CN107154049B - Vegetable leaf scab segmentation method based on color information and server - Google Patents

Vegetable leaf scab segmentation method based on color information and server Download PDF

Info

Publication number
CN107154049B
CN107154049B CN201710358725.1A CN201710358725A CN107154049B CN 107154049 B CN107154049 B CN 107154049B CN 201710358725 A CN201710358725 A CN 201710358725A CN 107154049 B CN107154049 B CN 107154049B
Authority
CN
China
Prior art keywords
color
scab
image
leaf
color space
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
CN201710358725.1A
Other languages
Chinese (zh)
Other versions
CN107154049A (en
Inventor
马浚诚
孙忠富
杜克明
褚金翔
郑飞翔
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.)
Institute of Environment and Sustainable Development in Agriculturem of CAAS
Original Assignee
Institute of Environment and Sustainable Development in Agriculturem of CAAS
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 Institute of Environment and Sustainable Development in Agriculturem of CAAS filed Critical Institute of Environment and Sustainable Development in Agriculturem of CAAS
Priority to CN201710358725.1A priority Critical patent/CN107154049B/en
Publication of CN107154049A publication Critical patent/CN107154049A/en
Application granted granted Critical
Publication of CN107154049B publication Critical patent/CN107154049B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/11Region-based segmentation
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/90Determination of colour characteristics
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10004Still image; Photographic image
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30181Earth observation
    • G06T2207/30188Vegetation; Agriculture

Abstract

The invention discloses a vegetable leaf scab segmentation method based on color information and a server. Wherein the method comprises the following steps: acquiring the scab leaf images of the vegetables, and preprocessing the scab leaf images; respectively extracting the color characteristics of the preprocessed scab leaf images in different color spaces; constructing comprehensive color feature detection filtering according to the color features of the different color spaces; detecting and filtering according to the comprehensive color features to obtain a comprehensive color feature map corresponding to the scab leaf image; and segmenting the scab in the scab leaf image according to the comprehensive color feature map. The server is used for executing the method. The vegetable leaf scab segmentation method and the server based on the color information improve the accuracy of scab segmentation in the scab leaf image.

Description

Vegetable leaf scab segmentation method based on color information and server
Technical Field
The invention relates to the technical field of image processing, in particular to a vegetable leaf scab segmentation method based on color information and a server.
Background
In greenhouse vegetable planting, diseases are one of main factors causing the reduction of vegetable quality, and the method accurately identifies and diagnoses the diseases and has important significance for greenhouse vegetable planting.
The key for accurately identifying the diseases is to accurately acquire the disease spot information, and the method for extracting the disease spot image information from the leaf images acquired in the field based on the image segmentation method is one of the most main ways for acquiring the disease information at present. The color information is the most direct information for distinguishing diseased leaf from normal leaf, but the color information is very easily influenced by the illumination condition. In the field, the image collected under the actual condition is interfered by the background, uneven illumination and the like of image collection, so that the identifiability of the scab in the collected blade image is reduced, and the difficulty in judging the scab image is caused.
Therefore, it is an urgent need in the art to provide a method for processing an acquired lesion leaf image under the conditions of non-uniform illumination conditions, complex leaf image acquisition background, and the like, so as to improve the accuracy of lesion segmentation in the lesion leaf image.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a vegetable leaf scab segmentation method based on color information and a server.
On one hand, the invention provides a vegetable leaf scab segmentation method based on color information, which comprises the following steps:
acquiring the scab leaf images of the vegetables, and preprocessing the scab leaf images;
respectively extracting the color characteristics of the preprocessed scab leaf images in different color spaces;
constructing comprehensive color feature detection filtering according to the color features of the different color spaces;
detecting and filtering according to the comprehensive color features to obtain a comprehensive color feature map corresponding to the scab leaf image;
and segmenting the scab in the scab leaf image according to the comprehensive color feature map.
In another aspect, the present invention provides a server, comprising:
the image preprocessing module is used for acquiring scab leaf images of vegetables and preprocessing the scab leaf images;
the characteristic extraction module is used for respectively extracting the color characteristics of the preprocessed scab blade images in different color spaces;
the filtering construction module is used for constructing comprehensive color characteristic detection filtering according to the color characteristics of the different color spaces;
the characteristic diagram obtaining module is used for detecting and filtering according to the comprehensive color characteristic to obtain a comprehensive color characteristic diagram corresponding to the scab leaf image;
and the scab segmentation module is used for segmenting the scab in the scab leaf image according to the comprehensive color feature map.
According to the vegetable leaf scab segmentation method and the server based on the color information, the acquired scab leaf image is preprocessed, the color features of the preprocessed scab leaf image in different color spaces are extracted, the comprehensive color feature detection filter is constructed, the comprehensive color feature image corresponding to the scab leaf image is obtained, the scab in the scab leaf image is segmented, and the accuracy of the scab segmentation in the scab leaf image is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for segmenting vegetable leaf spots based on color information according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating the relationship between ExR parameters and CCF ratio in a method for dividing vegetable leaf lesions based on color information according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a server according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a method for segmenting vegetable leaf spots based on color information according to an embodiment of the present invention, and as shown in fig. 1, the method for segmenting vegetable leaf spots based on color information according to the present invention includes:
s101, obtaining an image of a scab leaf of a vegetable, and preprocessing the image of the scab leaf;
specifically, a server may acquire an image of a lesion leaf of vegetables planted in a greenhouse through a camera, and then perform preprocessing on the acquired image of the lesion leaf, where the preprocessing includes image denoising and adjusting the size of the image of the lesion leaf, for example, to 800 × 600 pixels. Wherein, the scab blades are blades with scabs on the surfaces.
S102, respectively extracting color features of the preprocessed scab leaf images in different color spaces;
specifically, the server performs color feature extraction on the preprocessed scab leaf images in different color spaces. For example, when lighting conditions are relatively uniform, color features such as an extra Green feature (ExG) and an extra Red feature (ExR) in the RGB color space, the H component in the HSV color space and the b component in the la b color space can both distinguish lesions from normal leaves. However, ExG and ExR are susceptible to uneven lighting conditions, and the H component in the HSV color space and the b component in the la b color space can maintain the ability to distinguish between lesions and normal leaves. The server may extract the pre-processed hyper-red features of the lesion leaf image in the RGB color space, the H component in the HSV color space, and the b component in the L a b color space.
S103, constructing comprehensive color feature detection filtering according to the color features of the different color spaces;
specifically, the server constructs a comprehensive color feature detection filter based on the color features of the different color spaces. The comprehensive color feature detection filtering constructed by the invention comprises Gaussian Difference filtering (DoG) and circular region mean filtering, and ExR parameters can be introduced into the comprehensive color feature detection filtering because the RGB color space is easily influenced by illumination conditions, so that the influence of the uneven illumination conditions on lesion segmentation quality is reduced.
For example, based on the hyper-red feature of the preprocessed lesion leaf image in RGB color space, H component of HSV color space and b component in L a b color space, the integrated color feature detection rate wave is represented by a formula Constructing, wherein I is the preprocessed blade image, IExRSuper red feature image for RGB color space, Ib*B component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)H,σL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation.
S104, detecting and filtering according to the comprehensive color feature to obtain a comprehensive color feature map corresponding to the scab leaf image;
specifically, the server detects and filters according to the constructed comprehensive color feature to obtain a comprehensive color feature map corresponding to the scab leaf image.
For example, based on the hyper-red features of the preprocessed lesion leaf image in the RGB color space, the H component of the HSV color space, and the b component of the L a b color space, the integrated color feature map is according to the formula: CCF ═ exp (- β | f (I: r, σ)H,σL,αR) |) where CCF represents the integrated color profile, β is a drop rate parameter, αRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formulaIs obtained by calculation, M isNumber of pixels until light influences leaf area, N is number of pixels of normal leaf area, CCF (x)i,yi) And CCF (x)j,yj) And i and j are positive integers, wherein the CCF values of the pixel points on the comprehensive color feature image are represented by the I and the J. Wherein M and N can be obtained by manual labeling; it will be appreciated that the value of α is considered to be (0, 1)]In order to rapidly acquire αRA value of (0, 1)]Uniformly selecting a preset number of alpha values to calculate the R values of the preset number, selecting the minimum R value from the R values of the preset number, wherein the alpha value corresponding to the minimum R value is alphaR. Fig. 2 is a relationship diagram between ExR parameters and CCF ratio of a vegetable leaf scab segmentation method based on color information according to an embodiment of the present invention, and as shown in fig. 2, a CCF ratio R indicates how close a CCF value of an affected leaf area under an illumination condition is to a CCF value of a normal leaf area, and a smaller value of R indicates that the closer the CCF value of the affected leaf area under the illumination condition is to the CCF value of the normal leaf area, the smaller the influence of the illumination condition is. The preset number may be set according to an actual situation, and the embodiment of the present invention is not limited.
And S105, segmenting the scab in the scab leaf image according to the comprehensive color feature map.
Specifically, the server selects a growth position of an initial seed point in the comprehensive characteristic map according to the obtained comprehensive color characteristic map, and segments the scab in the scab leaf image by adopting a region growth method. Wherein the growth position of the initial seed point in the comprehensive characteristic map can be calibrated manually; in the process of segmenting the scab in the scab leaf image, a combined morphology operation can be adopted to optimize the result of the scab segmentation, wherein the combined morphology operation is that Wherein the content of the first and second substances,in order to do the operation of the dilation,for the erosion operator, bw is a binary image, S1And S2Is a structural element in morphology.
According to the vegetable leaf scab segmentation method based on the color information, the acquired scab leaf image is preprocessed, the color features of the preprocessed scab leaf image in different color spaces are extracted, comprehensive color feature detection filtering is constructed, a comprehensive color feature image corresponding to the scab leaf image is obtained, scabs in the scab leaf image are segmented, and therefore the accuracy of scab segmentation in the scab leaf image is improved.
On the basis of the foregoing embodiments, further, the separately extracting color features of the preprocessed lesion leaf images in different color spaces includes:
and respectively extracting the super red features of the preprocessed scab leaf images in an RGB color space, the H component in an HSV color space and the b component in an L, a, b and color space.
Specifically, when lighting conditions are relatively uniform, both the super-green and super-red features in the RGB color space, the H component in the HSV color space and the b component in the la b color space can achieve the distinction of the lesion leaf from the normal leaf. However, ExG and ExR are susceptible to uneven lighting conditions, and the H component in the HSV color space and the b component in the la b color space can maintain the ability to distinguish diseased from normal leaves. The server may extract ExR in RGB color space, H component in HSV color space and b component in L a b color space of the preprocessed lesion leaf image. ExR according to formula IExRExtracting (x, y) ═ 1.3R (x, y) -G (x, y), where (x, y) is pixel coordinates, and R (x, y), G (x, y) are color component values of RGB color space (x, y); the H component in HSV color space is according to formula H(x,y)=IH(x,y) Extracting, wherein (x, y) is pixel coordinate, and IH(x, y) is the value of the H component of (x, y) in the HSV color space; b component in L a b color space according to formula b(x,y)=Ib*(x, y) extraction, wherein (x, y) is pixel coordinate, Ib*(x, y) is the value of the b component of (x, y) in the color space.
On the basis of the foregoing embodiments, further, the constructing a comprehensive color feature detection filter according to the color features of the different color spaces includes:
according to the formula Constructing comprehensive color feature detection filtering, wherein I is the preprocessed blade image, and IExRSuper red feature image for RGB color space, Ib*B component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)H,σL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation.
Specifically, the server is used for processing the hyper-red characteristic of the lesion leaf image in the RGB color space, the H component of the HSV color space and the b component in the L A B color space according to a formula Constructing the comprehensive color characteristic detection rate wave, wherein I is the preprocessed blade image and IExRIs RGB color spaceSuper red characteristic image of cells, Ib*B component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)H,σL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation. Wherein a single pixel point in the preprocessed leaf image can be represented as I (x, y), and the integrated color feature detection filtering of the single pixel point can be represented as I (x, y) Wherein, IExR(x, y) is a hyper red feature component value of (x, y) of the RGB color space.
On the basis of the foregoing embodiments, further, the obtaining a comprehensive color feature map corresponding to the lesion leaf image according to the comprehensive color feature detection filtering includes:
according to the formula: CCF ═ exp (- β | f (I: r, σ)H,σL,αR) |) obtaining the integrated color feature map, wherein CCF represents the integrated color feature map, β is a descent rate parameter, and αRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formula Obtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) And i and j are positive integers, wherein the CCF values of the pixel points on the comprehensive color feature image are represented by the I and the J.
Specifically, the server, based on the hyper-red feature of the preprocessed lesion leaf image in the RGB color space, the H component of the HSV color space, and the b component in the L × a × b color space, according to a formula: CCF ═ exp (- β | f (I: r, σ)H,σL,αR) |) obtaining the comprehensive color feature map, wherein CCF represents the comprehensive color feature map, β is a descent rate parameter, and can be obtained through experiments, and α) isRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formulaObtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) Is the CCF value, CCF (x), of a pixel point on the integrated color feature mapi,yi)=exp(-β|f(I(xi,yi):r,σH,σL,αR)|),CCF(xj,yj)=exp(-β|f(I(xj,yj):r,σH,σL,αR) I) and j are positive integers. Wherein M and N can be obtained by manual labeling; it will be appreciated that the value of α is considered to be (0, 1)]In order to rapidly acquire αRA value of (0, 1)]Uniformly selecting a preset number of alpha values to calculate the R values of the preset number, selecting the minimum R value from the R values of the preset number, wherein the alpha value corresponding to the minimum R value is alphaR,. Fig. 2 is a relationship diagram between ExR parameters and CCF ratio of a vegetable leaf scab segmentation method based on color information according to an embodiment of the present invention, and as shown in fig. 2, a CCF ratio R indicates how close a CCF value of an affected leaf area under an illumination condition is to a CCF value of a normal leaf area, and a smaller value of R indicates that the closer the CCF value of the affected leaf area under the illumination condition is to the CCF value of the normal leaf area, the smaller the influence of the illumination condition is. The preset number may be set according to an actual situation, and the embodiment of the present invention is not limited.
In addition to the foregoing embodiments, the segmenting the lesion in the lesion leaf image according to the integrated color feature map further includes:
acquiring the growth position of an initial seed point in the comprehensive color characteristic diagram;
and according to the growth position of the initial seed point, segmenting the scab in the scab leaf image by adopting a region growth method.
Specifically, the server obtains the growth positions of initial seed points in the integrated color feature map, where there may be a plurality of growth positions of the initial seed points; wherein, the growth position of the seed point can be calibrated manually. And the server divides the scab in the scab leaf image by adopting a region growing method according to the growing position of the initial seed point.
Fig. 3 is a schematic structural diagram of a server according to an embodiment of the present invention, and as shown in fig. 3, the server provided in the present invention includes an image preprocessing module 301, a feature extraction module 302, a filtering construction module 303, a feature map obtaining module 304, and a lesion segmentation module 305, where:
the image preprocessing module 301 is configured to acquire an image of a scab leaf of a vegetable and preprocess the image; the feature extraction module 302 is configured to respectively extract color features of the preprocessed scab blade images in different color spaces; the filtering construction module 303 is configured to construct a comprehensive color feature detection filter according to the color features of the different color spaces; the feature map obtaining module 304 is configured to detect and filter according to the comprehensive color feature to obtain a comprehensive color feature map corresponding to the lesion leaf image; the lesion segmentation module 305 is configured to segment a lesion in the lesion leaf image according to the comprehensive color feature map.
Specifically, the image preprocessing module 301 may acquire an image of a lesion leaf of a vegetable planted in a greenhouse by using a camera, and then preprocess the acquired image of the lesion leaf, where the preprocessing includes image denoising and adjusting the size of the image of the lesion leaf, for example, to 800 × 600 pixels. Wherein, the scab blades are blades with scabs on the surfaces.
The feature extraction module 302 performs color feature extraction on the preprocessed scab blade image in different color spaces. For example, when lighting conditions are relatively uniform, color features such as super-green and super-red features in the RGB color space, the H component in the HSV color space and the b component in the la b color space can both effect differentiation of the lesion leaf from the normal leaf. However, ExG and ExR are susceptible to uneven lighting conditions, and the H component in the HSV color space and the b component in the la b color space can maintain the ability to distinguish diseased from normal leaves. The server may extract the pre-processed hyper-red features of the lesion leaf image in the RGB color space, the H component in the HSV color space, and the b component in the L a b color space.
The filter construction module 303 constructs a comprehensive color feature detection filter based on the color features of the different color spaces. The comprehensive color feature detection filtering constructed by the invention comprises Gaussian difference filtering and circular area mean filtering, and ExR parameters can be introduced into the comprehensive color feature detection filtering because the RGB color space is easily influenced by illumination conditions, so that the influence of the uneven illumination conditions on the lesion segmentation quality is reduced.
The feature map obtaining module 304 obtains a comprehensive color feature map corresponding to the lesion leaf image according to the constructed comprehensive color feature detection filtering.
The lesion segmentation module 305 selects a growth position of the initial seed point in the comprehensive characteristic map according to the obtained comprehensive color characteristic map, and segments the lesion in the lesion leaf image by using a region growth method. Wherein the growth position of the initial seed point in the comprehensive characteristic map can be calibrated manually; in the process of segmenting the scab in the scab leaf image, a combined morphology operation can be adopted to optimize the result of the scab segmentation, wherein the combined morphology operation is that Wherein the content of the first and second substances,in order to do the operation of the dilation,for the erosion operator, bw is a binary image, S1And S2Is a structural element in morphology.
According to the server provided by the invention, the acquired scab blade image is preprocessed, the color features of the preprocessed scab blade image in different color spaces are extracted, the comprehensive color feature detection filter is constructed, the comprehensive color feature image corresponding to the scab blade image is obtained, the scab in the scab blade image is segmented, and the accuracy of the scab segmentation in the scab blade image is improved.
On the basis of the foregoing embodiment, further, the feature extraction module 302 is specifically configured to:
and respectively extracting the super red features of the preprocessed scab leaf images in an RGB color space, the H component of the HSV color space and the b component of the preprocessed scab leaf images in an L, a, b and b color space.
Specifically, when lighting conditions are relatively uniform, both the super-green and super-red features in the RGB color space, the H component in the HSV color space and the b component in the la b color space can achieve the distinction of the lesion leaf from the normal leaf. However, ExG and ExR are susceptible to uneven lighting conditions, and the H component in the HSV color space and the b component in the la b color space can maintain the ability to distinguish diseased from normal leaves. The feature extraction module 302 may extract ExR in RGB color space, H component in HSV color space, and b component in L a b color space of the preprocessed lesion leaf image. ExR according to formula IExR(x, y) ═ 1.3R (x, y) -G (x, y) extracts, where (x,y) are pixel coordinates, R (x, y), G (x, y) are color component values of the RGB color space (x, y); the H component in HSV color space is according to formula H(x,y)=IH(x, y) extraction, wherein (x, y) is pixel coordinate, IH(x, y) is the value of the H component of (x, y) in the HSV color space; b component in L a b color space according to formula b(x,y)=Ib*(x, y) extraction, wherein (x, y) is pixel coordinate, Ib*(x, y) is the value of the b component of (x, y) in the color space.
On the basis of the foregoing embodiments, further, the filtering construction module 303 is specifically configured to:
according to the formula f (I: r, σ)H,σL,α)=α(pb(r)*IExR)+DoG(σH,σL)*IH+pb(r)*Ib*Constructing comprehensive color feature detection filtering, wherein I is the preprocessed blade image, and IExRSuper red feature image for RGB color space, Ib*B component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation ofAlpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation.
Specifically, the filtering construction module 303 performs filtering on the preprocessed lesion leaf image according to a formula based on the hyper-red feature in the RGB color space, the H component in the HSV color space, and the b component in the la b color space Constructing the healdCombining color characteristic detection rate waves, wherein I is the preprocessed blade image, and IExRSuper red feature image for RGB color space, Ib*B component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)H,σL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation. Wherein a single pixel point in the preprocessed leaf image can be represented as I (x, y), and the integrated color feature detection filtering of the single pixel point can be represented as I (x, y)IH(x,y)+pb(r)*Ib*(x, y) wherein IExR(x, y) is a hyper red feature component value of (x, y) of the RGB color space.
On the basis of the foregoing embodiments, further, the feature map obtaining module 304 is specifically configured to:
according to the formula: CCF ═ exp (- β | f (I: r, σ)H,σL,αR) |) obtaining the integrated color feature map, wherein CCF represents the integrated color feature map, β is a descent rate parameter, and αRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formula Obtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) And i and j are positive integers, wherein the CCF values of the pixel points on the comprehensive color feature image are represented by the I and the J.
In particular toThe feature map obtaining module 304 is configured to, based on the pre-processed hyper-red features of the lesion leaf image in the RGB color space, the H component in the HSV color space, and the b component in the la b color space, according to the formula: CCF ═ exp (- β | f (I: r, σ)H,σL,αR) |) obtaining the comprehensive color feature map, wherein CCF represents the comprehensive color feature map, β is a descent rate parameter, and can be obtained through experiments, and α) isRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formulaObtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) Is the CCF value, CCF (x), of a pixel point on the integrated color feature mapi,yi)=exp(-β|f(I(xi,yi):r,σH,σL,αR)|),CCF(xj,yj)=exp(-β|f(I(xj,yj):r,σH,σL,αR) I) and j are positive integers. Wherein M and N can be obtained by manual labeling; it will be appreciated that the value of α is considered to be (0, 1)]In order to rapidly acquire αRA value of (0, 1)]Uniformly selecting a preset number of alpha values to calculate the R values of the preset number, selecting the minimum R value from the R values of the preset number, wherein the alpha value corresponding to the minimum R value is alphaR. Fig. 2 is a relationship diagram between ExR parameters and CCF ratio of a vegetable leaf scab segmentation method based on color information according to an embodiment of the present invention, and as shown in fig. 2, a CCF ratio R indicates how close a CCF value of an affected leaf area under an illumination condition is to a CCF value of a normal leaf area, and a smaller value of R indicates that the closer the CCF value of the affected leaf area under the illumination condition is to the CCF value of the normal leaf area, the smaller the influence of the illumination condition is. The preset number may be set according to an actual situation, and the embodiment of the present invention is not limited.
On the basis of the above embodiments, further, the lesion segmentation module 305 includes an acquisition unit and a segmentation unit, wherein:
the acquisition unit is used for acquiring the growth position of the initial seed point in the comprehensive color feature map; and the segmentation unit is used for segmenting the scab in the scab leaf image by adopting a region growing method according to the growing position of the initial seed point.
Specifically, the obtaining unit obtains the growth positions of initial seed points in the integrated color feature map, where there may be a plurality of growth positions of the initial seed points; wherein, the growth position of the seed point can be calibrated manually. And the segmentation unit is used for segmenting the scab in the scab leaf image by adopting a region growing method according to the growing position of the initial seed point.
The embodiment of the server provided by the present invention may be specifically configured to execute the processing flows of the above method embodiments, and the functions of the embodiment are not described herein again, and refer to the detailed description of the above method embodiments.
The above-described server embodiments are only illustrative, and the units described as separate components may or may not be physically separate, and components displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (4)

1. A vegetable leaf scab segmentation method based on color information is characterized by comprising the following steps:
acquiring the scab leaf images of the vegetables, and preprocessing the scab leaf images;
respectively extracting the color characteristics of the preprocessed scab leaf images in different color spaces;
constructing comprehensive color feature detection filtering according to the color features of the different color spaces;
detecting and filtering according to the comprehensive color features to obtain a comprehensive color feature map corresponding to the scab leaf image;
according to the comprehensive color feature map, dividing the scab in the scab leaf image;
the segmenting the scab in the scab leaf image according to the comprehensive color feature map comprises the following steps:
acquiring the growth position of an initial seed point in the comprehensive color feature map;
according to the growth position of the initial seed point, a region growing method is adopted to segment the scab in the scab leaf image;
the preprocessing comprises image denoising and size adjustment of the scab blade image;
the color features of the lesion leaf images in different color spaces after the respective extraction preprocessing include:
respectively extracting the super red features of the preprocessed scab leaf images in an RGB color space, the H component in an HSV color space and the b component in an L, a, b and color space;
specifically, ExR of the preprocessed lesion leaf image in an RGB color space, an H component in an HSV color space and a b component in an L, a and b color space are extracted; ExR according to formula IExRExtracting (x, y) ═ 1.3R (x, y) -G (x, y), where (x, y) is pixel coordinates, and R (x, y), G (x, y) are color component values of RGB color space (x, y); the H component in HSV color space is according to formula H(x,y)=IH(x, y) extraction, wherein (x, y) is pixel coordinate, IH(x, y) is the value of the H component of (x, y) in the HSV color space; b component in L a b color space according to formulaExtracting, wherein (x, y) is pixel coordinate,taking values for the b component of (x, y) in L a b color space;
the constructing of the comprehensive color feature detection filter according to the color features of the different color spaces includes:
according to the formula Constructing comprehensive color feature detection filtering, wherein I is the preprocessed blade image, and IExRIs a super red feature image of the RGB color space,b component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)HL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation; wherein a single pixel point in the preprocessed leaf image can be represented as I (x, y), and the integrated color feature detection filtering of the single pixel point can be represented as I (x, y) Wherein, IExR(x, y) is a hyper red feature component value of (x, y) of the RGB color space.
2. The method of claim 1, wherein obtaining a composite color feature map corresponding to the lesion leaf image based on the composite color feature detection filtering comprises:
according to the formula: CCF ═ exp (- β | f (I: r, σ)HLR) |) obtaining the integrated color feature map, wherein CCF represents the integrated color feature map, β is a descent rate parameter, and αRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formulaObtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) And i and j are positive integers, wherein the CCF values of the pixel points on the comprehensive color feature image are represented by the I and the J.
3. A server, comprising:
the image preprocessing module is used for acquiring scab leaf images of vegetables and preprocessing the scab leaf images;
the characteristic extraction module is used for respectively extracting the color characteristics of the preprocessed scab blade images in different color spaces;
the filtering construction module is used for constructing comprehensive color characteristic detection filtering according to the color characteristics of the different color spaces;
the characteristic diagram obtaining module is used for detecting and filtering according to the comprehensive color characteristic to obtain a comprehensive color characteristic diagram corresponding to the scab leaf image;
the scab segmentation module is used for segmenting scabs in the scab leaf images according to the comprehensive color feature map;
the lesion segmentation module comprises an acquisition unit and a segmentation unit, wherein:
the acquiring unit is used for acquiring the growth position of the initial seed point in the comprehensive color feature map;
the segmentation unit is used for segmenting the scab in the scab leaf image by adopting a region growing method according to the growing position of the initial seed point;
the preprocessing comprises image denoising and size adjustment of the scab blade image;
the feature extraction module is specifically configured to:
respectively extracting the super red features of the preprocessed scab leaf images in an RGB color space, the H component in an HSV color space and the b component in an L, a, b and color space;
specifically, ExR of the preprocessed lesion leaf image in an RGB color space, an H component in an HSV color space and a b component in an L, a and b color space are extracted; ExR according to formula IExRExtracting (x, y) ═ 1.3R (x, y) -G (x, y), where (x, y) is pixel coordinates, and R (x, y), G (x, y) are color component values of RGB color space (x, y); the H component in HSV color space is according to formula H(x,y)=IH(x, y) extraction, wherein (x, y) is pixel coordinate, IH(x, y) is the value of the H component of (x, y) in the HSV color space; l a b color spaceB component according to formulaExtracting, wherein (x, y) is pixel coordinate,taking values for the b component of (x, y) in L a b color space;
the filtering construction module is specifically configured to:
according to the formula Constructing comprehensive color feature detection filtering, wherein I is the preprocessed blade image, and IExRIs a super red feature image of the RGB color space,b component image of L a b color space, IHIs an H component image of HSV color space, pb(r)Is a circular area mean filter with a radius r,is a standard deviation of (σ)HL) Alpha is a super red characteristic parameter of the RGB color space, and the value range is (0, 1)]Is a two-dimensional discrete convolution operation; wherein a single pixel point in the preprocessed leaf image can be represented as I (x, y), and the integrated color feature detection filtering of the single pixel point can be represented as I (x, y) Wherein, IExR(x, y) is a hyper red feature component value of (x, y) of the RGB color space.
4. The server according to claim 3, wherein the profile obtaining module is specifically configured to:
according to the formula: CCF ═ exp (- β | f (I: r, σ)HLR) |) obtaining the integrated color feature map, wherein CCF represents the integrated color feature map, β is a descent rate parameter, and αRIs the alpha value corresponding to the minimum R value, R is the CCF ratio, and is expressed by the formulaObtained by calculation, wherein M is the pixel number of the leaf area affected by illumination, N is the pixel number of the normal leaf area, and CCF (x)i,yi) And CCF (x)j,yj) And i and j are positive integers, wherein the CCF values of the pixel points on the comprehensive color feature image are represented by the I and the J.
CN201710358725.1A 2017-05-19 2017-05-19 Vegetable leaf scab segmentation method based on color information and server Active CN107154049B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710358725.1A CN107154049B (en) 2017-05-19 2017-05-19 Vegetable leaf scab segmentation method based on color information and server

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710358725.1A CN107154049B (en) 2017-05-19 2017-05-19 Vegetable leaf scab segmentation method based on color information and server

Publications (2)

Publication Number Publication Date
CN107154049A CN107154049A (en) 2017-09-12
CN107154049B true CN107154049B (en) 2020-12-25

Family

ID=59793052

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710358725.1A Active CN107154049B (en) 2017-05-19 2017-05-19 Vegetable leaf scab segmentation method based on color information and server

Country Status (1)

Country Link
CN (1) CN107154049B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184392A (en) * 2011-05-25 2011-09-14 中国水稻研究所 DSP (Digital Signal Processor)-based rice weed recognition system and method

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103247059B (en) * 2013-05-27 2016-02-17 北京师范大学 A kind of remote sensing images region of interest detection method based on integer wavelet and visual signature
CN103559511A (en) * 2013-11-20 2014-02-05 天津农学院 Automatic identification method of foliar disease image of greenhouse vegetable
CN103996185A (en) * 2014-04-29 2014-08-20 重庆大学 Image segmentation method based on attention TD-BU mechanism
CN104021544B (en) * 2014-05-07 2018-11-23 中国农业大学 A kind of greenhouse vegetable disease monitor video extraction method of key frame, that is, extraction system
CN103996041B (en) * 2014-05-15 2015-07-22 武汉睿智视讯科技有限公司 Vehicle color identification method and system based on matching
CN104063686B (en) * 2014-06-17 2017-08-29 中国科学院合肥物质科学研究院 Crop leaf diseases image interactive diagnostic system and method
CN104598908B (en) * 2014-09-26 2017-11-28 浙江理工大学 A kind of crops leaf diseases recognition methods
CN106023159B (en) * 2016-05-10 2019-02-26 中国农业大学 Facilities vegetable leaf portion scab image partition method and system
CN106250895B (en) * 2016-08-15 2019-07-26 北京理工大学 A kind of remote sensing image region of interest area detecting method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184392A (en) * 2011-05-25 2011-09-14 中国水稻研究所 DSP (Digital Signal Processor)-based rice weed recognition system and method

Also Published As

Publication number Publication date
CN107154049A (en) 2017-09-12

Similar Documents

Publication Publication Date Title
Welikala et al. Genetic algorithm based feature selection combined with dual classification for the automated detection of proliferative diabetic retinopathy
Budai et al. Robust vessel segmentation in fundus images
Zhao et al. Retinal vessels segmentation based on level set and region growing
Almazroa et al. Optic disc and optic cup segmentation methodologies for glaucoma image detection: a survey
GeethaRamani et al. Retinal blood vessel segmentation employing image processing and data mining techniques for computerized retinal image analysis
Aquino et al. Detecting the optic disc boundary in digital fundus images using morphological, edge detection, and feature extraction techniques
Chiu et al. Automatic segmentation of seven retinal layers in SDOCT images congruent with expert manual segmentation
Neto et al. An unsupervised coarse-to-fine algorithm for blood vessel segmentation in fundus images
Marin et al. Obtaining optic disc center and pixel region by automatic thresholding methods on morphologically processed fundus images
Martinez-Perez et al. Segmentation of blood vessels from red-free and fluorescein retinal images
Ege et al. Screening for diabetic retinopathy using computer based image analysis and statistical classification
Dorj et al. An yield estimation in citrus orchards via fruit detection and counting using image processing
Fraz et al. Application of morphological bit planes in retinal blood vessel extraction
JP4420821B2 (en) Histological assessment
Sekhar et al. Automated localisation of retinal optic disk using Hough transform
Zhu et al. Detection of the optic nerve head in fundus images of the retina using the hough transform for circles
Niemeijer et al. Comparative study of retinal vessel segmentation methods on a new publicly available database
US8811698B2 (en) Image processing apparatus, image processing method, and computer-readable recording medium
Rodrigues et al. Segmentation of optic disc and blood vessels in retinal images using wavelets, mathematical morphology and Hessian-based multi-scale filtering
US7689016B2 (en) Automatic detection of critical dermoscopy features for malignant melanoma diagnosis
Payne et al. Estimating mango crop yield using image analysis using fruit at ‘stone hardening’stage and night time imaging
Wang et al. Image recognition of plant diseases based on backpropagation networks
EP2888718B1 (en) Methods and systems for automatic location of optic structures in an image of an eye, and for automatic retina cup-to-disc ratio computation
Rabbani et al. Fully automatic segmentation of fluorescein leakage in subjects with diabetic macular edema
Liu et al. A novel transferable individual tree crown delineation model based on Fishing Net Dragging and boundary classification

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