CN110426045A - A kind of farmland spray machine device people vision guided navigation parameter acquiring method - Google Patents
A kind of farmland spray machine device people vision guided navigation parameter acquiring method Download PDFInfo
- Publication number
- CN110426045A CN110426045A CN201910740859.9A CN201910740859A CN110426045A CN 110426045 A CN110426045 A CN 110426045A CN 201910740859 A CN201910740859 A CN 201910740859A CN 110426045 A CN110426045 A CN 110426045A
- Authority
- CN
- China
- Prior art keywords
- guidance path
- parameter
- navigation
- navigational parameter
- machine device
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- 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)
- Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Automation & Control Theory (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The present invention discloses a kind of farmland spray machine device people vision guided navigation parameter acquiring method, introduce feedback and previewing mechanism, feedback navigational parameter, Present navigation parameter and previewing navigational parameter are combined, the navigational parameter acquisition methods based on serial closed loop BP network designed using the present invention, obtain high-precision navigational parameter.The location navigation image of acquisition is handled first, identifies that guidance path realizes self poisoning;Then navigational parameter is extracted according to feedback navigation information, Present navigation information and previewing navigation information;Finally using the navigational parameter of extraction as in combination component incoming serial closed loop BP network, high-precision navigational parameter is obtained.Realize the acquisition of farmland spray machine device people high-precision and high reliability navigational parameter.
Description
Technical field
The invention belongs to intelligent mobile robot vision navigation system field, more particularly, to a kind of for farmland operation
The vision guided navigation parameter acquiring method of spray robot.
Background technique
With mechanization of agriculture, informationization and intelligent development, more and more intelligence Agricultural Machinery Equipment tools come into being,
Its investment that can greatly reduce manual labor reduces cost, improves agricultural operation efficiency.Intelligent mobile spray robot tool
There is the features such as efficiently precisely spraying insecticide, save the activity duration and reduce pesticide waste pollution.High-precision vision guided navigation ginseng
Number is the important prerequisite condition of farmland spray machine device people's normal operation, under the farmland operation environment of winding complex, to improve spray
The accuracy of medicine robot navigation, it is necessary to sufficiently obtain and utilize known navigation information.However, current existing robot vision
Navigational parameter, which obtains, uses the navigation mechanism based on straight line path model, this can seriously affect the reliability of spray robot navigation
And precision.The present invention in this context, discloses a kind of farmland spray machine device people vision guided navigation parameter acquiring method, this method knot
It closes feedback and pre- view principle and obtains high-precision navigational parameter using serial closed loop BP network, there is high precision and high stable
The features such as property.
Summary of the invention
The purpose of the present invention is to provide a kind of farmland spray machine device people vision guided navigation parameter acquiring methods.This method will be anti-
Feedback and previewing mechanism are introduced into vision guided navigation parameter extraction, while being designed and being obtained accurately robot using serial closed loop BP network
Vision guided navigation parameter.
To achieve the above object, technical solution of the present invention is as follows: a kind of farmland spray machine device people vision guided navigation parameter acquisition
Method, which is characterized in that acquisition image first carries out image procossing, identifies general path information and determines spray robot
Relative position;Secondly, extracting Present navigation parameter, feedback navigational parameter and previewing navigational parameter;Finally, obtaining final navigation ginseng
Number.
Extraction Present navigation parameter, feedback navigational parameter and the previewing navigational parameter is using calculating guidance path
The mobile method of the method and ROI window recurrence of state curvature is realized;
The final navigational parameter of acquisition is realized using the method for obtaining navigational parameter based on serial closed loop BP network.
Further, what the mobile method of the ROI window recurrence obtained is that robot coordinate origin is cut to guidance path
The deflection angle of vertical range (lateral deviation) and robot of line;
Further, what the method for the state curvature of the calculating guidance path obtained is guidance path status information;
Further, it is described based on serial closed loop BP network obtain navigational parameter method include the network guidance path state BP and
The network navigational parameter BP and the control strategy for obtaining navigational parameter.
Detailed description of the invention
Fig. 1 is a kind of general frame signal of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention
Figure;
Fig. 2 is a kind of ROI window schematic diagram of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention;
Fig. 3 is a kind of calculating side of the state curvature of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention
Formula (1) schematic diagram;
Fig. 4 is a kind of calculating side of the state curvature of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention
Formula (2) schematic diagram;
Fig. 5 is that a kind of current and pre- viewed area of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention is led
Navigate parameter schematic diagram;
Fig. 6 is a kind of guidance path state BP net of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention
Road and navigational parameter BP network diagram;
Fig. 7 is a kind of closed loop control of the serial network of farmland spray machine device people vision guided navigation parameter acquiring method of the embodiment of the present invention
Make tactful schematic diagram.
Specific embodiment
Present embodiment is illustrated in conjunction with Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7.A kind of farmland spray machine
Device people's vision guided navigation parameter acquiring method includes the following steps:
(1) image procossing is carried out after obtaining visual pattern, to obtain the relative position of spray robot.
(2) ROI window recurrence is mobile.For dynamically track guidance path, region of interest window ROI is arranged in the present invention
(Region Of Interest).Only the path sections in R0I window are handled when obtaining navigational parameter, by the machine that sprays
The entire moving process of people is divided into multiple periods.
As shown in Figure 2.Y-axis direction is current spray robot direction of travel, and rectangle is current ROI window, if O1O2For
The destination path track that vision system extracts, wherein O1、O2It is 2 intersection points of ROI window and guidance path respectively.
The target point of spray robot is O2It is found down again if gone to after target point from current location by former direction of travel
One step diameter, then its run trace will have relatively large deviation with practical guidance path, so as to cause navigation failure.If spraying machine
Device people wants to return on correct guidance path, then must be to target point O2Direction is advanced, i.e. the deflection angle of spray robot is
OO2With the angle of y-axis, this makes it possible to obtain navigational parameter of the robot relative to guidance path that spray, whereinFor machine
People's coordinate origin to guidance path tangent line vertical range (lateral deviation),For the deflection angle of robot.
(3) the state curvature of guidance path is calculated.It calculates as shown in Figure 3, Figure 4.Determine current opposite of spray robot
Behind position, the state curvature of guidance path in current ROI window is calculated, to determine the status information of guidance path.
In the ROI window of current navigation path,Axis is the current driving direction of robot that sprays, edgeAxis direction will
ROI window is divided into n sub-regions, then the guidance path in each subregion can be approximately straight line.
If i+1 area navigation path is relative to the angle of ith zone guidance path(1≤i≤n), then lead to
Cross the state curvature that following formula calculates guidance path in ROI window:
WhereinIt is currentROIPath length in window.
Using identical method, any a part in current ROI Window Navigation path is calculatedState curvature,
Calculation such as following formula:
Wherein,Indicate the angle of the guidance path in m-th of region relative to the guidance path in n-th of region;It indicates
In current ROI window fromIt arrivesPath length.
(4) as shown in Figure 5.Current ROI window is divided into previewing navigation area and Present navigation region or more the area Liang Ge
Domain.Present navigation information in the feedback navigation information and current ROI window in a upper period, previewing navigation information are combined,
It is achieved with the spray accurate navigational parameter of robot.The horizontal axis of pre- viewed area is set as X1, in the level of current region
Axis is X2, guidance path and axis X1, ROI window axis, axis X2Intersection point be respectively 3 points of M, P and N, each target point
It can be indicated with the navigational parameter of spray robot, i.e.,、With.The Robot that sprays is each
Target point successively moves, so that it may realize the precision navigation of itself.、It navigates as two tunnel previewing
Parameter,For Present navigation parameter all the way, in conjunction with the feedback navigational parameter in a upper period, by this 4
Divide navigational parameter group component amount input step (5) that navigational parameter can be obtained.
(5) navigational parameter is obtained based on serial closed loop BP network.Two errors that the present invention devises serial structure are reversed
Propagation Neural Network, the i.e. network guidance path state BP and the network navigational parameter BP.As shown in fig. 6, both using 3 layers of feedforward
Structure.Guidance path is divided into 10 equal portions along the y axis, then can obtain 10 state characteristic points.Since ordinate presses ROI
Window equal part, therefore by 10 state Based on Feature Points are as follows: ……, ,
InWithThe lateral deviation and state curvature of respectively 10 state characteristic points.It will
This 10 state characteristic points are as input vectorComponent,
Middle P is the sample number of BP e-learning, and output vector then represents the state of guidance path.Therefore, the network guidance path state BP
It include the input layer of 10 nodes including one, a hidden layer comprising 12 nodes and one include the output of 2 nodes
Layer.Navigational parameter BP network includes the input layer comprising 4 nodes, a hidden layer and a packet comprising 8 nodes
Output layer containing 2 nodes.Navigational parameter will be fed back all the way, Present navigation parameter all the way, two tunnels it is pre-
Depending on navigational parameterWithGroup component amount input network can obtain navigational parameter.
Wherein, guidance path state BP network and navigational parameter BP network use serial Closed-loop Control Strategy, obtain final
Navigational parameter.As shown in fig. 7, control strategy is as follows
1) judge whether guidance path state is normal according to guidance path state BP neural network.
If 2) guidance path state is normal, navigational parameter is obtained by navigational parameter BP network, realizes spray machine
The precision navigation of people.
If 3) guidance path state is abnormal, makes the robot stopping walking that sprays, find new guidance path.
4) new guidance path state is rejudged after the new guidance path of discovery using Closed-loop Control Strategy.Such as
The new guidance path state of fruit is still abnormal, this process is jumped out in termination, otherwise carries out the acquisition of the navigational parameter of next step.
The above embodiments are only used to illustrate the present invention, and not limitation of the present invention, in relation to the common of technical field
Technical staff can also make a variety of changes and modification without departing from the spirit and scope of the present invention, therefore all
Equivalent technical solution also belongs to scope of the invention, and scope of patent protection of the invention should be defined by the claims.
Claims (6)
1. a kind of farmland spray machine device people vision guided navigation parameter acquiring method, which is characterized in that pass through ROI window recurrence movement side
Method dynamically track guidance path obtains robot relative position;Navigation is extracted by the state curvature estimation method of guidance path
Path state information;Final navigational parameter is obtained by obtaining the method for navigational parameter based on serial closed loop BP network.
2. a kind of farmland spray machine device people vision guided navigation parameter acquiring method according to claim 1, which is characterized in that institute
What the ROI window recurrence moving method stated obtained be robot coordinate origin to guidance path tangent line vertical range it is (laterally inclined
Difference) and robot deflection angle.
3. a kind of farmland spray machine device people vision guided navigation parameter acquiring method according to claim 1, which is characterized in that institute
The ROI window recurrence moving method stated sets y-axis direction as current spray robot direction of travel, and rectangle is current ROI window
Mouthful, ifFor vision system extract destination path track, wherein、It is 2 of ROI window and guidance path respectively
Intersection point.
4. a kind of farmland spray machine device people vision guided navigation parameter acquiring method according to claim 1, which is characterized in that institute
The state curvature estimation method for the guidance path stated is set in the ROI window of current navigation path, and y-axis is that spray robot works as
ROI window is divided into n sub-regions along the y-axis direction by preceding driving direction, and can be close by the guidance path in each subregion
It is seemingly straight line;If i+1 area navigation path is relative to the angle of ith zone guidance path, calculate the state curvature estimation method of guidance path in ROI window are as follows:
Calculate any a part in current ROI Window Navigation pathState curvature, calculation such as following formula:
Wherein,Indicate the angle of the guidance path in m-th of region relative to the guidance path in n-th of region;Expression is worked as
In preceding ROI window fromIt arrivesPath length.
5. a kind of farmland spray machine device people vision guided navigation parameter acquiring method according to claim 1, which is characterized in that institute
The method that navigational parameter is obtained based on serial closed loop BP network stated, including guidance path state BP network, navigational parameter BP net
Network and serial Closed-loop Control Strategy.
6. a kind of farmland spray machine device people vision guided navigation parameter acquiring method according to claim 5, which is characterized in that institute
The serial Closed-loop Control Strategy stated, as follows:
1) judge whether guidance path state is normal according to guidance path state BP neural network;
If 2) guidance path state is normal, navigational parameter is obtained by navigational parameter BP network, realizes spray robot
Precision navigation;
If 3) guidance path state is abnormal, makes the robot stopping walking that sprays, find new guidance path;
4) new guidance path state is rejudged, if newly after the new guidance path of discovery using Closed-loop Control Strategy
Guidance path state it is still abnormal, this process is jumped out into termination, otherwise carry out next step navigational parameter acquisition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910740859.9A CN110426045A (en) | 2019-08-12 | 2019-08-12 | A kind of farmland spray machine device people vision guided navigation parameter acquiring method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910740859.9A CN110426045A (en) | 2019-08-12 | 2019-08-12 | A kind of farmland spray machine device people vision guided navigation parameter acquiring method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110426045A true CN110426045A (en) | 2019-11-08 |
Family
ID=68415613
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910740859.9A Pending CN110426045A (en) | 2019-08-12 | 2019-08-12 | A kind of farmland spray machine device people vision guided navigation parameter acquiring method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110426045A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110825083A (en) * | 2019-11-12 | 2020-02-21 | 深圳创维数字技术有限公司 | Control method, apparatus, and computer-readable storage medium for vehicle |
CN111413986A (en) * | 2020-04-08 | 2020-07-14 | 安徽舒州农业科技有限责任公司 | Automatic driving control method and system based on agricultural machinery |
CN113093743A (en) * | 2021-03-30 | 2021-07-09 | 西北农林科技大学 | Navigation control method based on virtual radar model and deep neural network |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105182976A (en) * | 2015-09-17 | 2015-12-23 | 西北农林科技大学 | Visual navigation strategy of agricultural robot |
CN109885063A (en) * | 2019-03-15 | 2019-06-14 | 江苏大学 | A kind of application robot farmland paths planning method merging vision and laser sensor |
-
2019
- 2019-08-12 CN CN201910740859.9A patent/CN110426045A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105182976A (en) * | 2015-09-17 | 2015-12-23 | 西北农林科技大学 | Visual navigation strategy of agricultural robot |
CN109885063A (en) * | 2019-03-15 | 2019-06-14 | 江苏大学 | A kind of application robot farmland paths planning method merging vision and laser sensor |
Non-Patent Citations (1)
Title |
---|
TANG JINGLEI ET AL.: "Visual navigation control for agricultural robot using serial BP neural network", 《农业工程学报》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110825083A (en) * | 2019-11-12 | 2020-02-21 | 深圳创维数字技术有限公司 | Control method, apparatus, and computer-readable storage medium for vehicle |
CN110825083B (en) * | 2019-11-12 | 2023-02-28 | 深圳创维数字技术有限公司 | Control method, apparatus, and computer-readable storage medium for vehicle |
CN111413986A (en) * | 2020-04-08 | 2020-07-14 | 安徽舒州农业科技有限责任公司 | Automatic driving control method and system based on agricultural machinery |
CN113093743A (en) * | 2021-03-30 | 2021-07-09 | 西北农林科技大学 | Navigation control method based on virtual radar model and deep neural network |
CN113093743B (en) * | 2021-03-30 | 2022-08-30 | 西北农林科技大学 | Navigation control method based on virtual radar model and deep neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110426045A (en) | A kind of farmland spray machine device people vision guided navigation parameter acquiring method | |
CN107203214B (en) | A kind of cooperative self-adapted Intelligent planning method in carrying robot COMPLEX MIXED path | |
CN107544548B (en) | Method and device for controlling operation of unmanned aerial vehicle and unmanned aerial vehicle | |
CN105261034B (en) | The statistical method and device of vehicle flowrate on a kind of highway | |
CN105867379B (en) | A kind of motion control method and control system of robot | |
WO2021114764A1 (en) | Map correction method and system based on local map | |
CN103679774B (en) | A kind of polygon farmland operation zone boundary modeling method | |
CN109508007A (en) | A kind of agricultural machinery track following, obstacle avoidance system and method based on Multi-source Information Fusion | |
CN110502032A (en) | A kind of unmanned plane cluster formation flight method of Behavior-based control control | |
CN104575075B (en) | A kind of city road network vehicle coordinate bearing calibration based on the Big Dipper and device | |
CN107992078A (en) | A kind of autonomous paths planning method of plant protection unmanned plane | |
CN106168805A (en) | The method of robot autonomous walking based on cloud computing | |
CN107272680A (en) | A kind of automatic follower method of robot based on ROS robot operating systems | |
CN103714563A (en) | Modeling method of boundary of curve type farmland operation area | |
CN104808590A (en) | Mobile robot visual servo control method based on key frame strategy | |
CN103079059A (en) | Single-gun-camera-multi-dome-camera linkage method based on grid combination interpolation | |
CN107515004A (en) | Step size computation device and method | |
CN104197938A (en) | Indoor localization method based on foot inertia sensor | |
CN105786000A (en) | Positioning method and positioning system for robot in planned path | |
CN103941233A (en) | Radar intermittence alternate radiation control method based on multi-platform active and passive sensor collaborative tracking | |
CN104615150A (en) | Machine vision based adaptive precise mist spray device and method | |
CN106444757A (en) | EKF-SLAM (Extended Kalman Filter-Simultaneous Localization And Mapping) algorithm based on straight line feature map | |
CN101126806A (en) | Method for revising maximum likelihood registration based information infusion | |
CN108804539A (en) | A kind of track method for detecting abnormality under time and space double-visual angle | |
CN104864873A (en) | Method for helping map positioning by utilizing human motion features |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191108 |