CN102360428A - Automatic detection methods for trefoil stage and seven-leaf stage of corn - Google Patents

Automatic detection methods for trefoil stage and seven-leaf stage of corn Download PDF

Info

Publication number
CN102360428A
CN102360428A CN2011101626269A CN201110162626A CN102360428A CN 102360428 A CN102360428 A CN 102360428A CN 2011101626269 A CN2011101626269 A CN 2011101626269A CN 201110162626 A CN201110162626 A CN 201110162626A CN 102360428 A CN102360428 A CN 102360428A
Authority
CN
China
Prior art keywords
corn
image
leaf
tri
observation area
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN2011101626269A
Other languages
Chinese (zh)
Other versions
CN102360428B (en
Inventor
曹治国
余正泓
白晓东
吴茜
鄢睿丞
朱磊
张雪芬
薛红喜
李翠娜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN 201110162626 priority Critical patent/CN102360428B/en
Publication of CN102360428A publication Critical patent/CN102360428A/en
Application granted granted Critical
Publication of CN102360428B publication Critical patent/CN102360428B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Analysis (AREA)

Abstract

The invention provides automatic detection methods for a trefoil stage and a seven-leaf stage of a corn. According to the methods, segmentation is automatically carried out on a collected downward viewing image of a corn in the field and an image feature is extracted; and it is determined whether the corn in the image area enters a trefoil stage by utilizing the extracted image feature; and furthermore, on the basis of utilization of a feature of an initial image at the trefoil stage, it is automatically determined whether the corn enters a seven-leaf stage with regard to different sowing modes. According to the invention, an image feature parameter that characterizes the number of leaves of a corn is utilized as a determination basis; real-time determination is carried out on a growth period of the corn; and the detection result has high accuracy; therefore, the methods provided in the invention have an important guiding significance for an analysis of a relationship between a corn developmental phase and a meteorological condition, identification on an agricultural meteorological condition for corn growth as well as for farming activities on corns.

Description

The automatic testing method of a kind of corn tri-leaf period and seven leaf phases
Technical field
The invention belongs to Digital Image Processing and agrometeorological observation crossing domain; Be specifically related to the automatic testing method of a kind of corn tri-leaf period and seven leaf phases; Be object as sequence promptly, utilize characteristics of image to detect the method whether corn arrives tri-leaf period and seven leaf phases with view under the field corn of taking.
Background technology
Corn is one of main cereal crops of China, and cultivated area very extensively.In order to improve maize yield and quality, need understand its rate of development and process, and analyze the relation between its each puberty and the meteorological condition, thus the agrometeorological conditions of evaluation corn growth.Yet each budding observation mainly is the mode through artificial observation for corn, is observed the influence of personnel's subjective factor bigger for a long time; Because the corn planting region is wide, observation cycle is long, utilize manpower to observe also economical inadequately simultaneously.Therefore, through looking sequence image under the corn of taking,, its puberty observed automatically just seem very necessary by the means of Flame Image Process.Corn tri-leaf period and seven leaf phases are two important puberty links of corn nourishment growth phase; Effectively and accurate recognition these two periods; Be two important contents of agrometeorological observation, the present invention utilizes to look sequence image under the corn two puberties are discerned automatically.
Liu Hong in 2007 sees etc. and in the paper of delivering on " agriculture network information " " image processing techniques is being extracted the application on the corn image framework ", to utilize image processing techniques that the extractive technique of corn plant type skeleton is studied; But the method for this paper only is fit to the individual plant corn under the single background, and also inapplicable to the field condition of complicacy; Through the method for extracting the ground coverage population leaf area index (LAI) and dry-matter accumulation (DMA) value of summer corn are estimated in the paper " based on the summer corn colony growing way study on monitoring of image processing techniques " that the Li Rong spring in 2010 etc. deliver on " corn science "; Set up the regression relation model of coverage and leaf area index (LAI) and dry-matter accumulation (DMA); Thereby accomplish estimation, but only utilize this characteristics of image of coverage not observe exactly the puberty of corn to summer corn colony growing way; Ma Yanping proposed the measuring method based on the field corn growth parameter(s) of binocular stereo vision in Master's thesis " based on winter wheat, the technical research of summer corn growing way remote dynamic monitoring of digital picture " in 2010; Some growth parameter(s)s to winter wheat and summer corn are measured; But the hardware cost of this method is higher; And because the actual serious shielding of field crop, environmental impact factor is more, and the resolution of image is also limited; Therefore be difficult in practical application that field corn is carried out plant type and carry out three-dimensional reconstruction, so this method is infeasible in budding observation.In sum, although at present aspect the plant growth monitoring existing many correlation units technology occur, all because of certain limitation, the crop puberty that is difficult to apply it to actual land for growing field crops environment automatically observation come up.
Summary of the invention
The object of the invention is to provide the automatic testing method of a kind of corn tri-leaf period and seven leaf phases, and looking like with view under the corn of actual farmland collection is object, utilizes the maize seedling image characteristic point that obtains, and detects the time that corn arrives tri-leaf period and seven leaf phases exactly.
The automatic testing method in a kind of corn tri-leaf period is specially:
(1) to the corn map to be measured after emerging as segmented extraction n maize seedling connected region, the end of adding up each connected domain EndNum that counts j, j=1...n;
(2) count StemNum according to corn plants number and the corn plants that terminal mapping relations of counting obtain each connected domain j, said corn plants number and terminal mapping relations of counting are to utilize the historical sample graphical analysis statistics in corn tri-leaf period to obtain;
(3) the average end that calculates each connected domain is counted MeanEndNum j = EndNum j StemNum j ;
(4) the average end that calculates corn map picture to be measured is counted MeanEndNum = Σ j = 1 n MeanEndNum j n ;
(5) if MeanEndNum >=tri-leaf period judgment threshold, show that then corn has arrived tri-leaf period.
Further, said tri-leaf period judgment threshold span be [2,3].
The automatic testing method of a kind of seven leaf phases of corn; Confirm that according to the automatic testing method in aforesaid corn tri-leaf period corn gets into tri-leaf period; Utilize corn image in tri-leaf period to detect according to the following mode seven leaf phases of carrying out: seeding method is selected option A during for bunch planting for use, and seeding method is selected option b sowing time for use for bar;
Option A:
(A1) confirm the unique point sum and the seven leaf phase judgment thresholds of emphasis observation area, tri-leaf period image;
Said emphasis observation area is the barycenter M with n maize seedling connected region of corn image in tri-leaf period 1..., M nBe the center, R is a n square area of the length of side;
The unique point of said image adds up in the image characteristic of each connected region sum of counting, and the characteristic of said connected region is counted and counted and the trident sum of counting for the end of connected region;
Said seven leaf phase judgment thresholds comprise seven leaf phases arrival final singling threshold value T afterwards 1Arrive final singling threshold value T before with seven leaf phases 2, wherein
Figure BDA0000068871190000031
EndJunctionNum jThe characteristic that is j connected region is counted, k ∈ [2,3]; T 2=k*M*N, M are the cave number in the image in tri-leaf period, N be tri-leaf period the average characteristics of image count N = Σ j = 1 n EndJunctionNum i Σ j = 1 n StemNum j ;
(A2) the unique point sum B of statistics image to be detected, and judge the corn final singling whether in the image to be detected, if there is not final singling, then get into step (A3), otherwise, get into step (A4);
(A3) if B＞T 1, show that then the corn in the image to be detected arrived for seven leaf phases, finish;
(A4) if B＞T 2, show that then the corn in the image to be detected arrived for seven leaf phases, finish;
Said final singling is carried out according to following mode: the note corn is respectively P at i-1 and i image constantly I-1And P i, calculate J=1...n,
Figure BDA0000068871190000034
Be P I-1In j emphasis observation area in the maize leaves elemental area,
Figure BDA0000068871190000035
Be P iIn j emphasis observation area in the maize leaves elemental area, statistics Δ H j＞T 3Emphasis observation area number J, if
Figure BDA0000068871190000036
Then show P iIn corn final singling, T 3∈ [0.3,0.6], T 4∈ [0.2,0.4];
Option b:
(B1) according to the mode of step (A1) confirm the emphasis observation area and tri-leaf period image unique point sum C = Σ i = 1 n EndJunctionNum i ;
(B2) the sum D that counts of the characteristic in each emphasis observation area in the statistics image to be detected, distal point and trident in the observation area sum of counting of attaching most importance to of counting of the characteristic in the said emphasis observation area;
(B3) judge corn final singling whether in the image to be detected,, then get into step (B4) if there is not final singling, otherwise, step (B5) got into;
(B4) if D＞k*C shows that then the corn in the image to be detected arrived for seven leaf phases, finish;
(B5) Δ H is satisfied in rejecting j＞T 5The emphasis observation area, the characteristic of emphasis observation area connected domain of correspondence in corn image in tri-leaf period of sum E and the reservation sum F that counts that counts of the characteristic in the emphasis observation area that statistics keeps;
(B6) if E＞k*F shows that then the corn in the image to be detected arrived for seven leaf phases, finish.
Further, said cave number is confirmed according to following mode: the barycenter M that obtains each connected domain of image 1..., M n, calculate centroid distance dis I, j, i, if j=1...n is dis I, jLess than threshold alpha, then reject barycenter M iWith M j, and with M iWith M jBetween mid point be designated as barycenter, recomputate centroid distance, more relatively with itself and threshold alpha, so circulation, till all centroid distances were all greater than α, the final barycenter number that keeps was the cave number.
Technique effect of the present invention is embodied in: the present invention looks like to cut apart to view under the field corn of being gathered automatically; And characteristics of image extracted automatically; Utilize the characteristics of image that is extracted; Judge whether the corn in this image-region gets into tri-leaf period, and then utilize the characteristic of initial pictures in tri-leaf period to different seeding methods whether corn was got into for seven leaf phases and judge automatically.This method with the characteristics of image parameter that characterizes the maize leaf number as basis for estimation; In real time corn growing season is judged; The testing result accuracy rate is high; To analyzing the relation between corn puberty and the meteorological condition, identify the agrometeorological conditions of corn growth and the farming activities of corn is all had important directive significance.
Description of drawings
Fig. 1 is overall flow figure of the present invention;
Fig. 2 is the corn automatic training stage process flow diagram that detects in tri-leaf period;
Fig. 3 is corn history image in a tri-leaf period exemplary plot;
Fig. 4 carries out the figure as a result after the segmented extraction to Fig. 3;
Fig. 5 is the skeleton diagram after Fig. 4 handles through morphology;
Fig. 6 carries out the figure as a result after the skeleton distal point extracts to Fig. 5;
Fig. 7 is corn strain in tri-leaf period number and the end function curve fitted figure of counting;
Fig. 8 is the corn automatic detection-phase process flow diagram that detects in tri-leaf period;
Automatic testing process figure of seven leaf phases during Fig. 9 bunch planting;
Process flow diagram is confirmed in the emphasis observation area during Figure 10 bunch planting;
Emphasis observation area synoptic diagram during Figure 11 bunch planting;
During Figure 12 bunch planting tri-leaf period initial pictures unique point mark synoptic diagram;
Figure 13 cave number is confirmed process flow diagram;
During Figure 14 bunch planting tri-leaf period image average characteristics definite process flow diagram of counting;
Final singling decision flow chart during Figure 15 bunch planting;
Automatic testing process figure of Figure 16 bar seven leaf phases of sowing time;
Final singling sowing time of Figure 17 bar is judged and is upgraded emphasis area of observation coverage process flow diagram.
Embodiment
The present invention provides the automatic testing method of a kind of corn tri-leaf period and seven leaf phases, and looking like with view under the corn of actual farmland collection is object, utilizes the image characteristic point that extracts, and comes to detect exactly the time that corn arrives tri-leaf period and seven leaf phases.Specify embodiment of the present invention and implementation step below in conjunction with accompanying drawing.
Fig. 1 is overall flow figure of the present invention, is divided into two parts, and first part is through carrying out feature extraction to the image after emerging; Counting with the average end of image is basis for estimation; Automatically identify the time of taking place tri-leaf period, and then get into second part, promptly through the characteristics of image that will begin tri-leaf period as important indicator; Successively the image after tri-leaf period is carried out feature extraction, discern seven leaf phase time of origins.Therefore, the beginning image in tri-leaf period is a necessary condition of successfully discerning for seven leaf phases among the present invention.
Introduce the detection method of tri-leaf period and seven leaf phases below respectively:
1. the detection in tri-leaf period is divided into training and detects two stages:
(1) training stage: the fundamental purpose in this stage is through the corn history image data in tri-leaf period, sets up corn strain in tri-leaf period number and terminal funtcional relationship of counting, and its flow process is as shown in Figure 2, and concrete steps are following:
A. at first obtain several corns history image in tri-leaf period, as shown in Figure 3, then image is carried out segmented extraction maize seedling wherein, only contained the bianry image of seedling, as shown in Figure 4; Dividing method can adopt environment self-adaption dividing method (Lei F.Tian.Environmentally adaptive segmentation algorithm for outdoor image segmentation.Computers and electronics in agriculture; 1998,21:153～168), ultra green operator dividing method (D.M.Woebbecke, G.E.Meyer; K.Von Bargen; D.A.Mortensen.Color Indices for weed identification under various soil, residue, and lighting conditions.Transactions of the ASAE; 1995; 259～269), 38 (1): based on crop image partition method (Zheng L, Zhang J, the Wang Q.Mean-shift-based color segmentation of images containing green vegetation.Computers and Electronics in Agriculture of Mean Shift; 2009,65:93-98) or the like.
B. the bianry image that step a is obtained carries out morphology to be handled, and the connected domain in the image is carried out refinement, obtains its skeleton, as shown in Figure 5.Again through Cecilia Di Ruberto; Recognition of shapes by attributed skeletal graphs, Pattern Recognition 37,21-31; Method in 2004; Distal point (end points) quantity of extracting and adding up each connected domain skeleton, as shown in Figure 6, blockage is represented distal point among the figure.Suppose that certain width of cloth image has m connected domain, the connected domain end that statistics wherein is numbered i is counted and is EndNum i, i=1...m;
C. StemNum is counted in its corresponding strain of each connected domain complicate statistics among the step b i, i=1...m, (x, y), promptly when StemNum is x (x＞0), the end EndNum that counts is y={EndNum thereby can obtain strain number and terminal mapping of counting i| StemNum i=x, i=1...m}.
A. last, utilize conic fitting to go out that x is counted in strain and terminal Function Mapping of counting between the y concerns y=F (x), present embodiment fitting function relational expression is: y=1.5891x 2-3.773x+5.0172; As shown in Figure 7.
(2) detection-phase: the funtcional relationship through training obtains, the corn map after emerging is looked like to detect automatically, its flow process is as shown in Figure 8, and concrete steps are following:
A. at first, the corn map to be detected after emerging is looked like to carry out self-adaptation segmented extraction maize seedling, and the end of adding up each connected domain is counted the same training stage of method.Suppose that image has n connected domain, through statistics, the end that can obtain all connected domains EndNum that counts j, j=1...n;
B. the end of again step a the being added up EndNum that counts j, j=1...n is updated to functional relation y=F (x)=ax that the training stage match obtains respectively 2Among+the bx+c, through formula:
x = - b ± b 2 - 4 a ( c - EndNum j ) 2 a , J=1...n obtains the strain of each connected domain and counts x=StemNum j, x＞0, j=1...n;
C. follow the end of each connected domain that step a the is obtained EndNum that counts j, StemNum is counted in each connected domain strain that j=1...n and step b calculate j, the j=1...n correspondence is divided by, and can obtain the average end of each connected domain and count, promptly MeanEndNum j = EndNum j StemNum j ; J=1...n;
D. the average end of step c being obtained of all connected domains summation of counting; Divided by the connected domain number, calculating formula is again:
Figure BDA0000068871190000072
can obtain entire image the average end MeanEndNum that counts;
E. last, the MeanEndNum if the average end of entire image obtained of steps d is counted＞tri-leaf period judgment threshold representes that then this image arrives tri-leaf period, and the shooting time of this image is designated as time of arrival in tri-leaf period, judges tri-leaf period and finishes; Otherwise expression is no show tri-leaf period also, continues to handle next time chart picture.Tri-leaf period, judgment threshold needed to confirm that through experiment the value of present embodiment is 2.4 according to camera parameter and camera heights, and image resolution ratio is 1824 * 1368 pixels, and camera heights is 5 meters.
The automatic detection of seven leaf phases of corn will be because the difference of seeding method will be divided into two kinds of situation of bunch planting and drilling.Simultaneously, because before and after seven leaf phases arrived, have the farming activities of final singling, the seedling number of corn will change this moment, and in order to adapt to this situation, the present invention will discern the carrying out of final singling automatically, makes the automatic detection algorithm of seven leaf phases have better robustness.To divide two types of situation below, introduce automatic concrete technical scheme and the realization flow that detects of seven leaf phases of the present invention respectively:
(1) when the seeding method of corn is bunch planting, detect automatically according to following steps, its flow process is as shown in Figure 9:
A. at first confirm the emphasis observation area of image, the unique point sum A that obtains the image in tri-leaf period and the judgment threshold of confirming for seven leaf phases.
Further, the flow process of said definite emphasis observation area is shown in figure 10, and detailed process is: earlier with tri-leaf period image cut apart, and its connected domain is carried out mark, obtain the barycenter M of each connected domain 1..., M n, more respectively with this barycenter M i, i=1...n is the center, be that the length of side is made square with threshold value R, this square region observation area of promptly attaching most importance to is used for the leaf elemental area of this zone seedling is observed, and obtains n emphasis observation area the most at last, is designated as Ω 1..., Ω nWherein, the value of threshold value R can cover seedling length with this square region and be as the criterion to seven leaf phases, considers that simultaneously this value is also relevant with camera heights and picture size, therefore needs to set through experiment.Each parameter of present embodiment is: camera heights is that 5 meters, image size are that 309 * 609 pixels, R get 80 pixels; Shown in figure 11 is area of observation coverage synoptic diagram; Certain strain maize seedling with the intercepting regional area is an example, and square-shaped frame is an area of observation coverage scope, and round dot is represented the barycenter in this connection district;
Further, said tri-leaf period image the obtaining of unique point sum A, detailed process is: the segmentation result bianry image that will obtain in the time of will confirming the emphasis observation area carries out morphology to be handled, and the connected domain in the image is carried out refinement, obtains its skeleton.Again through Cecilia Di Ruberto; Recognition of shapes by attributed skeletal graphs, Pattern Recognition 37,21-31; Method in 2004; Extract and add up the distal point (end points) and triradius (junction points) sum of each connected domain skeleton, shown in figure 12, blockage is represented all unique points among the figure.Suppose that certain width of cloth image has n connected domain, the connected domain characteristic that statistics wherein is numbered i is counted and is EndJunctionNum i, i=1...n, then the unique point of this image adds up to Σ i = 1 n EndJunctionNum i ;
Further, confirming of said seven leaf phase judgment thresholds is specially: the difference according to the final singling time is divided into seven leaf phases arrival final singling threshold value T afterwards 1Arrive final singling threshold value T before with seven leaf phases 2, wherein seven leaf phases arrive after the final singling threshold value be tri-leaf period image unique point sum A k doubly, promptly
Figure BDA0000068871190000082
And seven leaf phases arrive before the final singling threshold value be k*M*N, wherein M is the cave number in the image in tri-leaf period, N be tri-leaf period the average characteristics of image count k ∈ [2,3].
What M was counted in the cave in the said tri-leaf period image confirms that flow process is shown in figure 13, and detailed process is: utilize the segmentation result binary map that obtains when confirming the emphasis observation area, carry out connected component labeling, suppose to have n connected domain; Obtain the barycenter M of each connected domain then 1..., M n, through range formula:
Figure BDA0000068871190000083
Calculate barycenter M i(x i, y i) and M j(x j, y j), i, the distance between the j=1...n obtains one apart from symmetry square matrix DIS, wherein the element dis in the square formation I, jPromptly represent barycenter M i(x i, y i) and M j(x j, y j) between distance; Judge the element value dis in the square formation I, jAnd the relation between the threshold alpha, if less than α, two then that its element is corresponding barycenter M iWith M jMerge into a barycenter M Ij, the coordinate of this barycenter does
Figure BDA0000068871190000084
Again this barycenter is joined in the barycenter sequence again, recomputate matrix D IS, so circulation, till the matrix element value was all greater than α, remaining at last barycenter number was the cave number.Cave and the cave distance in image was set when wherein the value of α was sowed according to reality, and present embodiment α value is 50 pixels;
Said tri-leaf period image the average characteristics points N confirm that flow process is shown in figure 14, detailed process is: at first to tri-leaf period image cut apart and the end of adding up each connected domain EndNum that counts j, j=1...n, the distal point with each connected domain is updated among the fitting function relational expression y=F in tri-leaf period (x) again, obtains the strain of each connected region and counts StemNum j, j=1...n, above method is with tri-leaf period detection-phase step a and step b; Utilize the unique point sum A of image in tri-leaf period to count sum divided by all connected domain strains again, expression formula does
Figure BDA0000068871190000091
Can obtain the average characteristics points N of image in tri-leaf period.
B. obtain next treating constantly again and observe the crop map picture, and add up the unique point sum B of this image, the total A of unique point that obtains the image in tri-leaf period among statistical method and the step a is identical;
C. judge whether image has carried out final singling; If there is not final singling, then get into steps d; If final singling has taken place, then get into step e;
The judgement of said final singling, flow process is shown in figure 15, and detailed process is: at first, the image constantly of i-1 after tri-leaf period cut apart, and the leaf elemental area of the crop seedling in the statistics emphasis observation area, be designated as
Figure BDA0000068871190000092
Take off the image of a moment i again, the leaf elemental area of the crop seedling in the same counterweight point observation zone is added up, and is designated as The leaf elemental area of above two corresponding observation areas of the moment is subtracted each other, remove the area area of carving i-1 above a period of time again, promptly J=1...n; Add up the Δ H that satisfies condition j＞T 3Connected domain number J, T 3∈ [0.3,0.6]; At last, when
Figure BDA0000068871190000095
The time, T 4∈ [0.2,0.4] then judges i final singling constantly, otherwise i also not final singling constantly.
D. if the unique point of present image sum B arrives final singling threshold value afterwards greater than seven leaf phases; Promptly
Figure BDA0000068871190000096
representes that then this image arrived for seven leaf phases; The shooting time of this image was designated as for seven time of arrival leaf phase, and seven leaf phases were judged end; Otherwise get into step b;
E. if the unique point of present image sum B arrive greater than seven leaf phases before the final singling threshold value, promptly k*M*N representes that then this image arrived for seven leaf phases, and the shooting time of this image was designated as for seven time of arrival leaf phase, seven leaf phases were judged and finish; Otherwise get into step b;
(2) seeding method when corn is bar sowing time, detects automatically according to following steps, and its flow process is shown in figure 16:
A. the emphasis observation area in definite image, the unique point sum C that obtains the image in tri-leaf period and the characteristic of each connected domain are counted;
Further, said definite emphasis observation area Ω 1..., Ω n, obtain the unique point sum of image in tri-leaf period
Figure BDA0000068871190000101
And the characteristic of each connected domain EndJunctionNum that counts i, i=1...n, all with bunch planting way flow process a in method identical;
B. obtain next treating constantly and observe the crop map picture, and the sum D that counts of the characteristic in the statistical picture emphasis observation area;
Further, the sum of counting of the characteristic in the said image emphasis observation area is meant the unique point EndJunction that obtains each connected domain in the current entire image i(x i, y i), after the i=1...k, statistics belongs to image emphasis observation area Ω again 1..., Ω nIn the characteristic sum of counting, promptly statistical nature point is numbered the number of i, wherein the i EndJunction that satisfies condition i(x i, y i) ∈ Ω j, i=1...k, j=1...n.
C. judge whether present image has carried out final singling; If there is not final singling, then get into steps d; If final singling has taken place, then get into step e;
Further, the judgement flow process of said final singling is identical, shown in figure 17 with final singling flow process among the bunch planting step c;
D. if the total characteristic in the present image emphasis observation area is counted D greater than k*C, represent that then this image arrived for seven leaf phases, the shooting time of this image was designated as for seven time of arrival leaf phase, seven leaf phases were judged end; Otherwise get into step b;
E. confirm the emphasis observation area again, flow process is shown in figure 17, and final singling takes place after, emphasis observation area Ω 1..., Ω nIn the leaf elemental area of some regional maize seedling die-off, promptly satisfy Δ H j＞T 5, wherein J=1...n, T 5∈ [0.3,0.6] therefore will satisfy the zone of above condition and reject, and remaining observation area is designated as Ω ' as new emphasis observation area.
F. obtaining in the present image the new interior characteristic of emphasis observation area Ω ' counts in sum E and the tri-leaf period image and the characteristic that the emphasis area of observation coverage the is numbered identical connected domain sum F that counts;
Further, the statistics of said unique point sum E is to carry out in the emphasis observation area Ω ' after renewal, and concrete steps are identical with step b.
Further, the statistics of unique point sum F, be through to tri-leaf period the image connectivity territory the unique point sum with
Figure BDA0000068871190000111
Calculate, wherein k ∈ { 1...n} and satisfied { k|k=i, Ω i∈ Ω ' }.
G. if the unique point of present image sum E greater than k*F, representes that then this image arrived for seven leaf phases, the shooting time of this image was designated as for seven time of arrival leaf phase, seven leaf phases were judged and finish; Otherwise get into step b.

Claims (4)

1. the automatic testing method in corn tri-leaf period is specially:
(1) to the corn map to be measured after emerging as segmented extraction n maize seedling connected region, the end of adding up each connected domain EndNum that counts j, j=1...n;
(2) count StemNum according to corn plants number and the corn plants that terminal mapping relations of counting obtain each connected domain j, said corn plants number and terminal mapping relations of counting are to utilize the historical sample graphical analysis statistics in corn tri-leaf period to obtain;
(3) the average end that calculates each connected domain is counted MeanEndNum j = EndNum j StemNum j ;
(4) the average end that calculates corn map picture to be measured is counted MeanEndNum = Σ j = 1 n MeanEndNum j n ;
(5) if MeanEndNum >=tri-leaf period judgment threshold, show that then corn has arrived tri-leaf period.
2. the automatic testing method in corn according to claim 1 tri-leaf period is characterized in that, said tri-leaf period, the span of judgment threshold was [2,3].
3. the automatic testing method of seven leaf phases of corn; Confirm that according to the automatic testing method in the described corn of claim 1 tri-leaf period corn gets into tri-leaf period; Utilize corn image in tri-leaf period to detect according to the following mode seven leaf phases of carrying out: seeding method is selected option A during for bunch planting for use, and seeding method is selected option b sowing time for use for bar;
Option A:
(A1) confirm the unique point sum and the seven leaf phase judgment thresholds of emphasis observation area, tri-leaf period image;
Said emphasis observation area is the barycenter M with n maize seedling connected region of corn image in tri-leaf period 1..., M nBe the center, R is a n square area of the length of side;
The unique point of said image adds up in the image characteristic of each connected region sum of counting, and the characteristic of said connected region is counted and counted and the trident sum of counting for the end of connected region;
Said seven leaf phase judgment thresholds comprise seven leaf phases arrival final singling threshold value T afterwards 1Arrive final singling threshold value T before with seven leaf phases 2, wherein
Figure FDA0000068871180000013
EndJunctionNum jThe characteristic that is j connected region is counted, k ∈ [2,3]; T 2=k*M*N, M are the cave number in the image in tri-leaf period, N be tri-leaf period the average characteristics of image count N = Σ j = 1 n EndJunctionNum i Σ j = 1 n StemNum j ;
(A2) the unique point sum B of statistics image to be detected, and judge the corn final singling whether in the image to be detected, if there is not final singling, then get into step (A3), otherwise, get into step (A4);
(A3) if B＞T 1, show that then the corn in the image to be detected arrived for seven leaf phases, finish;
(A4) if B＞T 2, show that then the corn in the image to be detected arrived for seven leaf phases, finish;
Said final singling is carried out according to following mode: the note corn is respectively P at i-1 and i image constantly I-1And P i, calculate
Figure FDA0000068871180000022
J=1...n,
Figure FDA0000068871180000023
Be P I-1In j emphasis observation area in the maize leaves elemental area,
Figure FDA0000068871180000024
Be P iIn j emphasis observation area in the maize leaves elemental area, statistics Δ H j＞T 3Emphasis observation area number J, if
Figure FDA0000068871180000025
Then show P iIn corn final singling, T 3∈ [0.3,0.6], T 4∈ [0.2,0.4];
Option b:
(B1) according to the mode of step (A1) confirm the emphasis observation area and tri-leaf period image unique point sum C = Σ i = 1 n EndJunctionNum i ;
(B2) the sum D that counts of the characteristic in each emphasis observation area in the statistics image to be detected, distal point and trident in the observation area sum of counting of attaching most importance to of counting of the characteristic in the said emphasis observation area;
(B3) judge corn final singling whether in the image to be detected,, then get into step (B4) if there is not final singling, otherwise, step (B5) got into;
(B4) if D＞k*C shows that then the corn in the image to be detected arrived for seven leaf phases, finish;
(B5) Δ H is satisfied in rejecting j＞T 5The emphasis observation area, the characteristic of emphasis observation area connected domain of correspondence in corn image in tri-leaf period of sum E and the reservation sum F that counts that counts of the characteristic in the emphasis observation area that statistics keeps;
(B6) if E＞k*F shows that then the corn in the image to be detected arrived for seven leaf phases, finish.
4. automatic testing method according to claim 3 is characterized in that, said cave number is confirmed according to following mode: the barycenter M that obtains each connected domain of image 1..., M n, calculate centroid distance dis I, j, i, if j=1...n is dis I, jLess than threshold alpha, then reject barycenter M iWith M j, and with M iWith M jBetween mid point be designated as barycenter, recomputate centroid distance, more relatively with itself and threshold alpha, so circulation, till all centroid distances were all greater than α, the final barycenter number that keeps was the cave number.
CN 201110162626 2011-06-16 2011-06-16 Automatic detection methods for trefoil stage and seven-leaf stage of corn Active CN102360428B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110162626 CN102360428B (en) 2011-06-16 2011-06-16 Automatic detection methods for trefoil stage and seven-leaf stage of corn

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110162626 CN102360428B (en) 2011-06-16 2011-06-16 Automatic detection methods for trefoil stage and seven-leaf stage of corn

Publications (2)

Publication Number Publication Date
CN102360428A true CN102360428A (en) 2012-02-22
CN102360428B CN102360428B (en) 2013-09-18

Family

ID=45585755

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110162626 Active CN102360428B (en) 2011-06-16 2011-06-16 Automatic detection methods for trefoil stage and seven-leaf stage of corn

Country Status (1)

Country Link
CN (1) CN102360428B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108647652A (en) * 2018-05-14 2018-10-12 北京工业大学 A kind of cotton development stage automatic identifying method based on image classification and target detection
US10186029B2 (en) 2014-09-26 2019-01-22 Wisconsin Alumni Research Foundation Object characterization
CN109345539A (en) * 2018-10-08 2019-02-15 浙江农林大学 Adaptive M ean-Shift standing tree image partition method based on image abstraction
CN111860038A (en) * 2019-04-25 2020-10-30 河南中原光电测控技术有限公司 Crop front-end recognition device and method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006320240A (en) * 2005-05-18 2006-11-30 Satake Corp Method for measuring crop information by remote sensing
CN101980249A (en) * 2010-11-12 2011-02-23 中国气象局气象探测中心 Automatic observation method and device for crop development and growth
CN102022984A (en) * 2009-09-15 2011-04-20 南开大学 Image-technique-based artificial wetland plant growth information extracting method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006320240A (en) * 2005-05-18 2006-11-30 Satake Corp Method for measuring crop information by remote sensing
CN102022984A (en) * 2009-09-15 2011-04-20 南开大学 Image-technique-based artificial wetland plant growth information extracting method
CN101980249A (en) * 2010-11-12 2011-02-23 中国气象局气象探测中心 Automatic observation method and device for crop development and growth

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10186029B2 (en) 2014-09-26 2019-01-22 Wisconsin Alumni Research Foundation Object characterization
CN108647652A (en) * 2018-05-14 2018-10-12 北京工业大学 A kind of cotton development stage automatic identifying method based on image classification and target detection
CN109345539A (en) * 2018-10-08 2019-02-15 浙江农林大学 Adaptive M ean-Shift standing tree image partition method based on image abstraction
CN109345539B (en) * 2018-10-08 2022-03-01 浙江农林大学 Self-adaptive Mean-Shift standing tree image segmentation method based on image abstraction
CN111860038A (en) * 2019-04-25 2020-10-30 河南中原光电测控技术有限公司 Crop front-end recognition device and method
CN111860038B (en) * 2019-04-25 2023-10-20 河南中原光电测控技术有限公司 Crop front end recognition device and method

Also Published As

Publication number Publication date
CN102360428B (en) 2013-09-18

Similar Documents

Publication Publication Date Title
Yu et al. Automatic image-based detection technology for two critical growth stages of maize: Emergence and three-leaf stage
Foerster et al. Crop type mapping using spectral–temporal profiles and phenological information
Southworth et al. Fragmentation of a landscape: Incorporating landscape metrics into satellite analyses of land-cover change
Hietel et al. Analysing land-cover changes in relation to environmental variables in Hesse, Germany
Zhong et al. Phenology-based crop classification algorithm and its implications on agricultural water use assessments in California’s Central Valley
CN110288647B (en) Method for monitoring irrigation area of irrigation area based on high-resolution satellite data
CN106355143A (en) Seed maize field identification method and system based on multi-source and multi-temporal high resolution remote sensing data
CN101858971A (en) Rice yield remote sensing estimation method based on MODIS data
CN102636486B (en) Method for automatically detecting three-true-leaf period of cotton
CN102592118B (en) Automatic detection method for time emergence of seedling of corns
CN106951836A (en) Crop cover degree extracting method based on priori threshold optimization convolutional neural networks
CN106971167A (en) Crop growth analysis method and its analysis system based on unmanned aerial vehicle platform
CN104424390A (en) Irrigation area monitoring method and device
CN104949981A (en) Automatic detection method and system for cotton five-euphylla period
WO2023116454A1 (en) Method and apparatus for identifying area having potential high risk of locust plagues, and device and storage medium
CN104778686A (en) Tissue culture prewarning method and system based on scab monitoring
CN102360428B (en) Automatic detection methods for trefoil stage and seven-leaf stage of corn
Kotsuki et al. SACRA–a method for the estimation of global high-resolution crop calendars from a satellite-sensed NDVI
CN113221806A (en) Cloud platform fusion multi-source satellite image and tea tree phenological period based automatic tea garden identification method
CN104951754A (en) Sophisticated crop classifying method based on combination of object oriented technology and NDVI (normalized difference vegetation index) time series
CN102542560A (en) Method for automatically detecting density of rice after transplantation
Sharma et al. Quantification and mapping of surface residue cover for maize and soybean fields in south central Nebraska
CN102663396A (en) Method for automatically detecting rice milky ripe stage
Dalposso et al. Spatial autocorrelation of NDVI and GVI indices derived from Landsat/TM images for soybean crops in the western of the state of Paraná in 2004/2005 crop season
CN116990491B (en) Automatic soil information monitoring system based on Internet of things

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant