식회사 드림씨엔지 Intelligent unmanned robotic weeder
CN104521936A (en) * 2015-01-15 2015-04-22 无锡北斗星通信息科技有限公司 Automatic weed cleaning system
CN106772631A (en) * 2016-12-01 2017-05-31 成都市宏德永兴养殖有限公司 Fruit tree for unmanned vehicle operation in orchard recognizes preventing collision method
CN107295818A (en) * 2017-07-20 2017-10-27 华南农业大学 Paddy field weed-killer machine automatic seedling avoiding system and method
CN207443361U (en) * 2017-10-31 2018-06-05 江西伊禾农产品科技发展有限公司 A kind of novel traction type woods orchard floating weeder
CN208113359U (en) * 2018-04-14 2018-11-20 宋国权 A kind of automatic weeder of agricultural
CN208387294U (en) * 2018-04-18 2019-01-18 石河子大学 Weeder between a kind of strain of orchard
CN208908534U (en) * 2018-05-18 2019-05-31 宁县海升现代农业有限公司 A kind of weeder allows tree device automatically
CN208490686U (en) * 2018-06-04 2019-02-15 南京农业大学 A kind of improved orchard weeder
CN110298914A (en) * 2019-05-29 2019-10-01 江苏大学 A kind of method of fruit tree canopy characteristic map in orchard establishing
CN110268817A (en) * 2019-08-01 2019-09-24 吉林大学 A kind of orchard ring-weeding is mechanical
CN110447374A (en) * 2019-09-11 2019-11-15 宁夏农林科学院枸杞工程技术研究所 Grass trimmer between a kind of fructus lycii special electronic induction plant

Also Published As

Publication number Publication date
CN111096107B (en) 2021-03-19

Similar Documents

Publication Publication Date Title
EP3836783B1 (en) Method for autonomous detection of crop location based on tool depth and location
Shalal et al. A review of autonomous navigation systems in agricultural environments
CN112056087B (en) Induction system of small-sized segmented crawler sugarcane harvester and control method
US11553636B1 (en) Spacing-aware plant detection model for agricultural task control
CN103891697A (en) Drug spraying robot capable of moving indoors autonomously and variable drug spraying method thereof
Visentin et al. A mixed-autonomous robotic platform for intra-row and inter-row weed removal for precision agriculture
US10959377B2 (en) Crop scanner
CN113906900B (en) Sugarcane harvester and method for adjusting position and posture of cutter head of sugarcane harvester based on multi-sensor fusion
CN108064560A (en) The automatic picker system of fruit and method based on Kinect depth of field cameras
Ahmadi et al. Towards autonomous visual navigation in arable fields
US20240300499A1 (en) Vehicle row follow system
English et al. Learning crop models for vision-based guidance of agricultural robots
CN115560754A (en) Visual navigation method based on weed removal
AU2021466764A1 (en) Vehicle row follow system
CN111096107B (en) Orchard weeding machine and dual-mode weeding method and weeding device under tree canopy
Mahmud et al. Measuring tree canopy density using a lidar-guided system for precision spraying
US20240324485A1 (en) Automated systems and methods for agricultural crop monitoring and sampling
US20230403964A1 (en) Method for Estimating a Course of Plant Rows
Hu et al. Extraction method for centerlines of crop row based on improved lightweight yolov4
EP3959955A1 (en) Agricultural machine such as harvester
Raja et al. A novel weed and crop recognition technique for robotic weed control in a lettuce field with high weed densities
CN115410193A (en) Machine vision-based corn sidewalk identification method
JP2024078006A (en) Harvesting vehicle
Patnaik et al. Weed removal in cultivated field by autonomous robot using LabVIEW
CN116645614B (en) A precision agricultural robot operation method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant