CN108362698A - A kind of rice shoot stem nodal test method and device - Google Patents
A kind of rice shoot stem nodal test method and device Download PDFInfo
- Publication number
- CN108362698A CN108362698A CN201810117281.7A CN201810117281A CN108362698A CN 108362698 A CN108362698 A CN 108362698A CN 201810117281 A CN201810117281 A CN 201810117281A CN 108362698 A CN108362698 A CN 108362698A
- Authority
- CN
- China
- Prior art keywords
- skeleton
- branch
- stem
- rice shoot
- contour line
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8466—Investigation of vegetal material, e.g. leaves, plants, fruits
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of rice shoot stem nodal test method and device, including：Obtain the skeletonizing image of rice shoot to be measured；Obtain the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch；According to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, determine whether to carry out beta pruning to each skeleton branch, if judging to know that all branches of the rice shoot to be measured need to carry out beta pruning there are branch, after completing cut operator, the skeletonizing image after beta pruning is obtained, and obtains all stem nodes in the stem region in the skeletonizing image after beta pruning.The present invention removes the skeleton branch of redundancy by beta pruning method, improves the recognition efficiency of skeleton line intersection point；It is approximately straight line by stem, by the way of the linear scanning within the scope of certain angle, improves the adaptability of algorithm；By the stem nodal test of bag of words, have the function of self study, improves the accuracy of identification on the whole.
Description
Technical field
The present invention relates to agriculture intelligent equipment fields, more particularly, to a kind of rice shoot stem nodal test method and dress
It sets.
Background technology
Currently, wisdom agricultural is agricultural and merging between artificial intelligence technology and modern information technologies, and it is complete at present
In ball agricultural development there is an urgent need for technology.The acquiring technology of plant growth information is one of important technology support of wisdom agricultural, is led to
The mathematical model for establishing the scientific evaluation crop growth conditions based on Multi-information acquisition is crossed, realizes the prison to plant growth situation
It surveys, evaluation, makes it possible that crop production carries out science, intelligent management and decision using computer.
Mainly include the acquisition of apparent information and internal information, apparent letter to the acquisition of plant growth information both at home and abroad at present
Breath includes the form identification etc. of leaf area, plant height and biomass measuring, blade, and internal information is obtained by means of external means
The information of physics and chemistry, such as nutritional information monitoring, blade and canopy surface temperature, leaf water potential, chlorophyll content.
The features such as the number of sheets, leaf length, height of seedling, stem thickness and the dry matter content of tomato seedling are used as diagnosis index, according to big
Quantifier elimination is found, rice shoot panel length, very sensitive on significantly affecting for environment-stress, including water shortage, night high temperature, lacks sun
Light and excessive nitrogen, therefore can be by tomato sprout internode away from the index as reflection seedling growth situation.
In the prior art, mainly by manually under different growing environment, the fruits and vegetables rice shoot of different stages of growth carries out
Quality qualitative observation is judged, and ununified standard also be easy to cause erroneous judgement.
Invention content
The present invention provides a kind of a kind of rice shoot stem section for overcoming the above problem or solving the above problems at least partly
Point detecting method and device.
According to an aspect of the present invention, a kind of rice shoot stem nodal test method is provided, including：
S1, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the rice shoot to be measured stem skeleton,
The contour line of all skeleton branches and the rice shoot to be measured on the skeleton line of the rice shoot to be measured；
S2, each corresponding sub- contour line of skeleton branch and the corresponding profile angle of each skeleton branch are obtained；Wherein, often
The corresponding sub- contour line of one skeleton branch is one section on the contour line, and the corresponding profile angle of each skeleton branch is each
The line of two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of skeleton branch is formed by angle, often
Intersection point or each skeleton branch and institute of the corresponding node of one skeleton branch between each skeleton branch and the stem skeleton
State the intersection point between other skeleton branches of rice shoot to be measured；
S3, according to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, determination is
No to carry out beta pruning to each skeleton branch, if judging to know all branches of the rice shoot to be measured, there are branches to be cut
Branch obtains the skeletonizing image after beta pruning, and obtain the stem area in the skeletonizing image after beta pruning after completing cut operator
All stem nodes in domain.
Preferably, in step S3, according to the corresponding sub- contour line of each skeleton branch and the corresponding wheel of each skeleton branch
Wide angle, it is determined whether beta pruning is carried out to each skeleton branch, is specifically included：For any skeleton in all skeleton branches point
Branch, if the ratio of the length of the corresponding sub- contour line of any skeleton branch and the corresponding profile angle of any skeleton branch
Value is less than predetermined threshold value, and beta pruning is carried out to any skeleton branch.
Preferably, the length of the corresponding sub- contour line of any skeleton branch is：
OL(bi)=max { Om (kj,kn):kj∈T(li),kn∈T(li),
Wherein, biIndicate the corresponding node of any skeleton branch, Om (kj,kn) indicate kjAnd knBetween contour line, T
(li) indicate on the contour line with biDistance be less than pre-determined distance all the points set.
Preferably, kjAnd knBetween contour line be：
Om(kj,kn)=min (d (kj,kn),d(k1,kN)-d(kj,kn)),
Wherein, d (kj,kn) indicate kjAnd knThe distance between, k1Indicate first point on the contour line, kNIndicate institute
State the last one point, d (k on contour line1,kN) indicate k1And kNThe distance between.
Preferably, in step S3, the stem region in the skeletonizing image after beta pruning is obtained by linear scanning algorithm.
Preferably, the stem region obtained by linear scanning algorithm in the skeletonizing image after beta pruning, it is specific to wrap
It includes：
Obtain the area-of-interest in the Two-dimensional Color Image of the rice shoot to be measured；
Using the minimum point of the rice shoot stem to be measured as an endpoint of scanned straight lines, according to preset increments successively described
Change the angle of inclination of scanned straight lines in area-of-interest；
The number for obtaining target point on the corresponding scanned straight lines in each angle of inclination, according to the largest number of scannings of target point
Skeletonizing image after straight line and beta pruning determines that stem region described in the skeletonizing image after beta pruning, the target point include
All pixels point of the rice shoot to be measured in the Two-dimensional Color Image.
Preferably, in step S3, by bag of words, the institute in the stem region in the skeletonizing image after beta pruning is obtained
There is stem node.
According to another aspect of the present invention, a kind of rice shoot stem nodal test device is provided, including：
Skeletonizing module, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the seedling to be measured
The stem skeleton of seedling, the rice shoot to be measured skeleton line on all skeleton branches and the rice shoot to be measured contour line；
Branch module, for obtaining the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch
Degree；Wherein, the corresponding sub- contour line of each skeleton branch is one section on the contour line, the corresponding profile of each skeleton branch
Angle for two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of each skeleton branch line institute shape
At angle, intersection point or each bone of the corresponding node of each skeleton branch between each skeleton branch and the stem skeleton
Intersection point between frame branch and other skeleton branches of the rice shoot to be measured；
Stem node module, for according to the corresponding sub- contour line of each skeleton branch and the corresponding wheel of each skeleton branch
Wide angle, it is determined whether beta pruning is carried out to each skeleton branch, knowing that all branches of the rice shoot to be measured exist divides if judging
Branch needs to carry out beta pruning, after completing cut operator, obtains the skeletonizing image after beta pruning, and obtain the skeletonizing figure after beta pruning
All stem nodes in stem region as in.
According to a further aspect of the invention, a kind of computer program product is provided, the computer program product includes
The computer program being stored in non-transient computer readable storage medium, the computer program include program instruction, work as institute
When stating program instruction and being computer-executed, the computer is made to execute a kind of rice shoot stem nodal test method.
According to another aspect of the present invention, a kind of non-transient computer readable storage medium, the non-transient meter are provided
Calculation machine readable storage medium storing program for executing stores computer instruction, and the computer instruction makes the computer execute a kind of rice shoot stem node
Detection method.
The present invention proposes that a kind of rice shoot stem nodal test method and device is gone by the beta pruning method to skeletonizing image
In addition to the skeleton branch of redundancy, the recognition efficiency of skeleton line intersection point is improved, while according to profile line length and contour line angle
Ratio, it is determined whether carry out beta pruning, there is inherent integral property, while also having good noiseproof feature.By the road of stem
Diameter is approximately straight line, by the way of the linear scanning within the scope of certain angle, improves the adaptability of algorithm, straight line is swept
The method of retouching improves the accuracy rate for detecting each stem node.By the stem nodal test of bag of words, have the function of self study,
The accuracy of identification is improved on the whole.
Description of the drawings
Fig. 1 is a kind of flow chart of rice shoot stem nodal test method of one embodiment of the invention；
Fig. 2 is the hardware system composition signal in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
Figure；
Fig. 3 is the signal of certain section of skeletonizing image in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
Figure；
Fig. 4 is the schematic diagram of beta pruning process in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention；
Fig. 5 is the signal that stem region is chosen in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
Figure；
Fig. 6 is to obtain stem by bag of words in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
The schematic diagram of node；
Fig. 7 is a kind of structural schematic diagram of rice shoot stem nodal test device of one embodiment of the invention.
Specific implementation mode
With reference to the accompanying drawings and examples, the specific implementation mode of the present invention is described in further detail.Implement below
Example is not limited to the scope of the present invention for illustrating the present invention.
Fig. 1 is a kind of flow chart of rice shoot stem nodal test method of one embodiment of the invention, as shown in Figure 1, this method
Including：
S1, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the rice shoot to be measured stem skeleton,
The contour line of all skeleton branches and the rice shoot to be measured on the skeleton line of the rice shoot to be measured；
S2, each corresponding sub- contour line of skeleton branch and the corresponding profile angle of each skeleton branch are obtained；Wherein, often
The corresponding sub- contour line of one skeleton branch is one section on the contour line, and the corresponding profile angle of each skeleton branch is each
The line of two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of skeleton branch is formed by angle, often
Intersection point or each skeleton branch and institute of the corresponding node of one skeleton branch between each skeleton branch and the stem skeleton
State the intersection point between other skeleton branches of rice shoot to be measured；
S3, according to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, determination is
No to carry out beta pruning to each skeleton branch, if judging to know all branches of the rice shoot to be measured, there are branches to be cut
Branch obtains the skeletonizing image after beta pruning, and obtain the stem area in the skeletonizing image after beta pruning after completing cut operator
All stem nodes in domain.
The skeletonizing image of rice shoot to be measured can be obtained by skeletonization method, the skeleton in skeletonizing image be connection,
And single pixel, but due to the unstability of skeleton boundary noise influence, it is meant that due to rice shoot boundary to be measured, there are micro-
The finite part of small protrusion or recess, the skeleton corresponding to it will change a lot, it will usually redundancy skeleton occur
The phenomenon that branch.In order to overcome the noise susceptibility of skeletonizing process, necessary measure to be taken to carry out the boundary of rice shoot to be measured
Pretreatment or directly to redundancy skeleton Zhi Jinhang beta prunings, that is, removes redundancy skeleton branch.
It is therefore desirable to judge whether each skeleton branch is redundancy skeleton branch in skeletonizing image, and judgment method is as follows：
By taking one of skeleton branch as an example, the corresponding sub- profile of a certain section of conduct skeleton branch in the contour line of rice shoot to be measured
Line, two endpoints of the corresponding sub- contour line of the skeleton branch are respectively A and B, A, B node O corresponding with the branch respectively
Line is that AO, BO according to the corresponding sub- contour line of the skeleton branch and are somebody's turn to do using ∠ AOB as the corresponding profile angle of the branch
The corresponding profile angle of branch, to judge the branch whether for redundancy skeleton branch.
Wherein, node O can be the intersection point of the skeleton branch and stem skeleton, can also be the skeleton branch and other bones
The intersection point of frame branch, if the intersection point of the skeleton branch and stem skeleton, then node O is exactly stem node at this time, if should
The intersection point of skeleton branch and other skeleton branches, then node O is exactly non-master stipes point at this time.
Known according to a large amount of practical experience, the length of the corresponding sub- contour line of general stem node is larger, corresponding wheel
Wide angle is smaller；Rather than the length and profile angle of the corresponding character wheel profile of stem node do not have apparent difference, therefore with son
The length of contour line and the ratio of profile angle as reference, if ratio is less than predetermined threshold value, illustrate the corresponding skeleton of the node
Branch is redundancy skeleton branch, and beta pruning processing is carried out to it.
By the above method, each skeleton branch is judged, judges whether the skeleton branch is redundancy skeleton
Branch, if so, being cut to it.After the completion of all skeleton branches all judge, the skeletonizing image after beta pruning has just been obtained.
Stem straight line is considered in skeletonizing image to be substantially the position where stem, after determining stem straight line, then goes
Except the node apart from stem straight line farther out, it is believed that these nodes cannot function as the both candidate nodes of branching-point not on stem.It is waiting for
Survey in the skeletonizing image after the beta pruning of rice shoot, the stem of rice shoot from top to root it is connectable at approximate straight line since really
Determine stem region.
It is that straight line is improved by the way of the linear scanning within the scope of certain angle by the path proximity of stem
The adaptability of algorithm, linear scanning method improve the accuracy rate for detecting each stem node.
Finally by bag of words, both candidate nodes all in stem region are obtained, by support vector machines to all
Both candidate nodes are classified, and stem node and non-master stipes point are obtained.
The embodiment of the present invention proposes that a kind of rice shoot stem nodal test method is gone by the beta pruning method to skeletonizing image
In addition to the skeleton branch of redundancy, the recognition efficiency of skeleton line intersection point is improved, while according to profile line length and contour line angle
Ratio, it is determined whether carry out beta pruning, there is inherent integral property, while also having good noiseproof feature.By the road of stem
Diameter is approximately straight line, by the way of the linear scanning within the scope of certain angle, improves the adaptability of algorithm, straight line is swept
The method of retouching improves the accuracy rate for detecting each stem node.By the stem nodal test of bag of words, have the function of self study,
The accuracy of identification is improved on the whole.
On the basis of the above embodiments, it is preferable that in step S3, according to the corresponding sub- contour line of each skeleton branch and
The corresponding profile angle of each skeleton branch, it is determined whether beta pruning is carried out to each skeleton branch, is specifically included：
For any skeleton branch in all skeleton branches, if the length of the corresponding sub- contour line of any skeleton branch
The ratio of degree profile angle corresponding with any skeleton branch cuts any skeleton branch less than predetermined threshold value
Branch.
On the basis of the above embodiments, it is preferable that the length of the corresponding sub- contour line of any skeleton branch is：
OL(bi)=max { Om (kj,kn):kj∈T(li),kn∈T(li),
Wherein, biIndicate the corresponding node of any skeleton branch, Om (kj,kn) indicate kjAnd knBetween contour line, T
(li) indicate on the contour line with biDistance be less than pre-determined distance all the points set.
OL(bi) indicate the corresponding sub- contour line of some skeleton branch, determining the corresponding sub- contour line of the skeleton branch
When, first determine the corresponding node b of the skeleton branchi, the corresponding T (l of the skeleton branch are then determined againi), T (li) indicate profile
On line with biDistance be less than pre-determined distance all the points set, the corresponding sub- contour line OL (b of the skeleton branchi) it is set T
(li) in maximum sub- contour line.
Specifically, kjAnd knBetween contour line be：
Om(kj,kn)=min (d (kj,kn),d(k1,kN)-d(kj,kn)),
Wherein, d (kj,kn) indicate kjAnd knThe distance between, k1Indicate first point on the contour line, kNIndicate institute
State the last one point, d (k on contour line1,kN) indicate k1And kNThe distance between.
For the contour line between any 2 points, which is defined as to the smaller contour line of two endpoints composition.
On the basis of the above embodiments, it is preferable that in step S3, the skeleton after beta pruning is obtained by linear scanning algorithm
Change the stem region in image.
The embodiment of the present invention utilizes linear scanning algorithm in stem regional choice, further improves main in stem region
The accuracy of stem node selection.
Specifically, the stem region obtained by linear scanning algorithm in the skeletonizing image after beta pruning, it is specific to wrap
It includes：
According to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, it is determined whether right
Each skeleton branch carries out beta pruning, specifically includes：
For any skeleton branch in all skeleton branches, if the length of the corresponding sub- contour line of any skeleton branch
The ratio of degree profile angle corresponding with any skeleton branch cuts any skeleton branch less than predetermined threshold value
Branch.
Fig. 2 is the hardware system composition signal in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
Figure, as shown in Fig. 2, the hardware system is mainly made of hardware such as PC machine 1, PCI-1428 image pick-up cards 2, visual sensors 3,
Visual sensor 3 includes that a camera obtains rice shoot side image to be measured.
Then ossified processing is carried out to side image, obtains skeletonizing image, Fig. 3 is that a preferred embodiment of the present invention is a kind of
The schematic diagram of certain section of skeletonizing image in rice shoot stem nodal test method, as shown in figure 3, the skeleton in skeletonizing image is to connect
It is logical and single pixel, but due to the unstability of skeleton boundary noise influence, it is meant that since rice shoot boundary to be measured exists
The finite part of small protrusion or recess, the skeleton corresponding to it will change a lot, it will usually redundancy bone occur
The phenomenon that frame branch.In order to overcome the noise susceptibility of skeletonizing process, to take necessary measure to the boundary of rice shoot to be measured into
Row pre-processes or directly to redundancy skeleton Zhi Jinhang beta pruning, that is, removes redundancy skeleton branch.
It is therefore desirable to judge whether each skeleton branch is redundancy skeleton branch in skeletonizing image, and judgment method is as follows：
By taking a skeleton branch in Fig. 3 as an example, as shown in figure 3, the corresponding node b of the skeleton branchiFor the skeleton branch and other bones
The intersection point of frame branch, the corresponding node b of the skeleton branchiBelong to non-master stipes point, according to each point arrives b on contour lineiAway from
From if distance is less than pre-determined distance, by the composition set T of all the points less than pre-determined distance (li), by set T (li) in arbitrary two
The contour line of the determining maximum length of point is as node biCorresponding sub- contour line, that is, the corresponding sub- profile of the skeleton branch
Line.
Specifically, the length of the corresponding sub- contour line of the skeleton branch is：
OL(bi)=max { Om (kj,kn):kj∈T(li),kn∈T(li),
Wherein, biIndicate the described corresponding node of skeleton branch, Om (kj,kn) indicate kjAnd knBetween contour line, T
(li) indicate contour line on biDistance be less than pre-determined distance all the points set.
Process due to seeking skeleton is exactly to seek the process of image axis, and skeleton is the lines of a pixel wide.Therefore
The intersection point of branch can centainly be found with the branch as the profile corresponding to axis.It is assumed that the contour line O of image object is by N number of
Orderly point composition O={ pi:i∈[1,N]}.Wherein, contour line is taken up an official post one or two point kjAnd knBetween contour line be defined as：
Om(kj,kn)=min (d (kj,kn),d(k1,kN)-d(kj,kn)),
Wherein, d (kj,kn) indicate kjAnd knThe distance between, k1Indicate first point on contour line, kNIndicate the wheel
The last one point, d (k on profile1,kN) indicate k1And kNThe distance between.
Two endpoints of the corresponding sub- contour line of the skeleton branch are kaAnd kb, then the corresponding profile angle of skeleton branch
For：
Wherein, dis (ka,kb) it is kaPoint and kbThe air line distance of point, dis (ka,bi) it is kaPoint and biThe air line distance of point,
dis(kb,bi) it is kbPoint and biThe air line distance of point judges that the skeleton branch is that the criterion of redundancy skeleton branch is：
Wherein, δ indicates that predetermined threshold value, λ indicate ratio.
Fig. 4 is the schematic diagram of beta pruning process in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention, such as
Shown in Fig. 4, (a) indicates that the original image of a certain section of rice shoot to be measured, (b) skeletonizing image (c) indicate the skeletonizing figure after beta pruning
Picture.
From the foregoing, it will be observed that by the beta pruning method to skeletonizing image, the skeleton branch of redundancy is eliminated, skeleton line intersection point is improved
Recognition efficiency, while according to the ratio of profile line length and contour line angle, it is determined whether carry out beta pruning, there is inherent product
Divide property, while also having good noiseproof feature.
Then the stem region in the skeletonizing image after beta pruning is chosen, is illustrated by taking tomato sprout as an example, Fig. 5 is this
The schematic diagram that stem region is chosen in a kind of rice shoot stem nodal test method of a preferred embodiment is invented, as shown in figure 5, stem
Straight line is considered to be substantially the position where stem in the picture, is exactly node of the removal apart from stem straight line farther out in next step,
Think that these nodes not on stem, cannot function as the both candidate nodes of skeleton branching-point.Skeleton after the beta pruning of tomato sprout
Change in image, the stem of tomato sprout is connectable at approximate straight line from top to root, and stem region is determined with this.According to
This, uses the stem extracted region detection algorithm based on linear scanning, is as follows herein：
Since the stem of tomato sprout is influenced by accidentalia, it cannot be guaranteed that itself is vertical, often occur as schemed institute
The inclination shown, if directly causing inefficiency using linear scanning method.Therefore set forth herein first determine whether tomato sprout two dimension
The left end position of rice shoot image is extracted in the minimum point of stem in coloured image, i.e. the centre position O of least significant end diameter, then scanning
A and right end position B is set, is chosen herein by the statistics to tomato sprout angle of inclination using OC as the bisector of ∠ AOB
Line centered on by bisector OC, 25 ° of angles of symmetrical selection are scanned.The sector region that AOB is determined is that the present invention is real
The area-of-interest of example is applied, and using the minimum point of stem described in the area-of-interest as an endpoint of scanned straight lines, is pressed
Change the angle of inclination of scanned straight lines successively according to preset increments, OD, OF are certain straight line in scanning process, and OO' is to need
The stem straight line to be found.
Using O points as endpoint when scanning, the preset increments of scan line are 2 °, generate Different Slope straight line, and statistics is fallen in straight line
On target point number, finally using comprising the most straight line of target point as stem straight line.
The number for obtaining target point on the corresponding scanned straight lines in each angle of inclination, according to the scanned straight lines of target point number
Determine that stem region, target point are all the points on the skeleton of the rice shoot to be measured.
Chosen by both candidate nodes in stem region, eliminate most interfering nodes, but stem region still
It is non-master stipes point there are some nodes.By observation, it can be seen that although there is differences in the skeleton branching-point region of different rice shoot
It is different, but we still can find some common places in these skeleton branching-point regions, for example some compare in branching-point etc.
Tiny position does not observe too big difference but, the stem position of joints feature and non-stem that we can be between different rice shoots
Position of joints feature extraction comes out, and as this classification target visual vocabulary is identified, that is, passes through bag of words (full name in English：
Bag-of-words) classify.
Fig. 6 is to obtain stem by bag of words in a kind of rice shoot stem nodal test method of a preferred embodiment of the present invention
The schematic diagram of node, as shown in fig. 6, by bag of words, the flow for obtaining stem node in stem region is as follows：
Since in local invariant feature in extracting image, most widely used algorithm, the present invention are implemented SIFT algorithms
Example extracts the SIFT feature in training set per piece image using SIFT algorithms.Image in training set includes that artificial choose is led
The image of the image and non-master stipes point feature region in stipes point feature region, its SIFT feature is all extracted to every piece image.
Clustering is carried out to the SIFT feature per piece image in training set using K-Means algorithms, K is obtained and gathers
Each Cluster merging is visual vocabulary similar in the meaning of a word, constructs a word list for including K vocabulary by class.
The number that each word occurs in the picture in word list is finally counted, is tieed up to which graphical representation is become a K
Numerical value vector, generates bag of words histogram.
For the skeletonizing image after rice shoot beta pruning to be measured, the SIFT feature of all nodes in the image is extracted, use is above-mentioned
SIFT feature is indicated to become numerical value histogram vector, is classified by support vector machines, which is seen by the word in word list
Node belongs to stem node, which node belongs to non-master stipes point.
It should be noted that support vector machines (Support Vector Machine, abbreviation SVM) is Corinna
Cortes and Vapnik is equal to what nineteen ninety-five proposed first, it is showed in solving small sample, the identification of non-linear and high dimensional pattern
Go out many distinctive advantages, and Function Fitting can be promoted the use of etc. in other machines problem concerning study.In machine learning, branch
It is supervised learning model related with relevant learning algorithm to hold vector machine (SVM goes back support vector network), can analyze number
According to recognition mode, for classification and regression analysis.
Fig. 7 is a kind of structural schematic diagram of rice shoot stem nodal test device of one embodiment of the invention, as shown in fig. 7, should
Device includes：
Skeletonizing module, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the seedling to be measured
The stem skeleton of seedling, the rice shoot to be measured skeleton line on all skeleton branches and the rice shoot to be measured contour line；
Branch module, for obtaining the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch
Degree；Wherein, the corresponding sub- contour line of each skeleton branch is one section on the contour line, the corresponding profile of each skeleton branch
Angle for two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of each skeleton branch line institute shape
At angle, intersection point or each bone of the corresponding node of each skeleton branch between each skeleton branch and the stem skeleton
Intersection point between frame branch and other skeleton branches of the rice shoot to be measured；
Stem node module, for according to the corresponding sub- contour line of each skeleton branch and the corresponding wheel of each skeleton branch
Wide angle, it is determined whether beta pruning is carried out to each skeleton branch, knowing that all branches of the rice shoot to be measured exist divides if judging
Branch needs to carry out beta pruning, after completing cut operator, obtains the skeletonizing image after beta pruning, and obtain the skeletonizing figure after beta pruning
All stem nodes in stem region as in.
The specific implementation process of present apparatus embodiment is identical as the implementation procedure of above method embodiment, specifically please refers to
Embodiment of the method is stated, details are not described herein.
On the basis of the above embodiments, it is preferable that described according to any skeleton branch section in the branch module
The corresponding sub- contour line of point and the corresponding profile angle of any branch, it is determined whether beta pruning is carried out to any branch,
It specifically includes：
If the length of the corresponding sub- contour line of any skeleton branch node is corresponding with any skeleton branch node
The ratio of profile angle be less than predetermined threshold value, beta pruning is carried out to any branch.
One embodiment of the invention discloses a kind of computer program product, and the computer program product is non-temporary including being stored in
Computer program on state computer readable storage medium, the computer program include program instruction, when described program instructs
When being computer-executed, computer is able to carry out the method that above-mentioned each method embodiment is provided, such as including：Obtain seedling to be measured
The skeletonizing image of seedling, the skeletonizing image include the skeleton line of the stem skeleton of the rice shoot to be measured, the rice shoot to be measured
The contour line of upper all skeleton branches and the rice shoot to be measured；Obtain the corresponding sub- contour line of each skeleton branch and each skeleton
The corresponding profile angle of branch；Wherein, the corresponding sub- contour line of each skeleton branch is one section on the contour line, each bone
The corresponding profile angle of frame branch is that two endpoints of the corresponding sub- contour line of each skeleton branch are corresponding with each skeleton branch
The line of node be formed by angle, the corresponding node of each skeleton branch be each skeleton branch and the stem skeleton it
Between intersection point or each skeleton branch and other skeleton branches of the rice shoot to be measured between intersection point；According to each skeleton branch
Corresponding sub- contour line and the corresponding profile angle of each skeleton branch, it is determined whether beta pruning is carried out to each skeleton branch, if
Judge to know that all branches of the rice shoot to be measured need to carry out beta pruning there are branch, after completing cut operator, obtains beta pruning
Skeletonizing image afterwards, and obtain all stem nodes in the stem region in the skeletonizing image after beta pruning.
One embodiment of the invention provides a kind of non-transient computer readable storage medium, and the non-transient computer is readable to deposit
Storage media stores computer instruction, and the computer instruction makes the computer execute the side that above-mentioned each method embodiment is provided
Method, such as including：The skeletonizing image of rice shoot to be measured is obtained, the skeletonizing image includes the stem bone of the rice shoot to be measured
Frame, the rice shoot to be measured skeleton line on all skeleton branches and the rice shoot to be measured contour line；Obtain each skeleton branch
Corresponding sub- contour line and the corresponding profile angle of each skeleton branch；Wherein, the corresponding sub- contour line of each skeleton branch is
One section on the contour line, the corresponding profile angle of each skeleton branch is the two of the corresponding sub- contour line of each skeleton branch
The line of a endpoint node corresponding with each skeleton branch is formed by angle, and the corresponding node of each skeleton branch is each
Other skeleton branches of intersection point or each skeleton branch between skeleton branch and the stem skeleton and the rice shoot to be measured it
Between intersection point；According to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, it is determined whether
Beta pruning is carried out to each skeleton branch, if judging to know that all branches of the rice shoot to be measured need to carry out beta pruning there are branch,
After completing cut operator, the skeletonizing image after beta pruning is obtained, and obtain the stem region in the skeletonizing image after beta pruning
In all stem nodes.
One of ordinary skill in the art will appreciate that：Realize that all or part of step of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can be stored in a computer read/write memory medium, the program
When being executed, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes：ROM, RAM, magnetic disc or light
The various media that can store program code such as disk.
Finally, method of the invention is only preferable embodiment, is not intended to limit the scope of the present invention.It is all
Within the spirit and principles in the present invention, any modification, equivalent replacement, improvement and so on should be included in the protection of the present invention
Within the scope of.
Claims (10)
1. a kind of rice shoot stem nodal test method, which is characterized in that including：
S1, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the stem skeleton, described of the rice shoot to be measured
The contour line of all skeleton branches and the rice shoot to be measured on the skeleton line of rice shoot to be measured；
S2, each corresponding sub- contour line of skeleton branch and the corresponding profile angle of each skeleton branch are obtained；Wherein, each bone
The corresponding sub- contour line of frame branch is one section on the contour line, and the corresponding profile angle of each skeleton branch is each skeleton
The line of two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of branch is formed by angle, each bone
Intersection point or each skeleton branch of the corresponding node of frame branch between each skeleton branch and the stem skeleton are waited for described
Survey the intersection point between other skeleton branches of rice shoot；
S3, according to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch, it is determined whether it is right
Each skeleton branch carries out beta pruning, if judging to know that all branches of the rice shoot to be measured need to carry out beta pruning there are branch,
After completing cut operator, the skeletonizing image after beta pruning is obtained, and obtain in the stem region in the skeletonizing image after beta pruning
All stem nodes.
2. method according to claim 1, which is characterized in that in step S3, according to the corresponding sub- profile of each skeleton branch
Line and the corresponding profile angle of each skeleton branch, it is determined whether beta pruning is carried out to each skeleton branch, is specifically included：
For any skeleton branch in all skeleton branches, if the length of the corresponding sub- contour line of any skeleton branch with
The ratio of the corresponding profile angle of any skeleton branch is less than predetermined threshold value, and beta pruning is carried out to any skeleton branch.
3. method according to claim 2, which is characterized in that the length of the corresponding sub- contour line of any skeleton branch
For：
OL(bi)=max { Om (kj,kn):kj∈T(li),kn∈T(li),
Wherein, biIndicate the corresponding node of any skeleton branch, Om (kj,kn) indicate kjAnd knBetween contour line, T (li)
Indicate on the contour line with biDistance be less than pre-determined distance all the points set.
4. method according to claim 3, which is characterized in that kjAnd knBetween contour line be：
Om(kj,kn)=min (d (kj,kn),d(k1,kN)-d(kj,kn)),
Wherein, d (kj,kn) indicate kjAnd knThe distance between, k1Indicate first point on the contour line, kNIndicate the wheel
The last one point, d (k on profile1,kN) indicate k1And kNThe distance between.
5. method according to claim 1, which is characterized in that in step S3, after obtaining beta pruning by linear scanning algorithm
Stem region in skeletonizing image.
6. method according to claim 5, which is characterized in that described to obtain the skeletonizing after beta pruning by linear scanning algorithm
Stem region in image, specifically includes：
Obtain the area-of-interest in the Two-dimensional Color Image of the rice shoot to be measured；
It is emerging in the sense successively according to preset increments using the minimum point of the rice shoot stem to be measured as an endpoint of scanned straight lines
Change the angle of inclination of scanned straight lines in interesting region；
The number for obtaining target point on the corresponding scanned straight lines in each angle of inclination, according to the largest number of scanned straight lines of target point
With the skeletonizing image after beta pruning, determine that stem region described in the skeletonizing image after beta pruning, the target point include described
All pixels point of the rice shoot to be measured in the Two-dimensional Color Image.
7. method according to claim 1, which is characterized in that in step S3, by bag of words, obtain the skeleton after beta pruning
Change all stem nodes in the stem region in image.
8. a kind of rice shoot stem nodal test device, which is characterized in that including：
Skeletonizing module, the skeletonizing image for obtaining rice shoot to be measured, the skeletonizing image include the rice shoot to be measured
Stem skeleton, the rice shoot to be measured skeleton line on all skeleton branches and the rice shoot to be measured contour line；
Branch module, for obtaining the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch；
Wherein, the corresponding sub- contour line of each skeleton branch is one section on the contour line, the corresponding profile angle of each skeleton branch
Degree is formed by the line of two endpoints node corresponding with each skeleton branch of the corresponding sub- contour line of each skeleton branch
Angle, intersection point or each skeleton of the corresponding node of each skeleton branch between each skeleton branch and the stem skeleton
Intersection point between branch and other skeleton branches of the rice shoot to be measured；
Stem node module, for according to the corresponding sub- contour line of each skeleton branch and the corresponding profile angle of each skeleton branch
Degree, it is determined whether beta pruning is carried out to each skeleton branch, there are branches to need if judging to know all branches of the rice shoot to be measured
Beta pruning is carried out, after completing cut operator, obtains the skeletonizing image after beta pruning, and obtain in the skeletonizing image after beta pruning
Stem region in all stem nodes.
9. a kind of computer program product, which is characterized in that the computer program product includes being stored in non-transient computer
Computer program on readable storage medium storing program for executing, the computer program include program instruction, when described program is instructed by computer
When execution, the computer is made to execute the method as described in claim 1 to 7 is any.
10. a kind of non-transient computer readable storage medium, which is characterized in that the non-transient computer readable storage medium is deposited
Computer instruction is stored up, the computer instruction makes the computer execute the method as described in claim 1 to 7 is any.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810117281.7A CN108362698B (en) | 2018-02-06 | 2018-02-06 | Method and device for detecting main stem nodes of seedlings |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810117281.7A CN108362698B (en) | 2018-02-06 | 2018-02-06 | Method and device for detecting main stem nodes of seedlings |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108362698A true CN108362698A (en) | 2018-08-03 |
| CN108362698B CN108362698B (en) | 2020-08-11 |
Family
ID=63004685
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810117281.7A Active CN108362698B (en) | 2018-02-06 | 2018-02-06 | Method and device for detecting main stem nodes of seedlings |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108362698B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102590772B1 (en) * | 2021-06-17 | 2023-10-19 | 라온피플 주식회사 | Apparatus and method for measuring length of leaf stem |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001048681A2 (en) * | 1999-12-23 | 2001-07-05 | National University Of Singapore | Automated fingerprint identification system |
| CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
| US20030235344A1 (en) * | 2002-06-15 | 2003-12-25 | Kang Sing Bing | System and method deghosting mosaics using multiperspective plane sweep |
| CN101853524A (en) * | 2010-05-13 | 2010-10-06 | 北京农业信息技术研究中心 | Method for generating corn ear panoramic image by using image sequence |
| JP4665191B2 (en) * | 2005-03-19 | 2011-04-06 | 康二 三宅 | Thinning method of binary image |
| CN103337092A (en) * | 2013-06-05 | 2013-10-02 | 北京农业信息技术研究中心 | An extraction method for a fruit tree limb skeleton |
| CN103729621A (en) * | 2013-12-20 | 2014-04-16 | 华南农业大学 | Plant leaf image automatic recognition method based on leaf skeleton model |
| CN104156997A (en) * | 2014-07-28 | 2014-11-19 | 北京航空航天大学 | Quick volume data skeleton extraction method based on rendering |
-
2018
- 2018-02-06 CN CN201810117281.7A patent/CN108362698B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001048681A2 (en) * | 1999-12-23 | 2001-07-05 | National University Of Singapore | Automated fingerprint identification system |
| CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
| US20030235344A1 (en) * | 2002-06-15 | 2003-12-25 | Kang Sing Bing | System and method deghosting mosaics using multiperspective plane sweep |
| JP4665191B2 (en) * | 2005-03-19 | 2011-04-06 | 康二 三宅 | Thinning method of binary image |
| CN101853524A (en) * | 2010-05-13 | 2010-10-06 | 北京农业信息技术研究中心 | Method for generating corn ear panoramic image by using image sequence |
| CN103337092A (en) * | 2013-06-05 | 2013-10-02 | 北京农业信息技术研究中心 | An extraction method for a fruit tree limb skeleton |
| CN103729621A (en) * | 2013-12-20 | 2014-04-16 | 华南农业大学 | Plant leaf image automatic recognition method based on leaf skeleton model |
| CN104156997A (en) * | 2014-07-28 | 2014-11-19 | 北京航空航天大学 | Quick volume data skeleton extraction method based on rendering |
Non-Patent Citations (8)
| Title |
|---|
| CHANG,F: "Feature Analysis Using Line Sweep Thinning Algorithm", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
| LI，CUILING: "Using hyperspectral imaging technology to identify diseased tomato leaves", 《INFRARED, MILLIMETER-WAVE, AND TERAHERTZ TECHNOLOGIES IV》 * |
| XUE-GUAN ZHAO: "Research on Node Detection Algorithm of Tomato Seedlings Based on Digital Image", 《2017 2ND INTERNATIONAL CONFERENCE ON COMPUTER, MECHATRONICS AND ELECTRONIC ENGINEERING》 * |
| YAMAMOTO,K: "Node Detection and Internode Length Estimation of Tomato Seedlings Based on Image Analysis and Machine Learning", 《SENSORS》 * |
| YAMAMOTO,K: "On Plant Detection of Intact Tomato Fruits Using Image Analysis and Machine Learning Methods", 《SENSORS》 * |
| 张国栋: "基于模糊距离变换的骨架剪枝算法", 《沈阳航空航天大学学报》 * |
| 张祯伟: "改进视觉词袋模型的快速图像检索方法", 《计算机系统应用》 * |
| 陈晓光: "应用图象处理技术进行蔬菜苗特征量识别", 《农业工程学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108362698B (en) | 2020-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Apolo-Apolo et al. | Deep learning techniques for estimation of the yield and size of citrus fruits using a UAV | |
| Bao et al. | Field-based architectural traits characterisation of maize plant using time-of-flight 3D imaging | |
| Song et al. | Kiwifruit detection in field images using Faster R-CNN with VGG16 | |
| CN113392775B (en) | Sugarcane seedling automatic identification and counting method based on deep neural network | |
| Aquino et al. | A new methodology for estimating the grapevine-berry number per cluster using image analysis | |
| CN109146948B (en) | Crop growth phenotype parameter quantification and yield correlation analysis method based on vision | |
| Liu et al. | Automated image-processing for counting seedlings in a wheat field | |
| Wosner et al. | Object detection in agricultural contexts: A multiple resolution benchmark and comparison to human | |
| Wang et al. | Deep learning approach for apple edge detection to remotely monitor apple growth in orchards | |
| Bosilj et al. | Analysis of morphology-based features for classification of crop and weeds in precision agriculture | |
| CN109684938A (en) | It is a kind of to be taken photo by plane the sugarcane strain number automatic identifying method of top view based on crop canopies | |
| CN108052886A (en) | A kind of puccinia striiformis uredospore programming count method of counting | |
| CN111160451A (en) | Flexible material detection method and storage medium thereof | |
| CN111860571B (en) | Cloud microparticle classification method based on CIP data quality control | |
| Syal et al. | A survey of computer vision methods for counting fruits and yield prediction | |
| Mirbod et al. | Automated measurement of berry size in images | |
| Zhang et al. | Automatic identification algorithm of the rice tiller period based on PCA and SVM | |
| CN110188657A (en) | Corn arid recognition methods based on crimping blade detection | |
| Boatswain Jacques et al. | Towards a machine vision-based yield monitor for the counting and quality mapping of shallots | |
| CN108362698A (en) | A kind of rice shoot stem nodal test method and device | |
| CN112488230A (en) | Crop water stress degree judging method and device based on machine learning | |
| Beguet et al. | Retrieving forest structure variables from very high resolution satellite images using an automatic method | |
| JP4270254B2 (en) | Image signal processing apparatus and image processing method | |
| Auleria et al. | A review on KN earest neighbour based classification for object recognition | |
| Khan et al. | Vision based classification of fresh fruits using fuzzy logic |
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 |