CN108133471A - Agriculture Mobile Robot guidance path extracting method and device based on artificial bee colony algorithm under the conditions of a kind of natural lighting - Google Patents
Agriculture Mobile Robot guidance path extracting method and device based on artificial bee colony algorithm under the conditions of a kind of natural lighting Download PDFInfo
- Publication number
- CN108133471A CN108133471A CN201611076483.9A CN201611076483A CN108133471A CN 108133471 A CN108133471 A CN 108133471A CN 201611076483 A CN201611076483 A CN 201611076483A CN 108133471 A CN108133471 A CN 108133471A
- Authority
- CN
- China
- Prior art keywords
- crop row
- crop
- image
- center line
- candidate
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the Agriculture Mobile Robot guidance path extracting methods based on artificial bee colony algorithm under the conditions of a kind of natural lighting and device, this method to include:Field-crop Image Acquisition;The processing of crop image gray processing;It will treated that gray level image is split is converted into bianry image;Crop row regional extent is obtained to carrying out gray scale vertical projection at the top and bottom of bianry image respectively;The crop row characteristic point in crop row regional extent is detected using based on the vertical projection method of moving window;The characteristics of according to crop row in the picture, establishes crop row center line solving model, carries out optimizing search to the crop row characteristic point in crop row regional extent by artificial bee colony algorithm and determines crop row center line;The guidance path between two crop rows is determined according to nearest adjacent two crop row center lines of range image center line;The method increase the accuracy of detection of guidance path, speed and the adaptability to natural lighting, solve the problems such as accuracy existing for traditional agriculture Mobile Robotics Navigation path extraction method is low, real-time is poor and changes sensitivity to natural lighting.
Description
Technical field
The present invention relates to Agriculture Mobile Robot field of navigation technology, and in particular to people is based under the conditions of a kind of natural lighting
The Agriculture Mobile Robot guidance path extracting method and device of work ant colony algorithm.
Background technology
Field self-navigation operation is completed using Agriculture Mobile Robot, farmland operation efficiency can be effectively improved, is reduced
Production cost avoids labourer from being exposed in the adverse circumstances such as high temperature, high humidity, harmful.At present, Agriculture Mobile Robot is led
The research of boat is concentrated mainly in machine vision and satellite positioning (GNSS) technology two ways, the moving machine based on machine vision
Device people has operating efficiency high, and the advantages such as real-time is good, system cost is low have become domestic and international precision agriculture research field
One hot spot.Guidance path identification is the premise of Agriculture Mobile Robot independent navigation operation, rapidly and accurately extracts crop row
Center line and guidance path can effectively improve the working efficiency and homework precision of mobile robot.For early period crops, due to
Less parallel between Mechanization sowing crop row, while crop row growth has continuity, is presented in whole tendency approximate straight
Line can extract crop row center line by carrying out processing to field-crop image and obtain guidance path;
Farm environment is complicated and changeable, has the characteristics that unstructured and open, influences Agriculture Mobile Robot vision guided navigation road
The principal element of diameter identification includes:Natural lighting variation and the Path Recognition algorithm performance of itself.For natural lighting problem, have
Scholar proposes a kind of navigation information acquisition method based on light durability and the unrelated figure of illumination, to reduce illumination variation to leading
The influence of bit path extraction.Some scholars under RGB, HIS and YCrCb color space, select 2G-R-B color factors, H respectively
Component and Cr components carry out gray processing processing to image, to improve adaptability of the image segmentation to illumination variation.But, R,
G, tri- components of B intercouple, with intensity of illumination change and change, therefore 2G-R-B color factors to illumination variation adaptability not
It is high;H components are converted into non-linear with RGB color, be easy to cause image fault, and error is generated to processing result image;
YCrCb color spaces lack the description to green component, are not suitable for processing farmland green crop image.In crop row center line
(Straight line)Approximating method selection on, some scholars using Hough transform carry out crop row identification, Hough transform strong robustness, but
It is limited to calculate complexity, accuracy of detection.Some scholars extract crop row straight line using least square method, have faster detection speed
Degree, but least square method is to very noisy point sensitivity, poor robustness.At present, scholar proposes the crop based on random device
Row Straight Line Extraction, which has the characteristics that computation complexity is low, real-time is good, but straight-line detection precision is largely
The upper selection dependent on random point.In addition, also scholar extracts crop skeleton using largest square, obtained by fitting a straight line
Crop row straight line since algorithm needs to search largest square to each target pixel points, causes computation complexity to increase, takes
It is more, it is difficult to meet requirement of the navigation operating system to real-time at a high speed.
Invention content
(One)Technical problems to be solved
The technical problem to be solved by the present invention is to how to carry out guidance path extraction under the conditions of natural lighting, it is mobile to improve agricultural
Accuracy, real-time and the adaptability of robot navigation's Path Recognition.
(Two)Technical solution
In order to solve the above-mentioned technical problem, in a first aspect, the present invention provides artificial bee colony is based under the conditions of a kind of natural lighting
The Agriculture Mobile Robot guidance path extracting method of algorithm, the described method comprises the following steps:
S01, crop Image Acquisition;Make video camera and horizontal direction in 60 °~70 ° angles, apart from ground vertical height about 1.2 ~
1.4m;
S03, image segmentation is carried out using maximum variance between clusters, converts gray images into bianry image;
S04, crop row is obtained in the edge position information on image top margin and image base using vertical projection method, is connected by straight line
Edge fit edge point forms crop row regional extent;
S05, crop row characteristic point in the crop row regional extent is extracted using the vertical projection method based on moving window;
S06, according to crop row in the picture the characteristics of, establish crop row center line solving model, existed by artificial bee colony algorithm
Extraction crop row center line in the crop row regional extent;
S07, it determines to be located at two crop row center lines according to two nearest crop row center lines of range image center line
Between guidance path.
The preferred step S02 specifically includes following steps:
(1)
(2)
(3)Cg processing is obtained according to ITU-R BT.601-6 standards(3)Formula:
(3)
(4)
(5)
(6)
The preferred step S06 specifically includes following steps:
(1)The characteristics of according to crop row in the picture, establishes crop row center line solving model, and so-called crop row center line is asked
It is exactly that field-crop row morphologically shows as near linear to solve model, and equation can be according to two crop row features in image
Point determines.If the crop row characteristic point data space that V expressions are obtained according to the step S05,WithFor two in V
A, then crop row center line equation can be expressed as:
(7)
Feature point number in the range of statistical distance straight line d, as evaluation straight line quality standard, by adjustingWithPosition is selected comprising the most straight line of characteristic point as crop row center line, wherein, the value range of d is
(1,5);
(2)The crop row characteristic point in crop row regional extent that the step S05 is obtained is divided into using 1/2 height of image as boundary
Upper and lower 2 part, using the crop row characteristic point of top half as the candidate start point of crop row center line, lower half portion crop row spy
Sign point establishes array and stores crop row center line candidate start point and candidate end point respectively as crop row center line candidate end point;
(3)According to crop row center line solving model, randomly choose 1 candidate start point and 1 candidate end point forms artificial bee colony
One nectar source of algorithm represents a candidate crop row center line.It initializes multiple nectar sources and forms a plurality of candidate crop row straight line,
The a certain range of feature point number of statistical distance candidate's straight line, as the fitness function of the candidate straight line quality of evaluation,
The candidate straight line for choosing fitness function maximum by the multiple search of artificial bee colony algorithm is used as crop row center line.
The present invention also provides the Agriculture Mobile Robot navigation based on artificial bee colony algorithm under the conditions of a kind of natural lighting
Path extraction device, including:
(2)Image segmentation module converts gray images into bianry image by maximum variance between clusters;
(3)Crop row regional extent determining module, for according to bianry image, crop row region model to be determined by vertical projection method
It encloses;
(4)Crop row characteristic point detection module, for according to crop row regional extent, passing through the upright projection based on moving window
Method extracts the crop row characteristic point in the crop row regional extent;
(5)Crop row center line extraction module the characteristics of according to crop row in the picture, establishes crop row center line and solves mould
Type extracts crop row center line by artificial bee colony algorithm in the crop row regional extent;
(6)Guidance path determining module, for determining to be located at according to two nearest crop row center lines of range image center line
Guidance path between two crop row center lines.
Preferably, described image gray processing processing module, including:
(8)
(9)
(3)Cg processing is obtained according to ITU-R BT.601-6 standards(10)Formula:
(10)
(11)
(12)
(13)
The preferred crop row center line extraction module includes:
(1)The characteristics of according to crop row in the picture, establishes crop row center line solving model, and so-called crop row center line is asked
It is exactly that field-crop row morphologically shows as near linear to solve model, and equation can be according to two crop row features in image
Point determines.If the crop row characteristic point data space that V expressions are obtained according to the crop row characteristic point detection module,WithFor two points in V, then crop row center line equation can be expressed as:
(14)
Feature point number in the range of statistical distance straight line d, as evaluation straight line quality standard, by adjustingWithPosition is selected comprising the most straight line of characteristic point as crop row center line, wherein, the value range of d is
(1,5);
(2)The crop row characteristic point in crop row regional extent that the crop row characteristic point detection module is obtained is with image 1/
2 height are divided into 2 parts up and down for boundary, using the crop row characteristic point of top half as the candidate start point of crop row center line, under
Half part crop row characteristic point establishes array and stores crop row center line candidate respectively as crop row center line candidate end point
Point and candidate end point;
(3)According to crop row center line solving model, randomly choose 1 candidate start point and 1 candidate end point forms artificial bee colony
One nectar source of algorithm represents a candidate crop row center line.It initializes multiple nectar sources and forms a plurality of candidate crop row straight line,
The a certain range of feature point number of statistical distance candidate's straight line, as the fitness function of the candidate straight line quality of evaluation,
The candidate straight line for choosing fitness function maximum by the multiple search of artificial bee colony algorithm is used as crop row center line.
(Three)Advantageous effect
The present invention is for insufficient existing for existing Agriculture Mobile Robot guidance path identification technology, it is proposed that a kind of natural lighting
Under the conditions of Agriculture Mobile Robot guidance path extracting method and device based on artificial bee colony algorithm.This method passes through to acquisition
Field-crop image carry out gray processing processing, will treated that gray level image is split is converted into bianry image;Using vertical
Straight sciagraphy obtains edge position information of the crop row on image top margin and image base, connects marginal point by straight line and is formed and makees
Object row regional extent;The crop row feature in the crop row regional extent is extracted using the vertical projection method based on moving window
Then according to crop row center line solving model, work is extracted by artificial bee colony algorithm in the crop row regional extent for point
Object row center line, and then obtain guidance path;This method reduce the influences that illumination variation identifies guidance path, improve and lead
The reliability of bit path extraction solves the extraction of Agriculture Mobile Robot guidance path real-time is poor, accuracy is low and to illumination
The problems such as variation is sensitive.
Description of the drawings
Fig. 1 is the agricultural movement based on artificial bee colony algorithm under the conditions of a kind of natural lighting that one embodiment of the invention provides
The flow diagram of robot navigation's path extraction method.
Fig. 2 is the schematic diagram of crop row image that one embodiment of the invention provides.
Fig. 3 is the crop row gradation of image that provides of one embodiment of the invention treated image schematic diagram.
Fig. 4 be one embodiment of the invention provide gray level image is split after bianry image schematic diagram.
Fig. 5 is the crop row gray scale vertical projection schematic diagram that one embodiment of the invention provides.
Nearest crop row is in image top margin at left and right sides of the range image center line that Fig. 6 is provided for one embodiment of the invention
With the position detection schematic diagram on base.
Fig. 7 is the schematic diagram of crop row regional extent detection that one embodiment of the invention provides.
Crop row characteristic point detects schematic diagram in the crop row regional extent that Fig. 8 is provided for one embodiment of the invention.
Fig. 9 is the crop row center line solving model feature schematic diagram that one embodiment of the invention provides.
Figure 10 is that the crop row center line that one embodiment of the invention provides and guidance path extract schematic diagram.
Figure 11 is the agriculture moving machine based on artificial bee colony algorithm under the conditions of the natural lighting that one embodiment of the invention provides
Device people's guidance path carries the structure diagram of device.
Specific embodiment
Below in conjunction with the accompanying drawings, the specific embodiment of invention is further described, following embodiment is only used for more clear
Illustrate to Chu technical scheme of the present invention, and be not intended to limit the protection scope of the present invention and limit the scope of the invention, in following embodiment selection
It is described in detail for corn image during ploughing under natural lighting item.
The agricultural based on artificial bee colony algorithm is moved under the conditions of Fig. 1 shows a kind of natural lighting provided in an embodiment of the present invention
The flow diagram of mobile robot guidance path extracting method, as shown in Figure 1, the step of this method includes is specific as follows.
S01, crop Image Acquisition.Adjustment is installed on camera height and angle on Agriculture Mobile Robot, makes camera shooting
Machine, in 60 °~70 ° angles, is 1.2 ~ 1.4m apart from ground vertical height, obtained image is as shown in Figure 2 with horizontal direction.
(1)
Cg components correspond to green and the difference of brightness:
(2)
Cg processing is obtained according to ITU-R BT.601-6 standards(3)Formula
(3)
(4)
(5)
(6)
S03, image segmentation is carried out using maximum variance between clusters, bianry image is converted gray images into, such as Fig. 4 institutes
Show.
S04, edge position information of the crop row on image top margin and image base is obtained using vertical projection method, by straight
Line connection marginal point forms crop row regional extent.The farmland image of usual visual sensor acquisition includes multirow-crop, this reality
Example is applied using two nearest crop row regional extents of vertical projection method's detecting distance picture centre line based on bianry image;
Crop row regional extent detecting step is as follows:
(1)Only include 2 kinds of white pixel and black picture element in the bianry image, white pixel gray value represents crop for 255
Information, black picture element gray value represent Soil Background information for 0.It is provided in image coordinate system, using the image upper left corner as coordinate
Origin is to the left horizontal axis positive direction, is longitudinal axis positive direction downwards.According to the view field set in bianry image, projection is calculated
The sum of each row grey scale pixel value and the view field is determined according to the gray value of pixel each in view field in region
The average gray value of interior all pixels, as shown in Figure 5.Wherein, the view field is top half 1/3 in the bianry image
1/3 region of region and lower half portion;
(2)In 1/3 view field of bianry image top half, projecting direction vertically upward, by the sum of row pixel grey scale
Project to image top margin.From left to right successively by the sum of described row pixel grey scaleIt is compared with the average gray value t, such as
FruitThe sum of described row pixel grey scale is then set as average gray value t, is otherwise set as the sum of described row pixel grey scale
0, wherein j are the integer more than or equal to 1, represent the row of image.The row pixel that the sum of row pixel for comparing jth row is arranged with jth+1
The sum of, the sum of row pixel that the sum of row pixel arranged according to the jth is arranged with jth+1 size judges crop row on image top
The left and right edges point on side, if the sum of pixel of jth row is less than the sum of pixel that jth+1 arranges, j is the left side edge of crop row
Point;If the sum of pixel of jth row is more than the sum of pixel that jth+1 arranges, j is the right side edge point of crop row;If jth arranges
The sum of pixel be equal to the sum of pixel that jth+1 arranges, then do not process;
(3)In 1/3 view field of described image lower half portion, projecting direction vertically downward, by the projection of the sum of row pixel grey scale
To image base.Calculate the difference of the sum of picture altitude value and the row pixel grey scale successively from left to right, and calculate figure
Image height angle value and the difference T of the average gray value successively will from left to rightIt is compared with T, ifMore than T then by institute
It states the sum of row pixel grey scale and is set as picture altitude value, the sum of described row pixel is otherwise set as T, wherein j is more than or equal to 1
Integer, represent the row of image.Compare the sum of the sum of the row pixel of jth row row pixel arranged with jth+1, according to the jth
The sum of the row pixel that the sum of row pixel of row is arranged with jth+1 size judges left and right edges point of the crop row on image base, if
The sum of pixel of jth row is more than the sum of pixel that jth+1 arranges, then j is the left side edge point of crop row;If the pixel of jth row
The sum of be less than the sum of pixel that jth+1 arranges, then j is the right side edge point of crop row;If the sum of pixel of jth row be equal to jth+
The sum of pixel of 1 row, then do not process;
(4)The left side edge of the crop row and right side edge difference are calculated, by the difference and default crop row width threshold value
It is compared, retains the crop row left side edge and right side edge if the difference is more than the crop row width threshold value
Otherwise information rejects the crop row left side edge and right side edge information;
(5)Crop row can be determined in the position on image top margin and base according to the crop row left side edge and right side edge information
Put region, found out at left and right sides of described image center line in image top margin and base respectively using picture centre line as boundary and
2 nearest crop row bands of position of range image center line, as shown in Figure 6;
(6)According to the nearest crop row position at left and right sides of picture centre line, by phase the same side crop row in image
Base left side edge point, right side edge point utilize straight line with crop row in image top margin left side edge point, right side edge point respectively
Connection forms crop row regional extent, as shown in Figure 7;
Specifically, for a widthThe image of pixel,Representative image width,Representative image height, ifIt represents
In imagePosition pixel gray level,For the sum of jth row pixel grey scale, T is pixel grey scale average value in view field,
The width threshold value of crop row to be detected is R,Expression formula with T is:
Wherein,Representative image width,Represent view field's height, that is, one row pixel number of upright projection;
Band of position detection algorithm of the crop row on image base is as follows:
1st, projecting direction vertically downward, calculates view field(View field is 1/3 region of image lower half portion in the present embodiment)
Interior row pixel and(j=1,2,…M)And average gray value T, as shown in figure 5, Grey curves are drop shadow curve, grey water
Flat line is average gray value;It establishes two-dimensional array A and initializes, for storing crop row number and location information, such as with() crop row number and location information are stored, whereinStorage theOn the left of a crop row
Marginal information,Storage theA crop row right side edge information;Initialize temporary variable m=1, j=1;
2nd, from left to right will in view fieldWithValue compares, ifThen=, otherwise=;
If the 3rd,>, then A [m] [0]=j represent crop row left side edge;If<, then A [m]
[1]=j represents crop row right side edge;If=, then it represents that non-crop row edge, without processing;
4th, crop row left and right edges difference, i.e. Ds=A [m] [1]-A [m] [0], if Ds are calculated>=R shows crop row location information
Effectively, retained;If Ds<R shows that crop row location information is invalid, is deleted;
5th, m=m+1, j=j+1 repeat step(2)~(5)Process,When stop search, EP (end of program);Table
Diagram picture from left to right completes the projection to all row pixels, is detected as crop row in the band of position on image base and terminates item
Part;
6th, nearest crop row position at left and right sides of the range image center line of image base is selected in A [m] [n], such as Fig. 6 institutes
Show;
Crop row is as follows in the band of position detection algorithm of image top margin:
1st, projecting direction vertically upward, calculates view field(View field is 1/3 region of image top half in the present embodiment)
Interior row pixel and(j=1,2,‥3,…M)And average gray value T, as shown in figure 5, Grey curves are drop shadow curve, ash
Color horizontal linear is average gray value;It establishes two-dimensional array A and initializes, for storing crop row number and location information, example
Such as use() crop row number and location information are stored, whereinStorage theA crop row is left
Lateral edges information,Storage theA crop row right side edge information;Initialize temporary variable m=1, j=1;
2nd, from left to right will in view fieldWithValue compares, ifThen=, otherwise=0;
If the 3rd,<, then A [m] [0]=j represent crop row left side edge;If>, then A [m] [1]=
J represents crop row right side edge;If=, then it represents that non-crop row edge, without processing;
4th, crop row left and right edges difference, i.e. Ds=A [m] [1]-A [m] [0], if Ds are calculated>=R shows crop row location information
Effectively, retained;If Ds<R shows that crop row location information is invalid, is deleted;
5th, m=m+1, j=j+1 repeat step(2)~(5)Process,When stop search, EP (end of program);Table
Diagram picture from left to right completes the projection to all row pixels, is detected as crop row in the band of position on image base and terminates item
Part;
6th, nearest crop row position at left and right sides of image top margin range image center line is selected in A [m] [n], such as Fig. 6 institutes
Show;
According to the nearest crop row position at left and right sides of picture centre line, by phase the same side crop row on image base
Left side edge point, right side edge point are connect respectively with crop in image top margin left side edge point, right side edge point using straight line, structure
Into crop row regional extent, as shown in Figure 7.
S05, crop row characteristic point in the crop row regional extent is extracted using the vertical projection method based on moving window,
As shown in Figure 8;
Crop row characteristic point detecting step is as follows in crop row regional extent:
(1)According to the crop row regional extent, a horizontal strip is divided in image apex, the horizontal strip with it is described
Crop row regional extent left and right sides edge straight line forms window, wherein the horizontal strip has certain altitude;
(2)Projecting direction vertically downward, scans by column the window from left to right, calculates the row pixel grey scale of each row
The sum of average gray value with all pixels in window;
(3)To the sum of described row pixel grey scale and the average gray value are compared successively from left to right in the window, if
The sum of described row pixel is greater than or equal to the average gray value, and the sum of described row pixel grey scale is set as average gray value,
Otherwise the sum of described row pixel is set as 0;
(4)Compare the sum of the sum of the row pixel of jth row row pixel arranged with jth+1, the sum of row pixel arranged according to the jth
The size of the sum of the row pixel arranged with jth+1, judges the left side edge point of crop row and right side edge point;If the row of jth row
The sum of pixel is less than the sum of pixel that jth+1 arranges, then j is crop row left side edge point;If the sum of pixel of jth row is more than the
The sum of pixel of j+1 row, then right side edge points of the j as crop row;If the sum of pixel of jth row is equal to the picture that jth+1 arranges
The sum of element does not process then;Wherein, j is the positive integer more than or equal to 1, represents the row of image pixel;
(5)It is compared according to the difference and the width threshold value of default crop row of the left hand edge of crop row point and right hand edge point
Judge, the crop row left side edge point and right side edge point if the difference is more than the default crop row width threshold value
For crop row edge feature point, it is otherwise pseudo-random numbers generation, is rejected;
(6)The intermediate point of the crop row left side edge point and right side edge point is calculated, using the intermediate point as crop row spy
Sign point;
(7)The horizontal strip moves down a pixel, repeats the above process, and calculates crop row characteristic point in the window, directly
Crop row region lowermost end is reached to the horizontal strip;
Specifically, for a widthThe image of pixel,Representative image width,Representative image height;IfIt represents
In imagePosition pixel gray level,For the sum of jth row pixel grey scale, T is pixel grey scale average value in window;It is provided as object
Capable width threshold value is R,Expression formula with T is:
Crop row is as follows in the band of position detection algorithm of image top margin:
1st, the horizontal strip that a height is h, the horizontal strip and the crop row region left and right sides are divided on bianry image top
Edge line forms a window;
2nd, the sum of per rows of pixels in calculation windowAnd average gray value T, from left to right will in the windowWith
Value compares, ifThen, otherwise;
If the 3rd,<, then j, which is arranged, represents candidate marginal on the left of crop row, if>Then j row represent
Candidate marginal on the right side of crop row, if the distance between right side candidate marginal and left side candidate marginal are more than given threshold
R, then it is assumed that this marginal point pair(Right side candidate marginal and left side candidate marginal)For the efficient frontier point of crop row, otherwise,
For pseudo-edge point, rejected;
4th, the midpoint of crop row left side edge point and right side edge point is calculated, using the point as crop row characteristic point;
5th, horizontal strip moves down a pixel, repeats the above process until the horizontal strip reaches crop row region most bottom
End.
S06, according to crop row in the picture the characteristics of, establish crop row center line solving model, calculated by artificial bee colony
Method extracts crop row center line in the crop row regional extent, and as shown in Figure 10, wherein grey filled lines are represented in crop row
Heart line, dash-dotted gray line are guidance path;
Specially:
(1)The characteristics of according to crop row in the picture, establishes crop row center line solving model, and so-called crop row center line is asked
It is exactly that field-crop row morphologically shows as near linear to solve model, and equation can be according to two crop row features in image
Point is determining, as shown in Figure 9.If the crop row characteristic point data space that V expressions are obtained according to the step S05,WithFor two points in V, then crop row center line equation can be expressed as:
(11)
Feature point number in the range of statistical distance straight line d, as evaluation straight line quality standard, by adjustingWithPosition is selected comprising the most straight line of characteristic point as crop row center line, wherein, the value range of d is
(1,5);
(2)The crop row characteristic point in crop row regional extent that the step S05 is obtained is divided into using 1/2 height of image as boundary
Upper and lower 2 part, using the crop row characteristic point of top half as the candidate start point of crop row center line, lower half portion crop row spy
Sign point establishes array and stores crop row center line candidate start point and candidate end point respectively as crop row center line candidate end point;
(3)According to crop row center line solving model, randomly choose 1 candidate start point and 1 candidate end point forms artificial bee colony
One nectar source of algorithm represents a candidate crop row center line.It initializes multiple nectar sources and forms a plurality of candidate crop row straight line,
The a certain range of feature point number of statistical distance candidate's straight line, as the fitness function of the candidate straight line quality of evaluation,
The candidate straight line for choosing fitness function maximum by the multiple search of artificial bee colony algorithm is used as crop row center line;
The minimum search model of artificial bee colony algorithm include nectar source, lead bee, follow bee and investigation four elements of bee and
It recruits honeybee and abandons the behavior of 2, nectar source, bee is led in algorithm, the quantity of bee is followed to be equal to nectar source quantity, basic principle is such as
Under:
If the problem of solving dimension is D, nectar source position represents the potential solution of problem, and the position of nectar source i is expressed as, then the mathematical model of artificial bee colony algorithm be:
1st, crop row number N and regional extent are obtained using vertical projection method, if(1 crop is included at least in image
Row)Characteristic point detection, otherwise EP (end of program) then are carried out to crop in bar-shaped zone;
3rd, crop row counting variable is initialized, distance threshold, nectar source quantity(It leads bee and follows bee quantity and honey
Source quantity is identical), local search threshold value limit, maximum iteration;Establish fitness function,It represents apart from straight lineIn the range of feature point number;
5th, bee is led according to formula(13)Neighborhood search is carried out, generates new nectar source, new nectar source fitness is calculated, ifIt adapts to
Degree is more thanThen, otherwiseIt remains unchanged;
6th, according to formula(14)It calculatesDependent probability, follow bee according toCarry out nectar source selection;Bee is followed to utilize formula(13)
Neighborhood search is carried out, generates new explanation, its fitness is calculated, ifFitness is more thanFitness, then, otherwise
It remains unchanged;
7th, after limit cycle, ifFitness does not change, then abandons the solution, corresponding bee to be led to be converted into detect
Bee is looked into, according to formula(12)A new explanation is generated to replace currently;
If the 9th, num+1>N, then it represents that all crop row regional extents all have stepped through, EP (end of program), otherwise num=num+1,
Return to step(4);
In this algorithm, nectar source quantity, to lead bee and follow bee quantity be m=30, and local search threshold value limit=10, maximum changes
Generation number C=50, air line distance threshold value d=2.
S07, guidance path extraction.Navigation road is calculated according to two nearest crop row straight lines of range image center line
Diameter equation, as shown in Figure 10, wherein dash-dotted gray line represents guidance path.If field-crop picture size is,Represent picture traverse,Represent picture altitude, the ordinate of image top margin point is 0, and the ordinate of image base point is, determining 2 points of crop row straight line can be calculated by artificial bee colony algorithm, then guidance path equation calculation is specific
Step is as follows:
(1)If on the left of picture centre line crop row straight line by(,)With(,)Two points determine, the straight line and image
Top margin and the intersection point on base are(,0)With(,), wherein,
,;
(2)If crop row straight line passes through on the right side of picture centre line(,)、(,)2 points, the straight line and image top margin and
The intersection point position on base(, 0)With(,), wherein,,;
(3)Calculate the midpoint between two crop row straight lines(, 0),(,), guidance path equation is solved using 2 point types, wherein, ,。
Figure 11 shows the agricultural based on artificial bee colony algorithm under the conditions of a kind of natural lighting provided in an embodiment of the present invention
Mobile Robotics Navigation path extraction device, including:
Image segmentation module M02 converts gray images into bianry image for passing through maximum variance between clusters;
Crop row regional extent determining module M03, for according to bianry image, crop row region model to be determined by vertical projection method
It encloses;
Crop row characteristic point detection module M04, for according to crop row regional extent, passing through the upright projection based on moving window
Method extracts the crop row characteristic point in the crop row regional extent;
Crop row center line extraction module M05 the characteristics of according to crop row in the picture, establishes crop row center line and solves mould
Type extracts crop row center line by artificial bee colony algorithm in the crop row regional extent;
Guidance path determining module M06, for determining to be located at according to two nearest crop row center lines of range image center line
Guidance path between two crop row center lines;
Above device is one-to-one relationship with the above method, and the present embodiment no longer carries out the implementation detail of above device detailed
It describes in detail bright.
Embodiment of above is merely to illustrate the present invention rather than limitation of the present invention, although with reference to embodiment to this hair
It is bright to be described in detail, it will be understood by those of ordinary skill in the art that, to technical scheme of the present invention carry out it is various combination,
Modification or equivalent replacement, without departure from the spirit and scope of technical solution of the present invention, the right that should all cover in the present invention is wanted
It asks in range.
Claims (6)
1. the Agriculture Mobile Robot guidance path extracting method based on artificial bee colony algorithm under the conditions of a kind of natural lighting, special
Sign is, the described method comprises the following steps:
S01, crop Image Acquisition;Make video camera and horizontal direction in 60 °~70 ° angles, apart from ground vertical height about 1.2 ~
1.4m;
S02, utilizationThe crop image of factor pair acquisition carries out gray processing processing, and colour is made
Object image is converted into gray level image;
S03, image segmentation is carried out using maximum variance between clusters, converts gray images into bianry image;
S04, crop row is obtained in the edge position information on image top margin and image base using vertical projection method, is connected by straight line
Edge fit edge point forms crop row regional extent;
S05, the crop row characteristic point in the crop row regional extent is extracted using the vertical projection method based on moving window;
S06, according to crop row in the picture the characteristics of, establish crop row center line solving model, existed by artificial bee colony algorithm
Extraction crop row center line in the crop row regional extent;
S07, it determines to be located at two crop row center lines according to two nearest crop row center lines of range image center line
Between guidance path.
2. according to the method described in claim 1, it is characterized in that, the step S02 specifically includes following steps:
(1)Crop image is transformed by RGB colorColor space:
(1)
(2)Build the Cg component unrelated with illumination on the basis of color space, Cg components correspond to green with it is bright
Spend the difference of signal:
(2)
(3)Cg processing is obtained according to ITU-R BT.601-6 standards(3)Formula:
(3)
(4)Respectively Cg, Cr, Cb component are normalized to obtainComponent:
(4)
(5)
(6)
(5)It utilizesFactor pair colour crop image carries out gray processing processing.
3. according to the method described in claim 1, it is characterized in that, the step S06 specifically includes following steps:
(1)The characteristics of according to crop row in the picture, establishes crop row center line solving model, and so-called crop row center line is asked
It is exactly that field-crop row morphologically shows as near linear to solve model, and equation can be according to two crop row features in image
Point determines, if the crop row characteristic point data space that V expressions are obtained according to the step S05,WithFor two in V
A, then crop row center line equation can be expressed as:
(7)
Feature point number in the range of statistical distance straight line d, as evaluation straight line quality standard, by adjusting
WithPosition is selected comprising the most straight line of characteristic point as crop row center line, wherein, the value range of d is(1,5);
(2)The crop row characteristic point in crop row regional extent that the step S05 is obtained is divided into using 1/2 height of image as boundary
Upper and lower 2 part, using the crop row characteristic point of top half as the candidate start point of crop row center line, lower half portion crop row spy
Sign point establishes array and stores crop row center line candidate start point and candidate end point respectively as crop row center line candidate end point;
(3)According to crop row center line solving model, randomly choose 1 candidate start point and 1 candidate end point forms artificial bee colony
One nectar source of algorithm represents a candidate crop row center line, initializes multiple nectar sources and forms a plurality of candidate crop row straight line,
The a certain range of feature point number of statistical distance candidate's straight line, as the fitness function of the candidate straight line quality of evaluation,
The candidate straight line for choosing fitness function maximum by the multiple search of artificial bee colony algorithm is used as crop row center line.
4. the Agriculture Mobile Robot guidance path extraction element based on artificial bee colony algorithm under the conditions of a kind of natural lighting, special
Sign is, including:
(1)Image gray processing processing module utilizesColored crop image is converted into ash by the factor
Spend image;
(2)Image segmentation module converts gray images into bianry image by maximum variance between clusters;
(3)Crop row regional extent determining module, for according to bianry image, crop row region model to be determined by vertical projection method
It encloses;
(4)Crop row characteristic point detection module, for according to crop row regional extent, passing through the upright projection based on moving window
Method extracts the crop row characteristic point in the crop row regional extent;
(5)Crop row center line extraction module the characteristics of according to crop row in the picture, establishes crop row center line and solves mould
Type extracts crop row center line by artificial bee colony algorithm in the crop row regional extent;
(6)Guidance path determining module, for determining to be located at according to two nearest crop row center lines of range image center line
Guidance path between two crop row center lines.
5. device according to claim 4, which is characterized in that described image gray processing processing module specifically includes:
(1)Crop image is transformed by RGB colorColor space:
(8)
(2)Build the Cg component unrelated with illumination on the basis of color space, Cg components correspond to green with it is bright
The difference of degree:
(9)
(3)Cg processing is obtained according to ITU-R BT.601-6 standards(10)Formula:
(10)
(4)Respectively Cg, Cr, Cb component are normalized to obtainComponent:
(11)
(12)
(13)
(5)It utilizesFactor pair colour crop image carries out gray processing processing.
6. device according to claim 4, which is characterized in that the crop row center line extraction module is used for:
(1)The characteristics of according to crop row in the picture, establishes crop row center line solving model, and so-called crop row center line is asked
It is exactly that field-crop row morphologically shows as near linear to solve model, and equation can be according to two crop row features in image
Point determines, if the crop row characteristic point data space that V expressions are obtained according to the crop row characteristic point detection module,WithFor two points in V, then crop row center line equation can be expressed as:
(14)
Feature point number in the range of statistical distance straight line d, as evaluation straight line quality standard, by adjusting
WithPosition is selected comprising the most straight line of characteristic point as crop row center line, wherein, the value range of d is(1,5);
(2)The crop row characteristic point in crop row regional extent that the crop row characteristic point detection module is obtained is with image 1/
2 height are divided into 2 parts up and down for boundary, using the crop row characteristic point of top half as the candidate start point of crop row center line, under
Half part crop row characteristic point establishes array and stores crop row center line candidate respectively as crop row center line candidate end point
Point and candidate end point;
(3)According to crop row center line solving model, randomly choose 1 candidate start point and 1 candidate end point forms artificial bee colony
One nectar source of algorithm represents a candidate crop row center line, initializes multiple nectar sources and forms a plurality of candidate crop row straight line,
The a certain range of feature point number of statistical distance candidate's straight line, as the fitness function of the candidate straight line quality of evaluation,
The candidate straight line for choosing fitness function maximum by the multiple search of artificial bee colony algorithm is used as crop row center line.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611076483.9A CN108133471B (en) | 2016-11-30 | 2016-11-30 | Robot navigation path extraction method and device based on artificial bee colony algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611076483.9A CN108133471B (en) | 2016-11-30 | 2016-11-30 | Robot navigation path extraction method and device based on artificial bee colony algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108133471A true CN108133471A (en) | 2018-06-08 |
CN108133471B CN108133471B (en) | 2021-09-17 |
Family
ID=62387328
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611076483.9A Active CN108133471B (en) | 2016-11-30 | 2016-11-30 | Robot navigation path extraction method and device based on artificial bee colony algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108133471B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107067430A (en) * | 2017-04-13 | 2017-08-18 | 河南理工大学 | A kind of wheatland crop row detection method of distinguished point based cluster |
CN108901540A (en) * | 2018-06-28 | 2018-11-30 | 重庆邮电大学 | Fruit tree light filling and fruit thinning method based on artificial bee colony fuzzy clustering algorithm |
CN111931789A (en) * | 2020-07-28 | 2020-11-13 | 江苏大学 | Linear crop row extraction method suitable for different illumination, crop density and growth backgrounds |
CN112395984A (en) * | 2020-11-18 | 2021-02-23 | 河南科技大学 | Method for detecting seedling guide line of unmanned agricultural machine |
CN112712534A (en) * | 2021-01-15 | 2021-04-27 | 山东理工大学 | Corn rootstock navigation datum line extraction method based on navigation trend line |
CN113111892A (en) * | 2021-05-12 | 2021-07-13 | 中国科学院地理科学与资源研究所 | Crop planting row extraction method based on unmanned aerial vehicle image |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2131184A1 (en) * | 2008-06-02 | 2009-12-09 | CNH Belgium N.V. | Crop particle discrimination methods and apparatus |
CN101604166A (en) * | 2009-07-10 | 2009-12-16 | 杭州电子科技大学 | A kind of method for planning path for mobile robot based on particle swarm optimization algorithm |
CN101807252A (en) * | 2010-03-24 | 2010-08-18 | 中国农业大学 | Crop row center line extraction method and system |
CN102999757A (en) * | 2012-11-12 | 2013-03-27 | 中国农业大学 | Leading line extracting method |
CN104866820A (en) * | 2015-04-29 | 2015-08-26 | 中国农业大学 | Farm machine navigation line extraction method based on genetic algorithm and device thereof |
-
2016
- 2016-11-30 CN CN201611076483.9A patent/CN108133471B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2131184A1 (en) * | 2008-06-02 | 2009-12-09 | CNH Belgium N.V. | Crop particle discrimination methods and apparatus |
CN101604166A (en) * | 2009-07-10 | 2009-12-16 | 杭州电子科技大学 | A kind of method for planning path for mobile robot based on particle swarm optimization algorithm |
CN101807252A (en) * | 2010-03-24 | 2010-08-18 | 中国农业大学 | Crop row center line extraction method and system |
CN102999757A (en) * | 2012-11-12 | 2013-03-27 | 中国农业大学 | Leading line extracting method |
CN104866820A (en) * | 2015-04-29 | 2015-08-26 | 中国农业大学 | Farm machine navigation line extraction method based on genetic algorithm and device thereof |
Non-Patent Citations (2)
Title |
---|
CONTRERAS-CRUZ M A 等: "Mobile robot path planning using artificial bee colony and evolutionary programming", 《APPLIED SOFT COMPUTING JOURNAL》 * |
黎竹娟: "人工蜂群算法在移动机器人路径规划中的应用", 《计算机仿真》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107067430A (en) * | 2017-04-13 | 2017-08-18 | 河南理工大学 | A kind of wheatland crop row detection method of distinguished point based cluster |
CN107067430B (en) * | 2017-04-13 | 2020-04-21 | 河南理工大学 | Wheat field crop row detection method based on feature point clustering |
CN108901540A (en) * | 2018-06-28 | 2018-11-30 | 重庆邮电大学 | Fruit tree light filling and fruit thinning method based on artificial bee colony fuzzy clustering algorithm |
CN111931789A (en) * | 2020-07-28 | 2020-11-13 | 江苏大学 | Linear crop row extraction method suitable for different illumination, crop density and growth backgrounds |
CN111931789B (en) * | 2020-07-28 | 2024-05-14 | 江苏大学 | Linear crop row extraction method suitable for different illumination, crop density and growth backgrounds |
CN112395984A (en) * | 2020-11-18 | 2021-02-23 | 河南科技大学 | Method for detecting seedling guide line of unmanned agricultural machine |
CN112395984B (en) * | 2020-11-18 | 2022-09-16 | 河南科技大学 | Method for detecting seedling guide line of unmanned agricultural machine |
CN112712534A (en) * | 2021-01-15 | 2021-04-27 | 山东理工大学 | Corn rootstock navigation datum line extraction method based on navigation trend line |
CN112712534B (en) * | 2021-01-15 | 2023-05-26 | 山东理工大学 | Corn rhizome navigation datum line extraction method based on navigation trend line |
CN113111892A (en) * | 2021-05-12 | 2021-07-13 | 中国科学院地理科学与资源研究所 | Crop planting row extraction method based on unmanned aerial vehicle image |
CN113111892B (en) * | 2021-05-12 | 2021-10-22 | 中国科学院地理科学与资源研究所 | Crop planting row extraction method based on unmanned aerial vehicle image |
Also Published As
Publication number | Publication date |
---|---|
CN108133471B (en) | 2021-09-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108133471A (en) | Agriculture Mobile Robot guidance path extracting method and device based on artificial bee colony algorithm under the conditions of a kind of natural lighting | |
Ward et al. | Deep leaf segmentation using synthetic data | |
Zhou et al. | Disease detection of Cercospora Leaf Spot in sugar beet by robust template matching | |
CN100394437C (en) | Dynamic character discriminating method of digital instrument based on BP nerve network | |
CN109635875A (en) | A kind of end-to-end network interface detection method based on deep learning | |
CN107492094A (en) | A kind of unmanned plane visible detection method of high voltage line insulator | |
CN103190224B (en) | Computer vision technique-based corn ear species test method, system and device | |
CN103530606B (en) | A kind of farm machinery navigation path extraction method under weeds environment | |
CN106384079A (en) | RGB-D information based real-time pedestrian tracking method | |
CN109376728A (en) | A kind of weeds in paddy field recognition methods and its application based on multiple features fusion and BP neural network | |
CN109726796A (en) | A kind of farm's management method neural network based | |
CN104239886B (en) | The extracting method on lawn and background line of demarcation based on graphical analysis | |
CN105825168B (en) | A kind of Rhinopithecus roxellana face detection and method for tracing based on S-TLD | |
CN108509928A (en) | For Cold region apple jujube garden field pipe operation vision guided navigation path extraction method | |
CN104751199B (en) | A kind of cotton splits bell phase automatic testing method | |
CN109522929A (en) | It is a kind of based on multi-source image fusion natural environment under wheat weeds recognition methods | |
CN109886155A (en) | Man power single stem rice detection localization method, system, equipment and medium based on deep learning | |
CN114067207A (en) | Vegetable seedling field weed detection method based on deep learning and image processing | |
CN102999757A (en) | Leading line extracting method | |
CN104034710A (en) | Chlorophyll fluorescence and imaging technology based plant disease detection method and detection device | |
CN105021196A (en) | Crop row detection method based minimum tangent circle and morphological principle | |
CN109191429A (en) | A kind of 3D printing nozzle detection method based on machine vision | |
CN107992856A (en) | High score remote sensing building effects detection method under City scenarios | |
CN109191461A (en) | A kind of Countryside Egg recognition methods and identification device based on machine vision technique | |
CN113920190A (en) | Ginkgo flower spike orientation method and system |
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 |