CN113570538B - Blade RGB image bias distribution parameter information acquisition and analysis method - Google Patents
Blade RGB image bias distribution parameter information acquisition and analysis method Download PDFInfo
- Publication number
- CN113570538B CN113570538B CN202010350189.2A CN202010350189A CN113570538B CN 113570538 B CN113570538 B CN 113570538B CN 202010350189 A CN202010350189 A CN 202010350189A CN 113570538 B CN113570538 B CN 113570538B
- Authority
- CN
- China
- Prior art keywords
- blade
- image
- color
- parameters
- distribution
- 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
Links
- 238000009826 distribution Methods 0.000 title claims abstract description 64
- 238000004458 analytical method Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000003491 array Methods 0.000 claims description 9
- 230000001186 cumulative effect Effects 0.000 claims description 9
- 229930002875 chlorophyll Natural products 0.000 claims description 7
- 235000019804 chlorophyll Nutrition 0.000 claims description 7
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000007405 data analysis Methods 0.000 claims description 3
- 238000013075 data extraction Methods 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims 1
- 238000000691 measurement method Methods 0.000 claims 1
- 230000008859 change Effects 0.000 abstract description 20
- 241000208125 Nicotiana Species 0.000 abstract description 9
- 235000002637 Nicotiana tabacum Nutrition 0.000 abstract description 9
- 241000196324 Embryophyta Species 0.000 description 5
- 230000002596 correlated effect Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000013488 ordinary least square regression Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 235000011299 Brassica oleracea var botrytis Nutrition 0.000 description 2
- 240000003259 Brassica oleracea var. botrytis Species 0.000 description 2
- 238000001727 in vivo Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 240000007124 Brassica oleracea Species 0.000 description 1
- 235000003899 Brassica oleracea var acephala Nutrition 0.000 description 1
- 235000011301 Brassica oleracea var capitata Nutrition 0.000 description 1
- 235000017647 Brassica oleracea var italica Nutrition 0.000 description 1
- 235000001169 Brassica oleracea var oleracea Nutrition 0.000 description 1
- 235000007340 Hordeum vulgare Nutrition 0.000 description 1
- 240000005979 Hordeum vulgare Species 0.000 description 1
- 240000008415 Lactuca sativa Species 0.000 description 1
- 235000003228 Lactuca sativa Nutrition 0.000 description 1
- 235000007688 Lycopersicon esculentum Nutrition 0.000 description 1
- 240000003768 Solanum lycopersicum Species 0.000 description 1
- 235000021307 Triticum Nutrition 0.000 description 1
- 241000209140 Triticum Species 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 210000003462 vein Anatomy 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
The invention discloses a blade RGB image bias distribution parameter information acquisition and analysis method, which adopts Photoshop software to cut and denoise a blade acquired by a digital camera, adopts MATLAB software to perform bias analysis on Red, green, blue channels and the blade color level distribution condition of a gray level image and extract relevant parameters, adopts tobacco single-blade color bias distribution parameters to construct a nonlinear correlation model with SPAD values, and compares the nonlinear correlation model with a regression model constructed by a traditional parameter traditional method, thereby verifying the feasibility and superiority of the bias analysis method in blade color-physiological index prediction. The method greatly enriches the information quantity of the RGB model based on the extracted bias distribution parameters of the blade RGB model with the bias distribution, can describe the change condition of the blade color from two aspects of the depth of the blade color and the uniformity of the distribution, and improves the accuracy of individual character description and model prediction.
Description
Technical Field
The invention relates to the field of plant leaf image information, in particular to a leaf RGB image bias distribution parameter information acquisition and analysis method.
Background
With the increasing maturity of digital image technology and the popularization of high-resolution image pickup equipment, the research of qualitative and quantitative description of plant appearance phenotype characteristics by using digital images is increasing. The digital camera can record the spectrum information of the visible wave band, and has high resolution and low cost. In addition, digital static color images contain a large amount of information on plant morphology and leaf color and are therefore often used to identify changes in leaf color.
The most common analysis method for digital color images is the RGB color model. Other methods (e.g., HSV color model, texture Analysis) are also converted from RGB data. In the RGB color model, three color sensors per pixel may be used to capture the light intensities in the red, green, and blue spectra, respectively, to describe the color of any pixel in the object image, all using existing software, such as (MathWorks inc.) or Image J.
The use of RGB for leaf color has been a long history. Since the beginning of RGB data in 1988 to determine chlorophyll content, the application of RGB data of leaves in plants has been focused on determining chlorophyll content and the index related to the change of chlorophyll content over several decades. In order to deeply dig the application value of RGB data, researchers have proposed a variety of derivative parameters such as (R-B)/(r+b), G/(r+g+b), R/(r+g+b), G/R, r+g+b, R-B, r+b, r+g, log sig ((G-R/3-B/3)/255) for determining chlorophyll content (KAWASHIMA AND NAKATANI 1998;Adamsen et al.1999;Hu et al.2010;Cai et al.2006;Ali et al.2012) in wheat, broccoli, cabbage, barley, tomato, lettuce and cauliflower, however, the problem that the information amount is small cannot be changed by combining R, G, B mean parameters, which becomes a bottleneck for RGB model application, greatly limiting its application range. Hyperspectral or multispectral are in turn investigated.
The existing RGB model has obvious limitation in leaf color analysis, and is mainly characterized in that: 1. the traditional image color level analysis method is based on analysis of data under normal distribution conditions, is a convenient approximate value taking method, and cannot comprehensively and truly reflect the distribution condition of leaf colors. 2. The information obtained from the RGB image is too small, and there are only Red, green, blue color-level averages of the three channels, and although there are several researchers who propose various combination parameters of these three base parameters, they cannot solve their physiological significance in describing the change of the leaf color. 3. The traditional research generally adopts a stepwise regression model to construct a leaf color-physiological parameter model, but some physiological indexes are not linearly related to leaf color change, and the model constructed by the method has poor fitting degree and low prediction precision.
Disclosure of Invention
The invention aims to provide a blade RGB image bias distribution parameter information acquisition and analysis method so as to describe the change condition of the depth and uniformity of the blade color more accurately.
The technical problems solved by the invention can be realized by adopting the following technical scheme:
A blade RGB image bias distribution parameter information acquisition and analysis method comprises the following steps:
Step one: blade image acquisition;
Step two: blade image cutting denoising: preprocessing the blade image acquired in the second step by Adobe Photoshop CS software, removing the background, saving the blade image as a PNG image mode, and adjusting the resolution of the image to 1000 x 1330;
Step three: and (3) carrying out data extraction and analysis on the color image by applying MATLAB software to the image preprocessed in the step two, wherein the process comprises the following steps: 1) The method comprises the steps of establishing color gradation arrays of different color channels, reading color images through imread functions, reading three channels of a blade color image Red, green, blue and Gray image gradation information through functions, and converting the color gradation arrays into double-precision arrays through double functions; 2) And (6) building a single blade color level cumulative histogram. Respectively acquiring three channels of a blade color image Red, green, blue and a Gray image tone scale cumulative histogram by using imhist functions; 3) And establishing a leaf color deviation parameter table, and respectively analyzing and acquiring parameters such as the mean value, the median value, the mode value, the standard deviation, the variance, the skewness, the kurtosis and the like of the double-precision arrays of the three channels of the leaf color image Red, green, blue and the Gray image through mean, median, mode, std, var, skewness, kurtosis functions to form the leaf color deviation distribution parameter table.
Preferably, the method also comprises space modeling, wherein the average value, the median, the mode, the skewness and the kurtosis of the three-channel and gray-scale image are used as input variables, the SPAD value is used as a dependent variable, and a SPAD value-leaf color space polynomial model is constructed by using the polynomial fitting in Curve FittingTool of MATLAB software.
Preferably, the image capturing in the step one adopts a high-resolution digital camera, and the resolution of the captured original digital image is 3840×5120.
According to the invention, after blades acquired by a digital camera are cut and denoised by adopting Photoshop software, the blade color level distribution conditions of Red, green, blue channels and gray images are subjected to bias analysis by adopting MATLAB software, related parameters are extracted, a nonlinear correlation model with SPAD values is constructed by adopting tobacco single-blade bias distribution parameters, and the nonlinear correlation model is compared with a regression model constructed by a traditional parameter traditional method, so that feasibility and superiority of a bias analysis method in blade color-physiological index prediction are verified. The method greatly enriches the information quantity of the RGB model based on the extracted bias distribution parameters of the blade RGB model with the bias distribution, can describe the change condition of the blade color from two aspects of the depth of the blade color and the uniformity of the distribution, and improves the accuracy of individual character description and model prediction.
Drawings
Fig. 1 is a leaf color RGB tone cumulative histogram of an embodiment.
FIG. 2 is a leaf color-SPAD Fourier fitting model based on bias parameters.
FIG. 3 is a leaf color-SPAD spatial high-order fitting model based on bias parameters.
Detailed Description
The present invention will be described in detail below with reference to the drawings and specific examples, but embodiments of the present invention are not limited thereto.
The method for acquiring and analyzing the information of the bias distribution parameters of the RGB image of the tobacco leaf is illustrated by taking the tobacco leaf as an example.
The in vivo leaf SPAD values were first measured. Tobacco leaves of different leaf ages (40 d, 50d, 60d, 65 d) were selected and the in vivo leaf SPAD values were measured. The instrument adopted by the invention for measuring the SPAD value of the living blade is se:Sup>A Hope brand hand-held TYS-A chlorophyll meter, and the emission window and the receiving window are ensured to be clean and free from foreign matters before measurement, and blank calibration is carried out; five-point measurement is adopted in the measurement process, 5 mesophyll SPAD values of the leaf tips, left and right directions of the centers of the blades and left and right directions of the handles of the blades are respectively measured, the sample is ensured to completely cover the receiving window in the measurement process, and only mesophyll tissues are measured while the veins are avoided. In data processing, the SPAD value for each blade is the average of the 5 azimuth data points measured.
Step one: blade image acquisition: and (3) immediately sending picked leaves into a room for image acquisition after the picked leaves are packaged in fresh bags, wherein the fresh She Li body duration is less than 3min. The image acquisition adopts a high-resolution digital camera, and the resolution of the acquired original digital image is 3840 x 5120. The collected cumulative histogram of the tone scale of the leaf color Red, green, blue THREE CHANNELS AND GRAY IMAGE is shown in fig. 1. It can be seen from fig. 1 that the cumulative histograms of the hues RGB all exhibit a biased distribution; distribution histograms of different leaf ages, different channels, also show different degrees of deviation.
Step two: blade image cutting denoising: and (3) preprocessing the blade image acquired in the step (I) by adopting Adobe Photoshop CS software, saving the blade image as a PNG image mode after removing the background, and adjusting the resolution of the image to 1000 x 1330.
Step three: and (3) carrying out data extraction and analysis on the color image by applying MATLAB software to the image preprocessed in the step two, wherein the process comprises the following steps: 1) And establishing color gradation arrays of different color channels. And reading a color image through imread functions, reading three channels of blade color images Red (Red channel), green (Green channel), blue (Blue channel) and Gray (Gray) image tone scale information through functions, and converting a tone scale array into a double-precision array through double functions. 2) And (6) building a single blade color level cumulative histogram. And respectively acquiring three channels of the blade color image Red, green, blue and Gray image tone scale cumulative histograms by using imhist functions. 3) And (6) establishing a leaf color deviation parameter table. And respectively analyzing and acquiring parameters such as the mean value, the median value, the mode value, the standard deviation, the variance, the skewness, the kurtosis and the like of the double-precision array of the blade color image Red, green, blue and the Gray image through mean, median, mode, std, var, skewness, kurtosis functions to form a blade color skewness distribution parameter table.
The color level distribution of each leaf color was performed Normal distribution test using the Lilliefors test, jarque-Bera test function provided in the MATLAB software statistics toolbox. The results show that the H value of the normal hypothesis of both checks is 1, namely, the negative normal distribution hypothesis; and the statistical distribution probability detection significance P values are all 0.001<0.05, which indicates that the leaf color level distribution of different channels of different leaf ages of the tobacco leaf does not accord with normal distribution, and the leaf color distribution is biased distribution as can be seen by combining with FIG. 1.
And respectively analyzing and obtaining the average value of the double-precision arrays of the three channels of the blade color image Red, green, blue through a mean function, and combining the three basic values according to R+G+ B, R/(R+G+B), G/(R+G+B), B/(R+G+B), R-G, R-B, G-B, R + G, R + B, B +G to form a three-channel color-level average value and a combination parameter table thereof.
And (3) taking the average value, the median, the mode, the skewness and the kurtosis of the three-channel gray image as input variables, taking the SPAD value as dependent variable, and constructing the SPAD value-leaf color space polynomial model by using the polynomial fitting in Curve Fitting Tool of MATLAB software.
Plant leaf color individual parameter and SPAD association model establishment and comparison based on bias distribution
The parameters of the leaf color RGB model can be expanded to the parameters of the mean value, the median, the mode, the variance, the standard deviation, the skewness, the kurtosis and the like of three-channel and gray images by using the analysis method of the skewness distribution. In the leaf color distribution, the variance, standard deviation, skewness and kurtosis all represent the uniformity of the leaf color distribution, and the variance and the standard deviation reflect the degree of data dispersion and are influenced by statistics; the skewness and kurtosis are distribution shape parameters for measuring the data set, and the influence of statistics is eliminated, so that in the statistical analysis, 20 parameters such as the mean value, the median, the mode, the skewness and the kurtosis of the three-channel and gray level images are taken as main parameters. After SPAD values of 50 blades of four blade ages and three-channel and gray image bias distribution parameters are taken for single-factor LSD variance analysis, obvious differences exist between the SPAD values and the bias distribution parameters of the blades of different blade ages, and the differences are shown in Table 1. After the SPAD values of the blades are subjected to correlation analysis with three-way mean values and combination parameters thereof adopted by the traditional research, table 2 is obtained, and it can be seen that 12 parameters are extremely obviously correlated with the SPAD values at the level of 0.01, and 1 is obviously correlated at the level of 0.05; and the SPAD values of the blades are subjected to correlation analysis with the bias distribution parameters of the three-channel and gray image to obtain a table 3, wherein 17 bias parameters are obviously correlated with the SPAD values at the 0.01 level. This suggests that the change in chlorophyll content is highly correlated with the change in leaf color, which is not only in the degree of leaf color depth, but also numerically related to mean, median, mode; also reflected in the unevenness of the leaf color distribution region, is numerically represented as being related to the skewness.
Table 1 leaf color parameter table of tobacco leaves of different ages (n=50)
TABLE 2 three channel means and combined parameters thereof SPAD value correlation analysis table
Note:**indicates significant correlation at the 0.01level(Two-tailed);*indicates significant correlation at the 0.05level(Two-tailed).The same below.
TABLE 3 analysis Table for correlation of three-channel and gray image bias distribution parameters and blade SPAD values
After the correlation between the SPAD value of the blade and the blade color parameter is determined, the built SPAD value-blade color model is compared in terms of fitting goodness and prediction accuracy by adopting different types of parameters and model building methods.
1) In the prior art, three-way mean values and combination parameters thereof in table 2 are generally used as input variables of the leaf color model, and a stepwise regression method based on a least squares method (Ordinary Least Square, OLS) is used to construct the correlation model. SPSS software was used to construct the SPAD-leaf color model Y 1 using conventional parameters Table 2:
Y1＝59.733-0.304×X1
Wherein Y 1 is the blade SPAD value, and X1 is the blade Red channel mean value.
It can be seen that in the conventional leaf color model, the available input variables have only Red, green, blue channel means, and the other 10 parameters are all obtained by combining the three parameters, so that there is a large autocorrelation among the parameters of the conventional leaf color model. In the invention, the SPSS software does not adopt combination parameters in the stepwise regression modeling process, the finally constructed model only uses the mean value of the Red channel, and the R 2 of the constructed model Y1 is lower, see Table 5, and only 0.583.
2) Further expanding the parameter range, and constructing a SPAD-leaf color model Y 2 by using the bias distribution parameters in table 3 as input variables:
Y2＝76.134-0.441×X1-11.203×X2-1.516×X3
Wherein Y 2 is the blade SPAD value, X 1 is the blade Red channel mean value, X 2 is the gray image skewness, and X 3 is the Green channel kurtosis.
As can be seen from the goodness of fit, R 2 of Y2 is 0.694, which is 19.04% higher than Y1, which shows that the model constructed by adopting the bias parameters is better than the model constructed by the traditional parameters, and the SPAD value can be better fitted.
3) As Y1 and Y2 are both linear models, the change of the SPAD value of the tobacco leaf blade in the table 1-40 d-65d can be seen, and the SPAD value shows the change trend of rising and then falling. To better reflect the correlation between SPAD values and leaf color variation, a fourier fit model Y 3 is built:
Y3＝19.38+7.972×cos(1.314×X)-6.747×sin(1.314×X)
Wherein Y 3 is the blade SPAD value, X is the blade Red channel median, and the fitted curve is shown in FIG. 2.
In the present study, it was found that the change of SPAD value is not only related to leaf color depth but also related to uneven leaf color distribution, and the Y3 model only uses a parameter reflecting leaf color depth, and cannot completely fit the function law of the variation gauge imhist of SPAD value, which is shown that R 2 of Y3 is superior to Y1 and inferior to Y2.
3) As can be seen from table 1, as the leaf age increases, both the shade parameters (mean, median, mode) and the distribution parameters (skewness, kurtosis) of the leaf color change, and the changes in the magnitude and direction of the changes are asynchronous. Non-linear functions using a single type of parameter (e.g., Y3 using only the Red channel median leaf color depth parameter) cannot fit the SPAD value variations perfectly. By combining Y2 and Y3 and applying the principle of leaf color change bi-directionality (depth and distribution) and the nonlinear law of SPAD value change, a space bi-vector-based polynomial fitting model Y 4 is constructed:
Y4＝0.3344+0.8709×x-177.3×y-0.005536×x2+2.876×x×y+8.515×y2-0.01227×x2×y-0.1398×x×y2+7.301×y3
Wherein Y4 is the SPAD value of the blade, x is the mean value of the Red channel of the blade, Y is the deflection of the Red channel of the blade, and the fitted curved surface is shown in FIG. 3.
As can be seen from Table 4, R-square and Adiusted R-square, SSE, RMSE of the Y 4 model are superior to those of other models, and R 2 of Y4 is improved by 23.33% compared with the traditional Y1 model. To verify the prediction accuracy of four models, we performed model verification with 168 leaves of another batch of four leaf ages collected at the same time as the object. As can be seen from Table 5, the prediction accuracy of the Y 4 model is higher than that of other models, and the prediction standard deviation is better than that of other models, so that the leaf SPAD change condition can be better simulated by adopting a space polynomial with bidirectional leaf color change (leaf color shade change and leaf color distribution change).
Table 4 Model Goodness of fit determination
Table 5 model predictive accuracy analysis table
Note:Predictive accuracy＝(1-|predictivevalue-measuredvalue|/measured)*100％
As can be seen from the above examples:
1. When the model is built by using the OLS-based stepwise regression method, the model Y2 built by adopting the bias distribution parameters is superior to the model Y1 built by the traditional parameters;
2. when the single leaf color depth parameters are used, the nonlinear model Y3 is adopted to be better than the traditional linear model Y1;
3. The space polynomial model Y4 fitting goodness constructed by combining the multidimensional parameter characteristics (leaf color depth parameters and leaf color distribution parameters) of the Y2 model and the nonlinear model advantages of the Y3 model is optimal in four models, the prediction accuracy is highest, and the parameter selection and model construction from two dimensions of leaf color depth and leaf color distribution are required to be carried out under the condition of leaf color bias distribution so as to more comprehensively fit and predict the change rule of the leaf color SPAD value.
According to the invention, through observing the color level distribution histograms of the RGB models of the fresh tobacco leaves of different channels with different leaf ages, the combined color level groups are found after normal inspection, and the leaf color level distribution shows the bias distribution. The leaf color parameters of the RGB model are expanded, the basic data are expanded from only 3 parameters of mean value, variance and standard deviation of each channel to 7 parameters of mean value, median, mode, standard deviation, variance, skewness and kurtosis, and the total of 28 parameters is added to the three color channels and one gray picture. Wherein the mean value, the median and the mode are numerical parameters, and the state of the leaf color depth can be quantitatively described; the standard deviation, variance, skewness and kurtosis are discrete parameters, and mainly reflect the uniformity of leaf color distribution. This makes it possible to accurately quantitatively describe the leaf color state in terms of both leaf color depth and leaf color distribution.
Because the tone scale distribution of the blade RGB model is the bias distribution, the information of the RGB model is greatly enriched based on the extracted bias distribution parameters (mean, median, mode, skewness, kurtosis and the like). The change condition of the leaf color is described from two aspects of leaf color depth and distribution uniformity, and the accuracy of individual character description and model prediction is improved.
The foregoing description is only a preferred embodiment of the present invention, and is not intended to limit the technical scope of the present invention, so any minor modifications, equivalent changes and modifications made to the above embodiments according to the technical spirit of the present invention still fall within the scope of the present invention.
Claims (2)
1. A blade RGB image bias distribution parameter information acquisition and analysis method comprises the following steps:
Step one: selecting a plurality of blades with different blade ages, acquiring the SPAD value of the blade by using a chlorophyll meter through a five-point measurement method, calculating the SPAD average value of a single blade, and manufacturing a corresponding relation table of the blade ages and the SPAD average value as a comparison standard of the bias distribution parameters;
Step two: blade image acquisition;
step three: blade image cutting denoising: preprocessing the blade image acquired in the second step by Adobe Photoshop CS software, removing the background, saving the blade image as a PNG image mode, and adjusting the resolution of the image to 1000 ANG 1330;
Step four: and (3) carrying out data extraction and analysis on the color image by applying MATLAB software to the image preprocessed in the step two, wherein the process comprises the following steps: 1) The method comprises the steps of establishing color gradation arrays of different color channels, reading color images through imread functions, reading three channels of a blade color image Red, green, blue and Gray image gradation information through functions, and converting the color gradation arrays into double-precision arrays through double functions; 2) Constructing a single blade color gradation cumulative histogram, and respectively obtaining three channels of a blade color image Red, green, blue and Gray image color gradation cumulative histograms by using imhist functions; 3) The method comprises the steps of establishing a leaf color deviation parameter table, respectively analyzing and obtaining parameters of the mean value, the median value, the mode value, the standard deviation, the variance, the deviation and the kurtosis of three channels of a leaf color image Red, green, blue and a Gray image double-precision array through mean, median, mode, std, var, skewness, kurtosis functions to form a leaf color deviation distribution parameter table;
step five: carrying out correlation analysis on the bias distribution parameter table in the fourth step and the corresponding relation table of the leaf ages and the SPAD mean values in the first step;
step six: and (3) space modeling, namely, taking 20 parameters of the mean value, the median, the mode, the skewness and the kurtosis of the three-channel gray image as input variables, taking the SPAD value as dependent variables, and constructing a SPAD value-leaf color space polynomial model by using the CurveFitting Tool polynomial fitting in MATLAB software.
2. The method for acquiring and analyzing the information of the skewed distribution parameters of the RGB image of the blade according to claim 1, wherein the method comprises the following steps: and step two, the image acquisition adopts a high-resolution digital camera, and the resolution of the acquired original digital image is 3840 ANG 5120.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010350189.2A CN113570538B (en) | 2020-04-28 | 2020-04-28 | Blade RGB image bias distribution parameter information acquisition and analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010350189.2A CN113570538B (en) | 2020-04-28 | 2020-04-28 | Blade RGB image bias distribution parameter information acquisition and analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113570538A CN113570538A (en) | 2021-10-29 |
| CN113570538B true CN113570538B (en) | 2024-05-17 |
Family
ID=78158040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010350189.2A Active CN113570538B (en) | 2020-04-28 | 2020-04-28 | Blade RGB image bias distribution parameter information acquisition and analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113570538B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114022389B (en) * | 2021-11-17 | 2022-08-09 | 河北农业大学 | Method for distinguishing plant leaf color grade based on digital camera |
| CN116486969B (en) * | 2023-06-25 | 2023-09-26 | 广东工业大学 | Genetic algorithm-based material optimal correlation relation acquisition method and application |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102968793A (en) * | 2012-11-20 | 2013-03-13 | 百年金海安防科技有限公司 | Method for identifying natural image and computer generated image based on DCT (Discrete Cosine Transformation)-domain statistic characteristics |
| CN109410232A (en) * | 2018-09-26 | 2019-03-01 | 漳州市佰佳贸易有限公司 | Based on conventional low resolution camera to the extracting method of Leaf color character value |
| CN110400293A (en) * | 2019-07-11 | 2019-11-01 | 兰州理工大学 | A kind of non-reference picture quality appraisement method based on depth forest classified |
-
2020
- 2020-04-28 CN CN202010350189.2A patent/CN113570538B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102968793A (en) * | 2012-11-20 | 2013-03-13 | 百年金海安防科技有限公司 | Method for identifying natural image and computer generated image based on DCT (Discrete Cosine Transformation)-domain statistic characteristics |
| CN109410232A (en) * | 2018-09-26 | 2019-03-01 | 漳州市佰佳贸易有限公司 | Based on conventional low resolution camera to the extracting method of Leaf color character value |
| CN110400293A (en) * | 2019-07-11 | 2019-11-01 | 兰州理工大学 | A kind of non-reference picture quality appraisement method based on depth forest classified |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113570538A (en) | 2021-10-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11454594B2 (en) | Method for identifying frostbite condition of grain seeds using spectral feature wavebands of seed embryo hyperspectral images | |
| CN110763698B (en) | Hyperspectral citrus leaf disease identification method based on characteristic wavelength | |
| CN102495005B (en) | Method for diagnosing crop water deficit through hyperspectral image technology | |
| AU2006264799B2 (en) | Method and device for analysing berries | |
| CN112541921B (en) | Urban green land vegetation information data accurate determination method | |
| CN111968080A (en) | Hyperspectrum and deep learning-based method for detecting internal and external quality of Feicheng peaches | |
| CN113570538B (en) | Blade RGB image bias distribution parameter information acquisition and analysis method | |
| CN109726705B (en) | Mangrove forest information extraction method and device and electronic equipment | |
| CN108734205A (en) | A kind of simple grain for different cultivars wheat seed pinpoints identification technology | |
| CN108520127A (en) | A kind of EO-1 hyperion inversion method of seeds leaf area index | |
| CN113109240B (en) | Method and system for determining imperfect grains of grains implemented by computer | |
| CN115495703B (en) | Tobacco maturity detection method and system based on airborne multispectral data | |
| CN111751376A (en) | Rice nitrogen nutrition estimation method based on canopy image feature derivation | |
| CN109870419B (en) | Method for predicting nitrogen, phosphorus and potassium content of black soil by adopting aviation hyperspectral data | |
| CN115015258A (en) | Crop growth and soil moisture association determination method and related device | |
| CN112488997B (en) | Method for detecting and evaluating color reproduction of ancient painting printed matter based on characteristic interpolation | |
| CN112329733B (en) | Winter wheat growth monitoring and analyzing method based on GEE cloud platform | |
| CN114136445B (en) | Method for improving airborne hyperspectral radiation correction precision of unmanned aerial vehicle | |
| CN115392305A (en) | Soil organic matter content high spectrum modeling method based on improved time convolution network | |
| CN111257854B (en) | Universal terrain correction optimization method based on remote sensing image segmentation unit | |
| CN112001910A (en) | Method and device for automatically identifying number of plant ears, electronic equipment and storage medium | |
| CN117309780B (en) | Method for determining content of germanium element in soil | |
| Cai et al. | Rapid measurement of potato canopy coverage and leaf area index inversion | |
| CN112329856B (en) | Wheat nitrogen accumulation amount prediction method based on spectrum and RGB-D fusion image | |
| CN113887493B (en) | Black and odorous water body remote sensing image identification method based on ID3 algorithm |
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 |