CN106022370B - Blade MEBO ribbon gauze monitoring method and system - Google Patents

Blade MEBO ribbon gauze monitoring method and system Download PDF

Info

Publication number
CN106022370B
CN106022370B CN201610326752.6A CN201610326752A CN106022370B CN 106022370 B CN106022370 B CN 106022370B CN 201610326752 A CN201610326752 A CN 201610326752A CN 106022370 B CN106022370 B CN 106022370B
Authority
CN
China
Prior art keywords
blade
image
cluster
fluorescent image
ribbon gauze
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.)
Active
Application number
CN201610326752.6A
Other languages
Chinese (zh)
Other versions
CN106022370A (en
Inventor
李明
孙文娟
陈梅香
明楠
赵丽
杨信廷
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.)
Beijing Research Center for Information Technology in Agriculture
Original Assignee
Beijing Research Center for Information Technology in Agriculture
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 Beijing Research Center for Information Technology in Agriculture filed Critical Beijing Research Center for Information Technology in Agriculture
Priority to CN201610326752.6A priority Critical patent/CN106022370B/en
Publication of CN106022370A publication Critical patent/CN106022370A/en
Application granted granted Critical
Publication of CN106022370B publication Critical patent/CN106022370B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/62Methods or arrangements for recognition using electronic means
    • G06K9/6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
    • G06K9/6218Clustering techniques
    • G06K9/622Non-hierarchical partitioning techniques
    • G06K9/6221Non-hierarchical partitioning techniques based on statistics
    • G06K9/6223Non-hierarchical partitioning techniques based on statistics with a fixed number of clusters, e.g. K-means clustering
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/36Image preprocessing, i.e. processing the image information without deciding about the identity of the image
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/62Methods or arrangements for recognition using electronic means
    • G06K9/6267Classification techniques
    • G06K9/6268Classification techniques relating to the classification paradigm, e.g. parametric or non-parametric approaches
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/36Image preprocessing, i.e. processing the image information without deciding about the identity of the image
    • G06K2009/363Correcting image deformation, e.g. trapezoidal deformation caused by perspective

Abstract

The present invention discloses a kind of blade MEBO ribbon gauze monitoring method and system, can relatively accurately calculate the MEBO ribbon gauze of blade, and whole process is not needed according to equipment adjustment is carried out the case where vine growth and development, calculating process is relatively simple.Method includes: S1, the fluorescent image for acquiring blade;S2, cluster segmentation is carried out to the fluorescent image using K mean cluster, and binaryzation is carried out to the result of the cluster segmentation, obtain binary image;S3, the binary image is corrected in such a way that opening and closing alternately filters;S4, the discrimination standard using preset drop shape feature and size as blade whether wet, using the support vector machines based on Statistical Learning Theory as the classifier distinguished whether blade fluorescent image moistens, image after the correction is identified, the wet amount of images of blade is obtained;The MEBO ribbon gauze of S5, the amount of images and photo opporunity interval calculation blade that are moistened according to the blade.

Description

Blade MEBO ribbon gauze monitoring method and system
Technical field
The present invention relates to plant disease monitoring technical fields, and in particular to a kind of blade MEBO ribbon gauze monitoring method and is System.
Background technique
Wet blade is one of many plant leaf portion disease infestations and popular dominant factor, when this also just allows blade to moisten Between become one of very crucial environment input factor in the early warning of greenhouse vegetable disease.In recent years, when people moisten blade Between research make some progress.Currently, at present there are mainly two types of the methods of monitoring blade MEBO ribbon gauze: sensor measurement And model prediction.Wherein, the blade sensor practical application of electron type is more.Blade sensor based on resistance moistens blade Situation is converted into the numerical value of voltage or electric current, so that it is determined that a dry and wet threshold value, to count blade MEBO ribbon gauze.But such sensing Device is not possible to the characteristic of complete simulating plant tissue, it is wet not to can accurately reflect blade caused by the reasons such as guttation, and sense Device arranges complexity in practical applications, due to need keep and the blade moment contact so that need in process of crop growth through Often according to movable sensors such as leaf orientation, Leaf inclinations, inconvenience is faced in practical applications.The method of model prediction is mainly divided For mechanism type and empirical.Model prediction is that relative air humidity, evaporation, radiation, temperature and wind speed are input in model, Blade MEBO ribbon gauze is estimated with this.Theoretical model is complicated, it is desirable that the parameter of input is more, some parameters are under existence conditions It is difficult to obtain.Although the input parameter of empirical requirement is less, limited by regional and artificial subjective factor.And it is current It flourishes to Imaging-PAM, played an important role in plant disease context of detection.Wulf etc. is using wavelength The laser excitation of 337nm obtains apple and carrot in far infrared, the fluorescence spectrum of red, green, blue wave band, analyze apple and The variation of carrot freshness in storage.Wetterich etc. uses LED light to acquire the mandarin orange for suffering from yellow twig as laser light source Tangerine leaf piece fluorescent image, and by image segmentation, texture feature extraction, support vector machines realize disease blade and just It is identified when the segmentation of normal blade.Yang Hao, which is instructed etc., excites cucumber live leaves fluorescence with 4 kinds of laser intensities respectively, obtains cucumber leaves The fluorescence spectrum of 504.080~899.872nm wave band carries out feature extraction to spectroscopic data, establishes branch in conjunction with principal component analysis It holds vector machine and tests and analyzes model, which can identify symptom caused by cucumber downy mildew and aphid damage.The research such as mass troops Influence of the verticillium wilt to the Chlorophyll Fluorescence of cotton leaf, is found, the serious journey of disease by the analysis to fluorescence parameter Degree has good correlation with chlorophyll fluorescence, it was demonstrated that fluorescent technique can be used for the detection of cotton verticillium wilt.
Summary of the invention
In view of this, the present invention provides a kind of blade MEBO ribbon gauze monitoring method and system, can relatively accurately calculate The MEBO ribbon gauze of blade, and whole process is not needed according to progress equipment adjustment, calculating process the case where vine growth and development It is relatively simple.
On the one hand, the embodiment of the present invention proposes a kind of blade MEBO ribbon gauze monitoring method, comprising:
S1, the fluorescent image for acquiring blade;
S2, cluster segmentation is carried out to the fluorescent image using K mean cluster, and the result of the cluster segmentation is carried out Binaryzation obtains binary image;
S3, the binary image is corrected in such a way that opening and closing alternately filters;
S4, the discrimination standard using preset drop shape feature and size as blade whether wet, using being based on The support vector machines of Statistical Learning Theory is as the classifier distinguished whether blade fluorescent image moistens, to the figure after the correction As being identified, the wet amount of images of blade is obtained;
The MEBO ribbon gauze of S5, the amount of images and photo opporunity interval calculation blade that are moistened according to the blade.
On the other hand, the embodiment of the present invention proposes a kind of blade MEBO ribbon gauze monitoring system, comprising:
Acquisition unit, for acquiring the fluorescent image of blade;
Cutting unit, for carrying out cluster segmentation to the fluorescent image using K mean cluster, and to the cluster segmentation Result carry out binaryzation, obtain binary image;
Filter unit, for being corrected in such a way that opening and closing alternately filters to the binary image;
Recognition unit, for the differentiation mark using preset drop shape feature and size as blade whether wet Standard, using the support vector machines based on Statistical Learning Theory as the classifier distinguished whether blade fluorescent image moistens, to institute Image after stating correction is identified, the wet amount of images of blade is obtained;
Computing unit, for according to the blade moisten amount of images and photo opporunity interval calculation blade it is wet when Between.
Blade MEBO ribbon gauze monitoring method and system provided in an embodiment of the present invention, using K mean cluster, binaryzation and are opened The mode for closing alternately filtering is split the fluorescent image of blade, then by the shape feature and size of preset water droplet As characteristic variable, the wet blade of support vector machines identification is reapplied, the MEBO ribbon gauze of blade can be relatively accurately calculated, And whole process is not needed according to equipment adjustment is carried out the case where vine growth and development, and calculating process is relatively simple.
Detailed description of the invention
Fig. 1 is the flow diagram of one embodiment of blade MEBO ribbon gauze monitoring method of the present invention;
Fig. 2 is the part flow diagram of another embodiment of blade MEBO ribbon gauze monitoring method of the present invention;
Fig. 3 is the structural schematic diagram that blade MEBO ribbon gauze of the present invention monitors one embodiment of system.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention In attached drawing, technical solution in the embodiment of the present invention is explicitly described, it is clear that described embodiment is the present invention A part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
As shown in Figure 1, the present embodiment discloses a kind of blade MEBO ribbon gauze monitoring method, comprising:
S1, the fluorescent image for acquiring blade;
S2, cluster segmentation is carried out to the fluorescent image using K mean cluster, and the result of the cluster segmentation is carried out Binaryzation obtains binary image;
S3, the binary image is corrected in such a way that opening and closing alternately filters;
S4, the discrimination standard using preset drop shape feature and size as blade whether wet, using being based on The support vector machines of Statistical Learning Theory is as the classifier distinguished whether blade fluorescent image moistens, to the figure after the correction As being identified, the wet amount of images of blade is obtained;
The MEBO ribbon gauze of S5, the amount of images and photo opporunity interval calculation blade that are moistened according to the blade.
The fluorescent image of the Portable fluorescence imager acquisition blade equipped with blue light probe can be used in the present invention.Image is adopted When collection, standard probe camera lens and interlobate distance are about 7cm, and entire phosphorimager can be connected with computer, to obtain With storage blade fluorescent image.
Because of the particularity of water droplet: transparent, small in size and distribution is dispersed, the reasons such as vein interference, under natural light, The digital picture of common digital camera acquisition will accomplish that Accurate Segmentation is very difficult.And green plants blade is in photosynthesis Feux rouges, blue light are absorbed, green light (about 20%) and near infrared light (about 50%) are reflected, and then absorption reflection to red light strongly is bluish-green for water Light.Phosphorimager matches the LED light source of blue light version, and chlorophyll molecule obtains energy after absorbing blue light, and electron transition is to higher Energy level, chlorophyll fluorescence is normally at red light district, so blade formed image in phosphorimager takes on a red color;And water pair Blue light absorption is less, and transition phenomenon will not occur for electronics, therefore the water on blade is real-time glimmering after receiving illumination for a period of time Light imaging is presented blue-green.Above-mentioned characteristic enhances the comparison of target and background, is conducive to effective segmentation of image, band blue light The Imaging-PAM of LED can provide such image-pickup method.
Acquired image, the color contrast for observing its target and background is more significant, finds after the operation such as smooth, filtering glimmering The contrast of light image does not increase significantly, while also enhancing the interference of vein, therefore directly selects the fluorescence of acquisition Image carries out subsequent image segmentation.
The fluorescent image of blade belongs to RGB image, in order to preferably according to color to image clustering divide, by image from RGB color is transformed into L*a*b* color space.The advantage of L*a*b* color space is: its described non-equipment is (aobvious Show device or digital camera) particular color required for color is generated, independent of equipment itself.The appreciable color of human eye Coloured silk can be transferred through L*a*b* color space and show.In addition, the advantages of L*a*b* color space, which also resides in it, compensates for RGB The deficiency of colour model COLOR COMPOSITION THROUGH DISTRIBUTION unevenness, thus according to different color blocks come in the partitioning algorithm that is clustered, L*a* B* color space is better than RGB color.RGB color cannot be converted directly to L*a*b* color space, need first to turn XYZ color space is changed to, then is converted by XYZ color space to L*a*b* color space, conversion formula is as follows:
A*=500 [f (X/X0)-f(Y/Y0)]
B*=200 [f (Y/Y0)-f(Z/Z0)],
In above formula, X0、Y0And Z0It is corresponding with reference to white point to respectively indicate X, Y and Z,
The color difference in a*b* 2-D data space is utilized in L*a*b color space, between Euclidean distance measurement pixel Similarity, image is clustered using K mean value, the flow chart of cluster process is as shown in Figure 2.Specifically, mean cluster is calculated Steps are as follows for the realization of method:
1) given pixel size is the sample space data set of n, selects K initial cluster center at random,Serial number in its bracket finds the sequence number of the interative computation of cluster centre.
2) sample space { X } that need to classify is distributed to by minimum distance criterion to some in K cluster centre one by oneCalculate each pixel X in sample space { X }iWithEuclidean distance, and as similarity distanceWherein, i=1,2 ..., n, j=1,2 ..., K.Each sample is calculated to the distance of cluster centre, sample is returned Class to where that cluster centre nearest from it.Wherein L is the sequence number of interative computation, first time iteration L=1, SjTable Show j-th of Clustering Domain, cluster centre Zj.Clustering criteria function J after calculating this time clusterL, calculation formula is
3) mean value of all pixels of each cluster centre is calculated according to the following formula
Wherein, j=1,2 ..., K, NjFor j-th of Clustering Domain SjIncluded in pixel quantity, willGather for new Class center is clustered, and calculates the clustering criteria function J after this time cluster(L+1)
4) judge whether clustering criteria function restrains.T ∈ { 1,2 ..., K } if it exists, so thatThen return 2) step, sample space data set is reclassified one by one, repeated iterative operation;IfThen Algorithmic statement, interative computation terminate.
K-means clustering algorithm is for the color characteristic that the basic ideas of blade fluorescent image segmentation are according to image, first Initial cluster center is selected from the image data objects of given size, for there is the image of water droplet on blade, observe image 3 color blocks are broadly divided into, select the initial cluster center in 3 regions, respectively droplet region, back from image data at random Scene area one, background area two then assign them to other remaining objects according to the similarity of they and cluster centre With its most similar cluster and be marked, to generate three kinds of new cluster marked regions, then calculate new cluster areas again Cluster centre (mean values of all data objects in cluster), continue to cluster, be repeated continuously this process until cluster Criterion function convergence, operation terminate.4 color blocks, respectively droplet region, back are divided into for the fluorescent image that leaf margin has water droplet Scene area one, background area two, background area three need to select 4 initial cluster centers, other algorithm steps are constant.Finally Generate the result that different cluster marked regions is image clustering.Cluster complete after, according to the result of cluster by water droplet from Split in image, obtain cluster segmentation image, image at this time be it is colored, first to its binaryzation, then use mathematics Opening and closing alternating filtering method in morphology is corrected, and is finally completed image segmentation.
After the image that cluster segmentation generates carries out binaryzation to it, image situations such as there are burr, isolated point and cavities, Binary image is further processed using morphologic method.Structural element select size for 3 × 3 collar plate shape structural elements Element, opening operation can play the hole noise formed in removal segmentation rear region, and closed operation can remove the spot noise in image, once The effect of opening operation or closed operation processing image is undesirable, and the two can all play the effect of smoothed image, in conjunction with can obtain To more preferably target image.Therefore it selects the complex method of opening and closing operation repeatedly to filter it, it is flat to finally obtain comparison Sliding water droplet image.The mathematical expression formal definition of opening and closing operation is as follows:
The original image for enabling input is A, and structural element B, Θ are etching operation, and ⊕ is expansive working, opening operation definition Are as follows:
The effect of opening operation be usually be to eliminate small objects, separating objects and smooth larger object be not again at very thin point It substantially change its size.
Closed operation is defined as:
The effect of closed operation is minuscule hole in filler body, connection adjacent object, smooth edges.
Alternately filtering operation obtains the target image of binaryzation for K mean cluster and opening and closing, at this time if it is desired to what differentiation gave Whether blade fluorescent image moistens, and selects the shape feature and size of water droplet image, with this as blade it is wet with No discrimination standard.There is good classification to small sample based on the support vector machines of Statistical Learning Theory, selects it as differentiation Classifier whether blade fluorescent image is wet.Classifier is determined as the wet picture of blade, otherwise it is -1 that exporting, which is 1, statistics The quantity that blade moistens image is N, when phosphorimager carries out Image Acquisition, has fixed photo opporunity between adjacent picture Interval, is set as t, then the MEBO ribbon gauze of blade is represented by T=(N-1) * t, and so far algorithm terminates.
Blade MEBO ribbon gauze monitoring method provided in an embodiment of the present invention is replaced using K mean cluster, binaryzation and opening and closing The mode of filtering is split the fluorescent image of blade, then using the shape feature of preset water droplet and size as spy Variable is levied, the wet blade of support vector machines identification is reapplied, the final MEBO ribbon gauze for calculating blade can be counted relatively accurately The MEBO ribbon gauze of blade is calculated, and whole process does not need to calculate according to manpower adjustment is carried out the case where vine growth and development Journey is relatively simple.
Optionally, it in another embodiment of blade MEBO ribbon gauze monitoring method of the present invention, before the S2, also wraps It includes:
The detection for having water droplet or leaf margin to have water droplet on blade is carried out to the fluorescent image;
Wherein, the S2, comprising:
According to the testing result, cluster segmentation is carried out to the fluorescent image using K mean cluster, wherein if described There is water droplet on blade in fluorescent image, then the quantity of cluster centre used in the K mean cluster is 3, including 1 water Drip region cluster centre and 2 background areas cluster centre, if the leaf margin in the fluorescent image have water droplet, institute The quantity for stating cluster centre used in K mean cluster is 4, cluster centre and 3 background areas including 1 water droplet region Cluster centre.
As shown in figure 3, the present embodiment discloses a kind of blade MEBO ribbon gauze monitoring system, comprising:
Acquisition unit 1, for acquiring the fluorescent image of blade;
Cutting unit 2, for carrying out cluster segmentation to the fluorescent image using K mean cluster, and to the cluster point The result cut carries out binaryzation, obtains binary image;
Filter unit 3, for being corrected in such a way that opening and closing alternately filters to the binary image;
Recognition unit 4, for the differentiation mark using preset drop shape feature and size as blade whether wet Standard, using the support vector machines based on Statistical Learning Theory as the classifier distinguished whether blade fluorescent image moistens, to institute Image after stating correction is identified, the wet amount of images of blade is obtained;
Computing unit 5, amount of images and photo opporunity interval calculation blade for being moistened according to the blade it is wet Time.
Blade MEBO ribbon gauze provided in an embodiment of the present invention monitors system, is replaced using K mean cluster, binaryzation and opening and closing The mode of filtering is split the fluorescent image of blade, then using the shape feature of preset water droplet and size as spy Variable is levied, the wet blade of support vector machines identification is reapplied, can relatively accurately calculate the MEBO ribbon gauze of blade, and whole A process is not needed according to equipment adjustment is carried out the case where vine growth and development, and calculating process is relatively simple.
Although the embodiments of the invention are described in conjunction with the attached drawings, but those skilled in the art can not depart from this hair Various modifications and variations are made in the case where bright spirit and scope, such modifications and variations are each fallen within by appended claims Within limited range.

Claims (7)

1. a kind of blade MEBO ribbon gauze monitoring method characterized by comprising
S1, the fluorescent image for acquiring blade;
S2, cluster segmentation is carried out to the fluorescent image using K mean cluster, and two-value is carried out to the result of the cluster segmentation Change, obtains binary image;
S3, the binary image is corrected in such a way that the opening and closing in mathematical morphology alternately filters;
S4, the discrimination standard using preset drop shape feature and size as blade whether wet, using based on statistics The support vector machines of the theories of learning as distinguish blade fluorescent image it is wet whether classifier, to the image after the correction into Row identification obtains the wet amount of images of blade;
The MEBO ribbon gauze of S5, the amount of images and photo opporunity interval calculation blade that are moistened according to the blade.
2. blade MEBO ribbon gauze monitoring method according to claim 1, which is characterized in that before the S2, further includes:
The detection for having water droplet or leaf margin to have water droplet on blade is carried out to the fluorescent image;
Wherein, the S2, comprising:
According to the testing result, cluster segmentation is carried out to the fluorescent image using K mean cluster, wherein if the fluorescence There is water droplet on blade in image, then the quantity of cluster centre used in the K mean cluster is 3, including 1 water droplet area The cluster centre of the cluster centre in domain and 2 background areas, if the leaf margin in the fluorescent image have water droplet, the K The quantity of cluster centre used in mean cluster is 4, cluster centre including 1 water droplet region and 3 background areas Cluster centre.
3. blade MEBO ribbon gauze monitoring method according to claim 2, which is characterized in that before the S2, further includes:
The fluorescent image is transformed into L*a*b* color space from RGB color.
4. blade MEBO ribbon gauze monitoring method according to claim 1, which is characterized in that the opening operation that the S3 is carried out Expression formula be
The expression formula of the closed operation carried out isWherein, A is original image, and B is structural element, and Θ is Etching operation,For expansive working.
5. blade MEBO ribbon gauze monitoring method according to claim 1, which is characterized in that the S5, comprising:
The MEBO ribbon gauze T of blade is calculated, calculation formula is T=(N-1) * t, wherein N is the amount of images that blade moistens, and t is institute State the time interval of fluorescent image acquisition.
6. blade MEBO ribbon gauze monitoring method according to claim 1, which is characterized in that the S1, comprising:
The fluorescent image is acquired using the portable IMAGING-PAM phosphorimager equipped with blue light version MINI- probe.
7. a kind of blade MEBO ribbon gauze monitors system characterized by comprising
Acquisition unit, for acquiring the fluorescent image of blade;
Cutting unit, for carrying out cluster segmentation to the fluorescent image using K mean cluster, and to the knot of the cluster segmentation Fruit carries out binaryzation, obtains binary image;
Filter unit, for carrying out school to the binary image in such a way that the opening and closing in mathematical morphology alternately filters Just;
Recognition unit is adopted for the discrimination standard using preset drop shape feature and size as blade whether wet Use the support vector machines based on Statistical Learning Theory as the classifier distinguished whether blade fluorescent image moistens, to the correction Image afterwards is identified, the wet amount of images of blade is obtained;
Computing unit, the MEBO ribbon gauze of amount of images and photo opporunity interval calculation blade for being moistened according to the blade.
CN201610326752.6A 2016-05-17 2016-05-17 Blade MEBO ribbon gauze monitoring method and system Active CN106022370B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610326752.6A CN106022370B (en) 2016-05-17 2016-05-17 Blade MEBO ribbon gauze monitoring method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610326752.6A CN106022370B (en) 2016-05-17 2016-05-17 Blade MEBO ribbon gauze monitoring method and system

Publications (2)

Publication Number Publication Date
CN106022370A CN106022370A (en) 2016-10-12
CN106022370B true CN106022370B (en) 2019-03-29

Family

ID=57097386

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610326752.6A Active CN106022370B (en) 2016-05-17 2016-05-17 Blade MEBO ribbon gauze monitoring method and system

Country Status (1)

Country Link
CN (1) CN106022370B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108776803A (en) * 2018-04-20 2018-11-09 中国农业大学 The method and system of weeds in a kind of removal farmland

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101976350A (en) * 2010-10-20 2011-02-16 中国农业大学 Grain storage pest detection and identification method based on video analytics and system thereof
CN104598908A (en) * 2014-09-26 2015-05-06 浙江理工大学 Method for recognizing diseases of crop leaves

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101976350A (en) * 2010-10-20 2011-02-16 中国农业大学 Grain storage pest detection and identification method based on video analytics and system thereof
CN104598908A (en) * 2014-09-26 2015-05-06 浙江理工大学 Method for recognizing diseases of crop leaves

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
"A novel image edge detection method based on mathematical morphology theory";Ruize Zhong;《WIT Transactions on Information and Communication Technologies》;20141231;895-901
"The Potential of Double K‐Means Clustering for Banana Image Segmentation";Menghan Hu et al;《Journal of Food Process Engineering》;20140228;第37卷(第1期);10-18
"基于K_means硬聚类算法的葡萄病害彩色图像分割方法";李冠林 等;《农业工程学报》;20101231;第26卷(第S2期);32-37
"基于同态滤波和K均值聚类算法的杨梅图像分割";徐黎明 等;《农业工程学报》;20150731;第31卷(第14期);202-208
"基于近红外光谱技术的叶面药液浓度检测";邱白晶 等;《农业机械学报》;20120930;第43卷(第9期);197-202
"融合K-means与Ncut算法的无遮挡双重叠苹果";王丹丹 等;《农业工程学报》;20150531;第31卷(第10期);227-234

Also Published As

Publication number Publication date
CN106022370A (en) 2016-10-12

Similar Documents

Publication Publication Date Title
Iqbal et al. An automated detection and classification of citrus plant diseases using image processing techniques: A review
Barbedo A review on the main challenges in automatic plant disease identification based on visible range images
Dorj et al. An yield estimation in citrus orchards via fruit detection and counting using image processing
Barbedo Digital image processing techniques for detecting, quantifying and classifying plant diseases
Payne et al. Estimating mango crop yield using image analysis using fruit at ‘stone hardening’stage and night time imaging
CN107240091B (en) Capsule endoscope image preprocessing system and method
Sannakki et al. Diagnosis and classification of grape leaf diseases using neural networks
Yu et al. Automatic image-based detection technology for two critical growth stages of maize: Emergence and three-leaf stage
Revathi et al. Classification of cotton leaf spot diseases using image processing edge detection techniques
Bhargava et al. Fruits and vegetables quality evaluation using computer vision: A review
CN104586364B (en) A kind of skin quality detection system and method
Zhou et al. Using colour features of cv.‘Gala’apple fruits in an orchard in image processing to predict yield
US8660342B2 (en) Method to assess aesthetic quality of photographs
CN107016405B (en) A kind of pest image classification method based on classification prediction convolutional neural networks
CN106406403B (en) A kind of agriculture managing and control system based on augmented reality
CN102759528B (en) Method for detecting diseases of crop leaves
Barbedo Using digital image processing for counting whiteflies on soybean leaves
CN105147274B (en) A kind of method that heart rate is extracted in the face video signal from visible spectrum
CN101881726B (en) Nondestructive detection method for comprehensive character living bodies of plant seedlings
El-Faki et al. Weed detection using color machine vision
Pydipati et al. Identification of citrus disease using color texture features and discriminant analysis
Hamuda et al. A survey of image processing techniques for plant extraction and segmentation in the field
Okamoto et al. Green citrus detection using hyperspectral imaging
McCarthy et al. Applied machine vision of plants: a review with implications for field deployment in automated farming operations
Burks et al. Classification of weed species using color texture features and discriminant analysis

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant