Response of Population Canopy Color Gradation Skewed Distribution Parameters of the RGB model to Micrometeorology Environment in Begonia

Pei Zhang (  78073954@qq.com ) Jiangsu Meteorological Bureau https://orcid.org/0000-0002-4157-7231 zhengmeng chen Longyan Company of Fujian Tobacco Corparation Fuzheng Wang Qingengren Modern Agricultural Science and Technology Development(Huan'an) Co., Ltd. Rong Wang Institute of Botany Jiangsu Province and Chinese Academy of Sciences Tingting Bao Jiangsu Meteorological Bureau Xiaoping Xie Jiangsu meteorological bureau Renzu Zhang Xuzhou Meteorological Bureau Ziyue An Carleton University Chunwei Liu Nanjing University of Information Science and Technology haidong jiang Nanjing Agricultural University


Background
Meteorological conditions are crucial factors that determine the crop growth and leaf development (Zhou et al., 2020;Nawaz R. et al., 2020). Facility agriculture is an e cient crop production through the precise regulating of microclimate within the greenhouse (Kempkes et al., 2017;Amani et al., 2020;Guo et al., 2021). Clarifying the relationships between plant physiological and ecological indicators and meteorological conditions are the basis method for quantitative assessing of agro-meteorological conditions, assessing the agro-meteorological disaster losses, predicting the yield, and applying the intelligent micrometeorology control of agricultural production. Thus, it is the right path to nd the determine parameters that have closely relationship with meteorological factors and re ect plant growth status sensitively.
The digital image technology is developed and the high-resolution camera equipment is widely used, the digital imaging technology has obvious advantages of high resolution and low cost in research of plant phenotype He et al., 2017). Digital color image is not only convenient to obtain (Sancho-Adamson et al., 2019), but also contains abundant information about plant morphology, structure, and color gradation Grosskinsky et al., 2018;Vasseur et al., 2018), which can re ect the internal status of plants. RGB color model is the popular used color representation for digital images (Barker et al., 2016). The parameters of traditional RGB model are mainly the mean values of each color channel and their combined values, which can approximately describe the color depth of the leaves and estimate the chlorophyll content of the leaves (Dey et al., 2016Yadav et al., 2010Adamsen et al., 1999;Hu et al., 2010;Ali et al., 2012;Han et al., 2014). Moreover, the RGB model re ects the soil moisture and nutrient level (Barbedo et al., 2019Humplík et al., 2015aBresson et al., 2018. However, the mean values of image parameters is not comprehensive enough in the description of color distribution (Li et al., 2014), as a result, RGB model is generally used in the studies of water, fertility and disease which may drastically change the leaf color (Neilson et al., 2015;Bai et al., 2018;Gracia-Romero et al., 2017;Sancho-Adamson et al., 2019). The RGB model is also an effective method to re ect the relationship of image information and the variation of the environmental factors.
The quantitative analysis of mega-high quality digital photographs could explain the genotype-by-environment interactions in plant phenotype (Tester M, 2010). The variation of meteorological factors affect the physiological and biochemical characters of plants slightly, such as in uencing the openness of stomata (James, 2000), changing the arrangement of the chloroplast (Min-Wha et al., 2006), and increasing or decreasing the photosynthesis rate of the leaves (Urban et al., 2017;Wang et al., 2020). All these affections change the plant color under visible spectrum (Gitelson, et al., 2003). Therefore, the image color information parameters are used as representative references to re ect meteorological environment of the growing plant. Skewed analysis expanded the RGB model parameters from the mean to the mean, median, mode, skewness and kurtosis, with the number of parameters from 4 to 20 (Chen et al., 2020). The color gradation skewness-distribution (CGSD) parameters describe leaf color information more accurately and comprehensively than the traditional mean parameters (Chen et al., 2020). Whether the color gradation distribution of single plant varies the same as the leaf, and the eld scale determines its sensitivity on the response of environmental meteorological factors. The CGSD for all scales varying consistently helps to construct the effctive model between the external meteorological factors and morphological parameters of plant. The quantitative evaluation of agro-meteorological conditions for plant can be achieved by combining canopy color information indexes corresponding to different morphological parameters of plants, so as to provide information supported precise management of plant production.
Begonia Fimbristipula Hance (BFH) is a vegetation type widely used for medicine and food. BFH has abundant anthocyanins, polyphenols and avonoids and a strong antioxidant capacity (Fukuoka et al., 2014;Ren et al., 2014;Shimizu et al., 2010a). BFH is cultivated in facilities which has warm and humid environment to ensure annual cultivation (Wu et al., 2015). After planting BFH in glass greenhouse and collecting the meteorological factors, canopy photographs, the relationship between changes in canopy color and the meteorological factors in facility environment was anlyzed. The purposes of this research are (1) quantifying the mathematical relationship between CGSD parameters and meteorological factors, and increasing the accuracy of quantifying the changes in population canopy color with meteorological changes; (2) constructing population canopy color -meteorological response models by using the relationship of CGSD parameters and meteorological factors, including temperature, humidity and solar radiation; (3) providing new methods from the RGB model to monitor the patterns of canopy color varies because of meteorological environment changes. The results may construct the crop growth model based on the CGSD parameters of RGB model, and provide a new approach for intelligent control of facility agricultural production.

Plant material and Growth conditions
The glass greenhouse for the experiment is in Guli Modern Agriculture Demonstration Park, Jiangning District, Nanjing, Jiangsu, China (24° 57'N, 116° 30'E).
The BFH were potted by rooting of cutting and planted with total nutrient matrix which was the proportion of N: P 2 O 5 : K 2 O in 100g matrix is 20:10:20, trace elements include Cu, Fe, Zn, Mn, B and Mo ≥1%, and pH 5.85. Moreover, the matrix contains active accelerator and sustained release agent of poly-fertilizer. The BFH experiment has 2 replicates, and each replicate contains 24 pots plant. The rst growing period of BFH is December 1, 2018 to February 17, 2019, and after cutting off the 10cm length top on February 18, BFH is continued growing at the second growth period from February 19, 2019 to April 11, 2019.

Meteorological Data Acquisition
Meteorological data were obtained from the agricultural meteorological observation station (DZZ4, Jiangsu Provincial Radio Science Research Institute Co., LTD., China) in the facility. The observed meteorological elements include including daily meteorological elements, such as the daily mean temperature (T dm ), daily mean relative humidity (RH dm ), daily mean ground temperature (TG dm ), daily mean soil temperature of 10cm (TS dm−10c ), daily mean vapor pressure (VP dm ), daily mean dew point temperature (TD dm ), daily total global radiation(GR dt ) and daily total photosynthetically active radiation (PAR dt ).
Accumulate temperature (AT), accumulate global radiation (AGR) and accumulate photosynthetically active radiation (APAR) were calculated from T dm , GR dt and PAR dt . Canopy color images of BFH were collected in facility by monitoring camera (DH-SD-65F630U-HN-Q, Zhejiang Dahua technology Co., Ltd, China) with an image resolution of 1920×1080, and the installation height was 280cm. The camera with vertical lens was set with xed focal length shooting and automatic white balance and adopted xed time shooting mode to take a photograph at set times of 9:03am every day. The 117 images from 3 groups without direct sunlight were selected for analysis, including 102 images in the rst growth period (December 1, 2018 to February 17, 2019) of 2 replicates with 51 pictures each, and 15 images in the second growth period (February19, 2019 to April 11, 2019).
Cutting and Denoising of the Image Adobe Photoshop CS software (San Jose, CA, USA) and MATLAB2016R software (referred to as MATLAB [Math Works, Natick, MA, USA]) was primarily used to cut and denoise the BFH original image.
1. Adobe Photoshop CS software (San Jose, CA, USA) was primarily used to intercept the range of 600*600 in the lower left corner of the image, and the processed image was save as JPG image format ( Fig.4-1).
2. The rgb2hsv function of MATLAB was used to convert RGB images into HSV images. Double cycle operation was used to set H value of the image background to 0 (i.e. black), while H value of the plant remains unchanged. The hsv2rgb function was used to convert the processed HSV image into RGB image ( Fig.4-2).
3. Double cycle operation of MATLAB was used again to lter the color threshold value of the image processed in the previous step. The color opacity of the black part of the image was adjusted to 0 (that is, completely transparent), and the color image of the target leaf or canopy with transparent background was saved as a PNG image mode ( Fig.4-3).

Information Collection of the RGB Image
Transformation of double precision arrays of image. MATLAB was used to extract and analyze BFH RGB images. After reading color images by using the imread function, the image (:, :, 1), image (:, :, 2), image (:, :, 3), rgb2gray function were respectively used to read every pixel gradation of red, green and blue channels, as well as gray-level images. The full circulation algorithm was used to retrieve and record the non-black pixel index codes of these pixels, which were combined to an array of leaf color gradation. Then the double function was used to transform it into double precision arrays again.
Establishment of canopy color gradation skewness-distribution (CGSD) parameters table. The mean, median, mode, std, skewness and kurtosis functions were used to acquire the mean median mode standard Deviation skewness and kurtosis of the double precision arrays of red, green and blue channels, as well as gray-level images (Chen et al., 2020). The CGSD parameters were obtained, including R Mean , R Median , R Mode , R Skewness , R Kurtosis , G Mean , G Median , G Mode , Array distribution normality testing. The lillietest and jbtest functions were used to conduct the Lilliefors and Jarque-Bera tests of normal distribution for the color gradation distribution of red, green and blue channels, as well as gray-level images, of BFH RGB images.

Prediction model construction and goodness of t detection
Correlation Analysis of 20 CGSD Parameters to Microclimate Factors. Cor package of R was used to analyze the relationship between 20 CGSD parameters of population canopy RGB images of BFH and the corresponding microclimate factors, including daily meteorological factors (T dm , RH dm , TG dm , TS dm-10c , VP dm , TD dm , GR dt and PAR dt ) and photothermal accumulation factors (AT, AGR and APAR), with double tail inspection were collected for signi cant examination.
Linear prediction models establishment. By using R, the linear prediction models (Y1-Y20) of 20 CGSD parameters of population canopy RGB images of BFH were established by a regression approach based on the least-square method, with meteorological factors as the independent variables. Probability of F-to-enter of the models was set to 0.050 or less, while Probability of F-to-remove was set to 0.100 or more. Decision coe cient optimization, signi cance test of regression model and regression coe cient, collinearity diagnosis of independent variables of the regression model were conducted for the alternative regression models in turn (Gai, 2000), and the optimal regression models were nally determined.
Spatial polynomial model establishment. 2 photothermal accumulation factors with the correlation coe cient of highest absolute value with CGSD parameters were selected as the independent variable, and the Spatial polynomial models (Z1-Z4) of skewness of red, green and blue channels, as well as gray-level images were established by using polynomial tting in Curve Fitting Tool of R.

Skew analysis of the distribution of leaf color gradation of the RGB images
After analyzing the cumulative distribution of color gradation of the 117 selected images of BFH, we found that all the red, green, blue channels, and the gray-level for single leaf, single plant and population canopy images showed a left skewed distribution (Fig. 3). Lilliefors and Jarque-Bera normality tests also indicated that color gradation data did not distribute normally (Supplementary Table S1). 20 color gradient skewed distribution (CGSD) parameters were obtained including the mean, median, mode, skewness and kurtosis for the red, green, blue channels, as well as the gray-level image, for each canopy image, respectively. These parameters can describe not only the depth of canopy color but also its distribution.
The main distribution of color gradation is positive skewness for RGB channels and the gray-level image near the value of 0. This increment of positive skewness as the canopy developing re ected that the leaf color transit from light green to dark green (Fig. 4).

Correlation Analysis Microclimate Factors in Glasshouse and Population Canopy CGSD Parameters
The relationship among 20 CGSD parameters and the microclimate factors including daily meteorological factors (T dm , RH dm , TG dm , TS dm-10c , VP dm , TD dm , GR dt and PAR dt ) and accumulative factors (AT, AGR and APAR) were shown in Fig. 5. Fig. 5 indicate that temperature, humidity, ground temperature, soil temperature of 10cm and solar radiation were signi cantly correlated with CGSD parameters. The accumulated temperature and solar radiation (AT, AGR and APAR) had extremely signi cant correlation with 20 CGSD parameters, with the correlation coe cients higher than 0.8. The parameters mean, median, and mode which represented color depth showed signi cant positive correlation with the meteorological factors, while the parameters skewness and kurtosis represented color uniformity showed obviously negative correlation.

Multiple linear relationships of the Microclimate Factors and Population Canopy CGSD Parameters
Multiple linear relationships between population canopy CGSD parameters and microclimate factors were established by stepwise regression method using the ordinary least square method (OLS) ( Table 1). Table 1 showed that all the multivariate determination coe cient (R 2 ) of the population canopy colormeteorological response models equation were greater than 0.7, indicating that the equation well explained the response of BFH population canopy CGSD parameters to microclimate factors. The R 2 of the red and green channels, as well as the gray-level image, was 0.692-0.954, which was better than those of blue channel with 0.674-0.791. The model between the mean of green channel and accumulated temperature, daily mean relative humidity, and the daily total photosynthetically active radiation has the highest R 2 0.954. Moreover, all the models was signi cant (p<0.001).
All the models were veri ed by between-group and outside-group samples ( Table 2 and Figure 6-8) , which indicated that the population canopy color tting models performed well in predicting the mean, median, mode, and the kurtosis. The accuracy of prediction of the skewness was about 80%, while other CGSD parameters exceeded 90%, with the mean of blue channel, as well as the gray-level image exceeded 95% (Table 2) in the same modeling group samples. Similarly, the accuracy of prediction of the mean, median and mode of RGB channels and the gray-level were 90%, and the kurtosis was about 75%, the skewness 53-58% in Between-group prediction. The accuracy of prediction of other CGSD parameters was 80% with the exception of the skewness and kurtosis of red channel, the mode and skewness of green channel, and the median of gray-level image in outside-group. Moreover, the accuracy of prediction of the kurtosis of green, blue channels and gray-level image exceeded 85%.
Optimization of canopy color skewness-meteorological models Although the skewness of the red, green channels, and the gray-level image perform well, the between-group and outside-group prediction accuracy were generally lower than other CGSD parameters. The skewness is non-linear to microclimate factors (Chen et al., 2020). Therefore, we t the polynomial Z1-Z4 that incorporates spatial bidirectional patterns for the RGB channels, and the gray-level image, by taking 2 accumulation factors (the accumulate temperature (x 1 ), the accumulate photosynthetically active radiation (x 5 ) as the independent variables, which had the highest absolute value of the correlation coe cient.
The spatial bidirectional models for the skewness has 5.7 percentage points higher accuracy in prediction, and the number of outliers signi cantly decreased. Compared to the regression model, the average accuracy of the spatial bidirectional models was improved by 7.63, 5.75 and 6.25 percentage points in the modeling group, between-group and outside-group, respectively. The RGB channels and gray-level image was all improved by 6.90, 10.87, 3.47 and 1.72 percentage points, respectively.

Discussion
The high quality and e cient production depend on the environment regulation focusing on the plant growth in facility. The changing physiological parameters are result from variation of temperature, radiation, humidity rapidly and instantaneously (Hand, 1988;Shamshiri et al., 2018Li at al., 2018Chand Singh et al., 2018. Non-effectively environmental factors controlling restrict the improvement of glass greenhouse yield and crop quality (Zhang et al., 2020). Therefore, Coupling the relationship between plant growth state indicated by color parameters and meteorological factors controlling help promoting the high quality and e ciency production in facility plants, realizing the precise controlling of facility environment and decreasing the cost.
Digital color images information can re ect the growing state of plants (Vasseur et al., 2018), which provides a convenient way of plant growth monitoring. Researches on conducting qualitative and quantitative descriptions of phenotypic traits for plant appearance by using digital imaging technology are increased on smart agriculture . The RGB color model has been widely used to process digital color images to study chlorophyll content and its related plant nutritional status and stress response ( Zhu et al., 2019;Neilson et al., 2015;Chaerle et al., 2007). The RGB model and leaf color parameters can describe changes in tobacco leaf color depth and homogeneity (Chen et al., 2020). Now we found that images of the RGB channels, and the gray-level all showed a skewed distribution at the single leaf, plant and population canopy images scale in BFH.
Environmental changes lead to changes in the color of plants leaves (Humplik et al., 2015b;Gous et al., 2015;Cai et al., 2016;Schmalko et al., 2005;Gracia-Romero et al., 2017). Our work indicated that temperature, humidity, ground temperature, soil temperature of 10cm and solar radiation in glass greenhouse were signi cantly correlated with multiple CGSD parameters of canopy images in BFH. Moreover, the accumulative value of temperature and solar radiation (AT, AGR and APAR) had extremely signi cant correlation with 20 CGSD parameters, with the correlation coe cients beyond 0.8. The mean, median, and mode parameters representing color depth has positive relationship on the accumulative parameters, while the skewness and kurtosis parameters representing the color distribution showed negative correlation with the accumulative factors signi cantly. The deeply canopy color of BFH images has higher accumulative factors depending on the analysis of the peak changes of the cumulative distribution diagram of canopy color and providing a new approach to quantitative the plant growing conditions.
The canopy color -meteorological responding models have high predict accuracy after calculating the image mean, median, mode and kurtosis parameters and skewness distribution using the spatial bidirectional models. The change of accumulative factors is generally linear, while skewness is calculated from higher-order function. Although the skewness was signi cantly related to microclimate factors, it cannot be simply described and tted by linear tting (Chen et al., 2020). Comparing to the multiple stepwise regression models, the spatial bidirectional models had better tness and higher accuracy.
Digital color image has been applied to the automatic observation in agrometeorology (Shibayama et al., 2011;Sritarapipat et al., 2014), and proposed as a new technical approach that can be combined with crop growth models (Sun and Shen, 2019). Agriculture meteorological services are automatically and intelligently in future services (Yu et al., 2013;Li et al., 2016). The relationship between the CGSD parameters of the population canopy and the environment meteorological factors makes it possible to use the canopy color information to construct a plant growth model from digital images, which can provide new methods for quantitative evaluating agricultural meteorological conditions, assessing agricultural meteorological disaster loss, predicting yield and quality, and supporting precise management of crop production.

Conclusion
Analyzing the canopy images color histograms skewed distribution of BFH in glasshouse, we found 20 population canopy CGSD parameters were sensitive to the microclimate factors, especially the radiation and temperature accumulation factors. The canopy color -meteorological responding models showed that the CGSD parameters have close relationship with microclimate factors, which expands the potential application of the RGB model in monitoring the variations of plant or leaf color. The CGSD parameters can describe the canopy color information comprehensively and accurately, and consider as a new indicator for quantitative evaluating the plant growth conditions. These parameters help constructing a plant growth condition model, and providing a new method for agrometeorological assessment and prediction and ultimately supporting the precise management of crop production.
Declarations Zhang, S. H., Guo, Y., Zhao, H. J., Wang, Y., Chow, D., Fang, Y., 2020 Note x 1 is the accumulate temperature, x 2 is the daily mean relative humidity, x 3 is the daily total photosynthetically active radiation, x 4 is the daily mean soil temperature of 10cm, x 5 is the accumulate photosynthetically active radiation, x 6 is the daily minimum ground temperature, x 7 is the accumulate global radiation. 9   RGB model color gradation distribution of single leaf, plant canopy and population canopy images of BFH. The cumulative frequency histogram of red, green and blue channels, as well as gray-level images, were drawn using the imhist function of MATLAB.. The X-axis is the color gradation value, and the Yaxis is the cumulative frequency, Figure 4 RGB model color gradation distribution of population canopy images of BFH on 3 different dates. The cumulative frequency histogram of red, green and blue channels, as well as gray-level images, were drawn using the imhist function of MATLAB. The X-axis is the color gradation value, and the Y-axis is the cumulative frequency, Figure 5 Spearman correlation coe cient of population canopy CGSD parameters of BFH and microclimate factors. A Spearman correlation analysis of SPSS software was used on 20 CGSD parameters (RMean, RMedian, RMode, RSkewness, RKurtosis, GMean,GMedian,GMode, GSkewness, GKurtosis, BMean, BMedian, BMode, BSkewness, BKurtosis, YMean, YMedian, YMode, YSkewness and YKurtosis) and the microclimate factors including daily meteorological factors (Tdm, RHdm, TGdm, TSdm-10c, VPdm, TDdm, GRdt and PARdt) and photothermal accumulation factors (AT, AGR and APAR) (n = 51). The correlation coe cient obtained by related analysis was drawn to CGSD parameters-daily meteorological factors heat map. The positive correlation coe cient is shown in red, and the negative correlation coe cient is shown in blue. ** indicates signi cant correlation according to a two-tailed test (p 0.01), * indicates signi cant correlation according to a two-tailed test (p 0.05), the same as below.

Figure 6
Prediction accuracy interval analysis of the population canopy color-meteorological response models of modeling group. The Boxplot function of MATLAB was used to draw the accuracy percentile distribution diagram of population canopy CGSD parameters prediction model. The blue box in the gure represents the prediction accuracy of the model distributed between 25% and 75%. The shorter the box, the better the degree of data concentration. The red line in the box represents the median distribution of prediction accuracy, and the dotted line at both ends outside the box represents the endpoint value (maximum and minimum) of the prediction data distribution. The shorter the distance between the two endpoint values, the better the degree of data concentration. The red-cross outside the box represents abnormal data. The fewer abnormal data, the better the degree of data distribution concentration.

Figure 7
Prediction accuracy interval analysis of the population canopy color-meteorological response models of between-group.

Figure 8
Prediction accuracy interval analysis of the population canopy color-meteorological response models of outside-group.