CN105046229B - A kind of recognition methods of crops row and device - Google Patents
A kind of recognition methods of crops row and device Download PDFInfo
- Publication number
- CN105046229B CN105046229B CN201510446384.4A CN201510446384A CN105046229B CN 105046229 B CN105046229 B CN 105046229B CN 201510446384 A CN201510446384 A CN 201510446384A CN 105046229 B CN105046229 B CN 105046229B
- Authority
- CN
- China
- Prior art keywords
- crops
- row
- characteristic point
- crops row
- linear equation
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of recognition methods of crops row and devices, this method converts the crop map picture of acquisition to bianry image by fuzzy clustering method first, then interested region ROI image is extracted from the bianry image of acquisition, wherein crops characteristic point is obtained by horizontal stripe method, and the crops characteristic point to being extracted carries out linear regression, fit crops row linear equation, the characteristic point for finally rejecting mistake by the method for multiple linear regression again, is modified crops row linear equation.The inventive system comprises fuzzy clustering module, feature point extraction module and fitting modules.The method and device of the present invention, the accuracy of identification crops row is high, speed of service block, strong antijamming capability.
Description
Technical field
The invention belongs to technical field of crop cultivation more particularly to a kind of identifications of the crops row based on image procossing
Method and device.
Background technology
China has a vast territory, and landform, climate type complexity are various, and the torrid zone, subtropical zone, temperate zone are divided into from south to north and is trembled with fear
Band, staple food crop have rice, wheat, corn and soybean etc., industrial crops to have cotton, peanut, rape, sugarcane and beet etc..
However China human mortality is numerous, cultivated area is relatively fewer, therefore agriculture especially planting industry is extremely important in the status in China, closes
It is entire national economy.With mechanization of agriculture and information-based rapid development, realize that the requirement of agricultural automation is increasingly compeled
It cuts.
Machine vision coordinates large and medium-sized agricultural machinery also more and more extensive in the utilization of agriculture field, is especially led in vision
In terms of boat and crop identification, accuracy and expense are obtained for larger improvement.Therefore the automatic weeding of crops, harvest,
In the work such as fertilising, trimming, ploughing and weeding, the identification of crops row is carried out based on image procossing to be particularly important.
The recognition methods of existing crops row mainly divides image using super green method, maximum variance between clusters, passes through
Hough transform identifies crops row.But these method image segmentations are inaccurate, cannot distinguish crops and weeds well,
And it is computationally intensive, the requirement of real-time is not achieved.In the presence of especially having a large amount of weeds in crops, it cannot obtain
Ideal result.
Invention content
The object of the present invention is to provide a kind of recognition methods of crops row and devices, divide to avoid prior art image
Inaccuracy, the not high technical problem of recognition efficiency.
To achieve the goals above, technical solution of the present invention is as follows：
A kind of recognition methods of crops row, the recognition methods include：
It converts the crop map picture of acquisition to bianry image by fuzzy clustering method；
Interested region ROI image is extracted from the bianry image of acquisition, it is special to obtain wherein crops by horizontal stripe method
Sign point；
Linear regression is carried out to the crops characteristic point extracted, fits crops row linear equation；
The characteristic point that mistake is rejected by the method for multiple linear regression, is modified crops row linear equation.
Preferably, described that the crop map picture of acquisition is converted by bianry image by fuzzy clustering method, it is to acquire
Crop map picture as entire sample, using the percentage shared by the value in the channels pixel G as sample elements, by crops and
The cluster centre of background is initialized as 0.35~0.40 and 0.30~0.35 progress fuzzy clustering and obtains respectively.
The present invention using crop map picture as entire sample, can directly with crop map as pixel pixel R G B
The value of triple channel usually carries out fuzzy clustering as sample.It is preferable that being hundred shared by the value using the channels pixel G
Ratio is divided to be initialized as the cluster centre of crops and background respectively to reduce the dimension of cluster as sample elements
0.35~0.40 and 0.30~0.35 progress fuzzy clustering obtains bianry image, improves cluster speed.
Further, the width of the crop map picture of the acquisition is W pixels, and the width of a height of H pixels, the ROI image is w
=W/2 pixels, width h=H/2 is described to obtain wherein crops characteristic point by horizontal stripe method, including：
ROI image is divided into Q items horizontal stripe of same size, uses Sp,qIndicate that white pixel occurs in the q articles horizontal stripe pth row
Number, the wherein value of p from 1 to w, w be ROI image width pixel；
For the q articles horizontal stripe, it is corresponding with threshold value uq, threshold value uqFor all S in the q articles horizontal stripep,qMean value；
Work as Sp,qLess than or equal to uqAnd Sp+1,qMore than uqWhen, it is believed that enter crops row, the row coordinate recorded at this time is p1；
Work as Sp,qMore than or equal to uqAnd Sp+1,qLess than uqWhen, it is believed that crops row is left, the row coordinate recorded at this time is p2；
Calculate difference DELTA=p of columns when entering and leaving crops row2-p1If Δ is more than the constant d of setting, recognize
For on horizontal stripe q from pth1To p2Section be crops, and take this section of midpoint be crops characteristic point；
All horizontal stripes are traversed, crops characteristic point all in ROI image is obtained；
Wherein, the value range of the constant d is：W/20<d<W/15.
Further, described that linear regression is carried out to the crops characteristic point extracted, fit crops row straight line side
Journey, including：
According to the distribution of characteristic point, characteristic point is divided into different crops rows；
For any crops row, if crops row linear equation is：
Y=kx+b
Wherein b is oblique distance, and k is slope, and calculating belongs to all characteristic points of the crops row to the distance l of the straight line:
It further calculates, belongs to all characteristic points of the crops row to the square distance and l ' of the straight line:
It is 0 to seek local derviation to k and b to the molecule of above formula and enable it, is obtained：
Wherein M is the quantity for all characteristic points for belonging to the crops row, and r belongs to 1~M, and the coordinate of r-th of characteristic point is
(xr, yr), lrIt is r-th of characteristic point at a distance from crops row straight line, the solution for solving k and b in above formula is respectively and will
It, which is brought linear equation into and obtains crops linear equation, is：
Further, the method by multiple linear regression rejects the characteristic point of mistake, to crops row straight line side
Journey is modified, including：
According to crops row linear equation, the characteristic point for belonging to the crops row is calculated to the distance of the crops row, is picked
Except distance is more than the characteristic point of the constant of setting；
After the characteristic point for rejecting mistake, crops row linear equation is fitted according to remaining characteristic point again, and again
The crops row linear equation that secondary basis newly fits calculates characteristic point to the distance of the straight line, rejects distance and be more than setting
The characteristic point of constant；
So cycle returns number until reaching maximum linear, or changes when the characteristic point quantity that linear regression is rejected is 0
In generation, stops.
The invention also provides a kind of identification device of crops row, described device includes：
Fuzzy clustering module, for converting the crop map picture of acquisition to bianry image by fuzzy clustering method；
Feature point extraction module passes through horizontal stripe for extracting interested region ROI image from the bianry image of acquisition
Method obtains wherein crops characteristic point；
Fitting module fits crops row straight line side for carrying out linear regression to the crops characteristic point extracted
Journey, and wrong characteristic point is rejected by the method for multiple linear regression, crops row linear equation is modified.
A kind of recognition methods of crops row proposed by the present invention and device divide acquisition image by fuzzy clustering
It cuts, determines characteristic point, and the affiliated crops row of position judgment using characteristic point in x-axis using horizontal stripe method, returned using linear
Return and find out transition crops row, in the characteristic point for rejecting mistake by multiple linear regression, obtains final crops row.This hair
Bright identification crops row accuracy is high, speed of service block, strong antijamming capability.
Description of the drawings
Fig. 1 is the flow chart of crops row recognition methods of the present invention；
Fig. 2 is bianry image schematic diagrames of the embodiment of the present invention；
Fig. 3 is ROI image schematic diagram of the embodiment of the present invention；
Fig. 4 is characteristic point schematic diagram in ROI image of the embodiment of the present invention.
Specific implementation mode
Technical solution of the present invention is described in further details with reference to the accompanying drawings and examples, following embodiment is not constituted
Limitation of the invention.
The general thought of profit of the invention is to use computer vision technique, to the collected crops row figure of image capture device
As carrying out analyzing processing, to identify crop row.The present embodiment is illustrated by taking corn seedling as an example.
As shown in Figure 1, a kind of recognition methods of crops row, includes the following steps：
Step S1, the crop map picture of acquisition is converted by bianry image by fuzzy clustering method.
The present embodiment indicates all samples with X, that is, the value in the channels RBG of each pixel of the crop map picture acquired
As a sample elements xi, xiCorresponding to a pixel.It is crops and background two parts by entire sample X points, so poly-
Pixel in image is divided into two class of crops and background by class number c=2.
Use fijIndicate xiThe degree of membership for belonging to jth class, uses vjThe cluster centre for indicating jth class, to each x in sample Xi
It is iterated, all sample elements, which calculate, to be completed to be followed successively by iteration successively, by the cluster for iterating to calculate out crops and background
Center.
It is needed before the iteration to cluster centre vjIt is initialized.Since crops part should be in the ideal case
Crops cluster centre v is arranged in green, the present embodiment1={ 0,255,0 }；Remaining background Green channel is logical compared to red
Road and blue channel need not occupy an leading position, and v is arranged2={ 255,0,128 }.
Degree of membership fijWith with cluster centre vjIterative formula it is as follows：
Wherein, dijIt is xiTo cluster centre vjEuclidean distance, λ be referred to as index weight, λ>1, n is the capacity of sample X.
The process iterate until | | fij(t+1)-fij(t) | | < ε have arrived at specified iterations tmax。
Two polymerization sites will be obtained after iteration：v1={ V1R, V1G, V1BAnd v2={ V2R, V2G, V2B, due to v1It is
The cluster centre of crops, the present embodiment is according to v1To calculate segmentation threshold, the preset=V of segmentation1G/(V1R+V1G+V1B)。
For arbitrary pixel xi, when the value of its RGB triple channel meets G/ (R+G+B) more than segmentation threshold, by the point
Pixel is set to 255, i.e., white, is otherwise set to 0, i.e. black, to which original image is divided into bianry image.Such as according to Fig. 2
The secondary bianry image that acquisition image obtains, wherein white pixel represent crops, and black picture element represents background.
In different situations, extraction crop information can be very good by cluster segmentation, compared to it is common super green
Method is compared with the method for the combination of Da-Jin algorithm, and fuzzy clustering can preserve more details, such as weeds, fallen leaves etc., and not
Under the conditions of illumination, accurately crops and background can be distinguished.
It is worth noting that, after cluster, for each cluster centre Vj, the percentage shared by the value in the channels G
Val is：
Data through a large number of experiments, as shown in table 1：
| Background v1(%) | Crop v2(%) |
| 0.345 | 0.403 |
| 0.335 | 0.347 |
| 0.345 | 0.386 |
| 0.336 | 0.353 |
| 0.339 | 0.412 |
| 0.345 | 0.386 |
Table 1
It is seen that the corresponding percentage val of crops cluster centre is mostly between 0.35 and 0.40, in background cluster
The corresponding percentage val of the heart is between 0.30 and 0.35.And in practical cluster process, what is played a decisive role is also the channels G
Value shared by percentage.In order to improve computational efficiency, it is preferable that directly using the percentage shared by the value in the channels G as sample
Element, therefore the dimension of sample drops to 1 from 3, and before each cluster starts, respectively just by the cluster centre of crops and background
Beginning turns to 0.35~0.40 and 0.30~0.35, and computing repeatedly when avoiding clustering iteration every time improves cluster speed.
Step S2, interested region ROI is extracted from the bianry image of acquisition, and wherein crops are obtained by horizontal stripe method
Characteristic point.
First, interested region ROI (Region Of Interest) is extracted from the bianry image of acquisition, for width
For W pixels, the acquisition image of a height of H pixels, a length of w=W/2 of the ROI generally extracted, width h=H/2, the present embodiment extraction
ROI as shown in figure 3, extraction ROI include at least a crops row, the present embodiment include two crops rows.
Horizontal stripe method is used to ROI, ROI image is divided into Q items horizontal stripe of same size, is w pixels, a height of h pictures for width
The ROI image of element, is classified as Q horizontal stripe, uses Sp,qIndicate the number that white pixel occurs in the q articles horizontal stripe pth row, wherein p
Value from 1 to w.
For the q articles horizontal stripe, it is corresponding with threshold value uq, threshold value uqFor all S in the q articles horizontal stripep,qMean value：
Traverse SP, q, with the characteristic point of following procedure extraction crops row：
(1) work as Sp,qLess than or equal to uqAnd Sp+1,qMore than uqWhen, illustrate to enter crops row, the row coordinate recorded at this time is
p1；
(2) work as Sp,qMore than or equal to uqAnd Sp+1,qLess than uqWhen, illustrate to leave crops row, the row coordinate recorded at this time is
p2；
(3) when entering and leaving crops row every time, difference DELTA=p of columns when entering and leaving crops row is calculated2-
p1If Δ is more than the constant d of setting, then it is assumed that from pth on horizontal stripe q1To p2Section be crops, and take this section of midpoint be agriculture
Crop characteristic point.
Wherein, d is constant, it can be understood as the width of crops row, value range are：W/20<d<W/15.
By traversing all horizontal stripes, several characteristic points can be obtained, as shown in figure 4, wherein stain is characterized a little.
Step S3, linear regression is carried out to the crops characteristic point extracted, obtains crops row linear equation.
According to the distribution of characteristic point, characteristic point is divided into different crops rows.It is general in the ROI that the present embodiment creates
Only retain two crops rows, judge the x-axis coordinate of all characteristic points, if coordinate is less than the half of width, which is attributed to
Otherwise the crops row in left side is attributed to right side.
For any crops row, crops are obtained by linear regression according to all characteristic points for belonging to the crops row
Row linear equation, process are as follows：
Assuming that the coordinate of characteristic point is (x, y), the equation of crops row straight line is：
Y=kx+b
Wherein, b is oblique distance, and k is slope.
Then the distance l of characteristic point to the crops row straight line is:
The quadratic sum l ' of all feature distance between beeline and dot is:
Wherein M is characterized quantity a little.
As it can be seen that if characteristic point (x, y) on this line, l 0, but can not possibly all characteristic points all fallen just at this
On straight line, by asking k and b that l ' is made to be minimized, i.e., pairMolecule k and b seek local derviation and enable its be 0：
Wherein M is the quantity for all characteristic points for belonging to the crops row, and r belongs to 1~M, and the coordinate of r-th of characteristic point is
(xr, yr), lrIt is r-th of characteristic point at a distance from crops row straight line, solves the solution of above formulaWithIt willWithBring straight line into
Equation, the equation for obtaining crops row straight line are：
So that characteristic point is substantially distributed near the straight line both sides.
Step S4, the characteristic point that mistake is rejected by the method for multiple linear regression carries out crops row linear equation
It corrects.
In an ideal case, it can have been obtained compared with subject to by the crops row linear equation obtained in step S3 at this time
True crops row, but may not include only two crops rows, and illumination, weather, the external worlds such as weeds in ROI in actual conditions
Factor can also influence accuracy.Accurate crop row in order to obtain retains correct current embodiment require that rejecting the characteristic point of mistake
Characteristic point.
In all external environments, the intrusion of other rows and the influence of weeds are maximum in ROI, and the crops of these intrusions
Row and weeds are spectrally close with target, so the present embodiment judges which characteristic point needs using the position relationship of point and line
It rejects.
The method that the present embodiment introduces multiple linear regression has been obtained for crops row linear equation by S3And it is picked by the distance l that distance between beeline and dot formula calculates characteristic point to the straight line when l is more than W/15
Except the point, W is the width of image.
After the characteristic point for rejecting mistake, the linear equation of crops row is fitted again according to remaining characteristic point, and
Again according to the linear equation of the crops row newly fitted, the distance l that characteristic point arrives the straight line is calculated, when l is more than W/15,
Reject the point.
So cycle returns number until reaching maximum linear, and the present embodiment is 10 times, or as the spy of linear regression rejecting
Iteration stopping when sign point quantity is 0.
The present embodiment separately carries out linear regression in order to improve efficiency, to two crops rows, when on a crops row
Characteristic point without mistake when, directly skip and another crops row solved.
The present embodiment also proposed a kind of identification device of crops row corresponding to the above method, and described device includes：
Fuzzy clustering module, for converting the crop map picture of acquisition to bianry image by fuzzy clustering method；
Feature point extraction module passes through horizontal stripe for extracting interested region ROI image from the bianry image of acquisition
Method obtains wherein crops characteristic point；
Fitting module fits crops row straight line side for carrying out linear regression to the crops characteristic point extracted
Journey, and wrong characteristic point is rejected by the method for multiple linear regression, crops row linear equation is modified.
Operation performed by each module is corresponding with above-mentioned crops row recognition methods, and which is not described herein again.
The prior art is as shown in table 2 using the comparison for the multiple linear regression that Hough transform and the present embodiment propose：
Table 2
Numerical value in table 2 is mean values.As can be seen from the table, it is either still taken in accuracy, Hou Zhejun
Better than the former, and when image gradually increases, the calculation amount increase tendency of Hough transform is also faster than the method for the present embodiment.
Experimental data after being tested to corn seedling is shown, under conditions of different weather environment, identifies the accurate of crops row
Rate reaches 96%, and for error at 2 ° or so, the speed of service can reach the requirement of real-time within 10ms.
The above embodiments are merely illustrative of the technical solutions of the present invention rather than is limited, without departing substantially from essence of the invention
In the case of refreshing and its essence, those skilled in the art make various corresponding changes and change in accordance with the present invention
Shape, but these corresponding change and deformations should all belong to the protection domain of appended claims of the invention.
Claims (8)
1. a kind of recognition methods of crops row, which is characterized in that the recognition methods includes：
It converts the crop map picture of acquisition to bianry image by fuzzy clustering method；
Interested region ROI image is extracted from the bianry image of acquisition, and wherein crops characteristic point is obtained by horizontal stripe method；
Linear regression is carried out to the crops characteristic point extracted, fits crops row linear equation；
The characteristic point that mistake is rejected by the method for multiple linear regression, is modified crops row linear equation；
Wherein, the width of the crop map picture of the acquisition is W pixels, and the width of a height of H pixels, the ROI image is w=W/2 pictures
Element, a height of h=H/2 is described to obtain wherein crops characteristic point by horizontal stripe method, including：
ROI image is divided into Q items horizontal stripe of same size, uses Sp,qIndicate time that white pixel occurs in the q articles horizontal stripe pth row
Number, for the wherein value of p from 1 to w, w is the width pixel of ROI image；
For the q articles horizontal stripe, it is corresponding with threshold value uq, threshold value uqFor all S in the q articles horizontal stripep,qMean value；
Work as Sp,qLess than or equal to uqAnd Sp+1,qMore than uqWhen, it is believed that enter crops row, the row coordinate recorded at this time is p1；
Work as Sp,qMore than or equal to uqAnd Sp+1,qLess than uqWhen, it is believed that crops row is left, the row coordinate recorded at this time is p2；
Calculate difference DELTA=p of columns when entering and leaving crops row2-p1If Δ is more than the constant d of setting, then it is assumed that horizontal
From pth on q1To p2Section be crops, and take this section of midpoint be crops characteristic point；
All horizontal stripes are traversed, crops characteristic point all in ROI image is obtained；
Wherein, the value range of the constant d is：W/20<d<W/15.
2. the recognition methods of crops row according to claim 1, which is characterized in that described to be incited somebody to action by fuzzy clustering method
The crop map picture of acquisition is converted into bianry image, is using the crop map picture of acquisition as entire sample, with the channels pixel G
Value shared by percentage as sample elements, the cluster centre of crops and background is initialized as to 0.35~0.40 He respectively
0.30~0.35 progress fuzzy clustering obtains.
3. the recognition methods of crops row according to claim 1, which is characterized in that described special to the crops extracted
Sign point carries out linear regression, fits crops row linear equation, including：
According to the distribution of characteristic point, characteristic point is divided into different crops rows；
For any crops row, if crops row linear equation is：
Y=kx+b
Wherein b is oblique distance, and k is slope, and calculating belongs to all characteristic points of the crops row to the distance l of the straight line:
It further calculates, belongs to all characteristic points of the crops row to the square distance and l ' of the straight line:
It is 0 to seek local derviation to k and b to the molecule of above formula and enable it, is obtained：
Wherein M is the quantity for all characteristic points for belonging to the crops row, and r belongs to 1~M, and the coordinate of r-th of characteristic point is (xr,
yr), lrIt is r-th of characteristic point at a distance from crops row straight line, the solution for solving k and b in above formula is respectivelyWithBy its band
Enter linear equation and obtain crops row linear equation and is：
4. the recognition methods of crops row according to claim 3, which is characterized in that described by multiple linear regression
Method rejects the characteristic point of mistake, is modified to crops row linear equation, including：
According to crops row linear equation, calculating belongs to the characteristic point of the crops row and arrives the distance of the crops row, rejecting away from
Characteristic point from the constant more than setting；
After the characteristic point for rejecting mistake, crops row linear equation, and root again are fitted according to remaining characteristic point again
According to the crops row linear equation newly fitted, characteristic point is calculated to the distance of the straight line, rejects the constant that distance is more than setting
Characteristic point；
So cycle returns number until reaching maximum linear, or iteration is stopped when the characteristic point quantity that linear regression is rejected is 0
Only.
5. a kind of identification device of crops row, which is characterized in that described device includes：
Fuzzy clustering module, for converting the crop map picture of acquisition to bianry image by fuzzy clustering method；
Feature point extraction module is obtained for extracting interested region ROI image from the bianry image of acquisition by horizontal stripe method
Take wherein crops characteristic point；
Fitting module fits crops row linear equation for the crops characteristic point progress linear regression to being extracted, and
The characteristic point that mistake is rejected by the method for multiple linear regression, is modified crops row linear equation；
Wherein, the width of the crop map picture of the acquisition is W pixels, and the width of a height of H pixels, the ROI image is w=W/2 pictures
Element, a height of h=H/2 when the feature point extraction module obtains wherein crops characteristic point by horizontal stripe method, execute following behaviour
Make：
ROI image is divided into Q items horizontal stripe of same size, uses Sp,qIndicate time that white pixel occurs in the q articles horizontal stripe pth row
Number, for the wherein value of p from 1 to w, w is the width pixel of ROI image；
For the q articles horizontal stripe, it is corresponding with threshold value uq, threshold value uqFor all S in the q articles horizontal stripep,qMean value；
Work as Sp,qLess than or equal to uqAnd Sp+1,qMore than uqWhen, it is believed that enter crops row, the row coordinate recorded at this time is p1；
Work as Sp,qMore than or equal to uqAnd Sp+1,qLess than uqWhen, it is believed that crops row is left, the row coordinate recorded at this time is p2；
Calculate difference DELTA=p of columns when entering and leaving crops row2-p1If Δ is more than the constant d of setting, then it is assumed that horizontal
From pth on q1To p2Section be crops, and take this section of midpoint be crops characteristic point；
All horizontal stripes are traversed, crops characteristic point all in ROI image is obtained；
Wherein, the value range of the constant d is：W/20<d<W/15.
6. the identification device of crops row according to claim 5, which is characterized in that the fuzzy clustering module is passing through
It is using the crop map picture of acquisition as entire sample when fuzzy clustering method converts the crop map picture of acquisition to bianry image
This is distinguished the cluster centre of crops and background initial using the percentage shared by the value in the channels pixel G as sample elements
0.35~0.40 and 0.30~0.35 progress fuzzy clustering is turned to obtain.
7. the identification device of crops row according to claim 5, which is characterized in that the fitting module is to being extracted
Crops characteristic point carries out linear regression, when fitting crops row linear equation, executes following operation：
According to the distribution of characteristic point, characteristic point is divided into different crops rows；
For any crops row, if crops row linear equation is：
Y=kx+b
Wherein b is oblique distance, and k is slope, and calculating belongs to all characteristic points of the crops row to the distance l of the straight line:
It further calculates, belongs to all characteristic points of the crops row to the square distance and l ' of the straight line:
It is 0 to seek local derviation to k and b to the molecule of above formula and enable it, is obtained：
Wherein M is the quantity for all characteristic points for belonging to the crops row, and r belongs to 1~M, and the coordinate of r-th of characteristic point is (xr,
yr), lrIt is r-th of characteristic point at a distance from crops row straight line, the solution for solving k and b in above formula is respectivelyWithBy its band
Enter linear equation and obtain crops row linear equation and is：
8. the identification device of crops row according to claim 7, which is characterized in that the fitting module is by multiple
The method of linear regression rejects the characteristic point of mistake, when being modified to crops row linear equation, executes following operation：
According to crops row linear equation, calculating belongs to the characteristic point of the crops row and arrives the distance of the crops row, rejecting away from
Characteristic point from the constant more than setting；
After the characteristic point for rejecting mistake, crops row linear equation, and root again are fitted according to remaining characteristic point again
According to the crops row linear equation newly fitted, characteristic point is calculated to the distance of the straight line, rejects the constant that distance is more than setting
Characteristic point；
So cycle returns number until reaching maximum linear, or iteration is stopped when the characteristic point quantity that linear regression is rejected is 0
Only.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510446384.4A CN105046229B (en) | 2015-07-27 | 2015-07-27 | A kind of recognition methods of crops row and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510446384.4A CN105046229B (en) | 2015-07-27 | 2015-07-27 | A kind of recognition methods of crops row and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105046229A CN105046229A (en) | 2015-11-11 |
| CN105046229B true CN105046229B (en) | 2018-11-02 |
Family
ID=54452758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510446384.4A Expired - Fee Related CN105046229B (en) | 2015-07-27 | 2015-07-27 | A kind of recognition methods of crops row and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105046229B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102524360B1 (en) * | 2016-07-18 | 2023-04-24 | 강원대학교산학협력단 | Determination technique of corn grade using fuzzy clustering for the control of sterilization at producing corn sterilized at high temperature and pressure |
| 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 |
| CN114002951B (en) * | 2021-09-16 | 2023-12-29 | 江苏农林职业技术学院 | Fuzzy control irrigation method for hard rice seedling raising |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521596A (en) * | 2011-12-09 | 2012-06-27 | 中国科学院长春光学精密机械与物理研究所 | Method for identifying and dividing crop objects under limited cognitive condition |
| CN103632381A (en) * | 2013-12-08 | 2014-03-12 | 中国科学院光电技术研究所 | Method for tracking extended targets by means of extracting feature points by aid of frameworks |
-
2015
- 2015-07-27 CN CN201510446384.4A patent/CN105046229B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521596A (en) * | 2011-12-09 | 2012-06-27 | 中国科学院长春光学精密机械与物理研究所 | Method for identifying and dividing crop objects under limited cognitive condition |
| CN103632381A (en) * | 2013-12-08 | 2014-03-12 | 中国科学院光电技术研究所 | Method for tracking extended targets by means of extracting feature points by aid of frameworks |
Non-Patent Citations (1)
| Title |
|---|
| "基于最小二乘法的早期玉米作物行检测研究";汪博 等;《浙江理工大学学报（自然科学版）》;20150710;第33卷(第4期);第547-551页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105046229A (en) | 2015-11-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bao et al. | Field-based architectural traits characterisation of maize plant using time-of-flight 3D imaging | |
| CN108846832B (en) | Multi-temporal remote sensing image and GIS data based change detection method and system | |
| Xiang et al. | Recognition of clustered tomatoes based on binocular stereo vision | |
| Tang et al. | Weed detection using image processing under different illumination for site-specific areas spraying | |
| Vidović et al. | Crop row detection by global energy minimization | |
| CN109146948B (en) | Crop growth phenotype parameter quantification and yield correlation analysis method based on vision | |
| Wu et al. | Detection and counting of banana bunches by integrating deep learning and classic image-processing algorithms | |
| CN111368736B (en) | Rice refined estimation method based on SAR and optical remote sensing data | |
| CN104598908A (en) | Method for recognizing diseases of crop leaves | |
| CN102663397B (en) | Automatic detection method of wheat seedling emergence | |
| CN111340826A (en) | Single tree crown segmentation algorithm for aerial image based on superpixels and topological features | |
| Ma et al. | Automatic detection of crop root rows in paddy fields based on straight-line clustering algorithm and supervised learning method | |
| CN108364300A (en) | Vegetables leaf portion disease geo-radar image dividing method, system and computer readable storage medium | |
| Ji et al. | In-field automatic detection of maize tassels using computer vision | |
| CN105046229B (en) | A kind of recognition methods of crops row and device | |
| CN107680098A (en) | A kind of recognition methods of sugarcane sugarcane section feature | |
| CN113536958B (en) | Navigation path extraction method, device, agricultural robot and storage medium | |
| Liu et al. | A lightweight method for grape berry counting based on automated 3D bunch reconstruction from a single image | |
| Zhou et al. | Autonomous detection of crop rows based on adaptive multi-ROI in maize fields | |
| CN102855485A (en) | Automatic wheat earing detection method | |
| Bai et al. | A fast and robust method for plant count in sunflower and maize at different seedling stages using high-resolution UAV RGB imagery | |
| CN110188657A (en) | Corn arid recognition methods based on crimping blade detection | |
| CN110610438B (en) | Crop canopy petiole included angle calculation method and system | |
| Izzuddin et al. | Analysis of multispectral imagery from unmanned aerial vehicle (UAV) using object-based image analysis for detection of ganoderma disease in oil palm | |
| Che et al. | The effect of growth stage and plant counting accuracy of maize inbred lines on LAI and biomass prediction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181102 Termination date: 20200727 |