CN113866104A

CN113866104A - Method for detecting nitrogen nutrition status of greenhouse tomatoes based on digital camera

Info

Publication number: CN113866104A
Application number: CN202111135500.2A
Authority: CN
Inventors: 石小虎; 李超
Original assignee: Beijing Fuping Chuangyuan Agricultural Science And Technology Development Co ltd; Taiyuan University of Technology
Current assignee: Beijing Fuping Chuangyuan Agricultural Science And Technology Development Co ltd; Taiyuan University of Technology
Priority date: 2021-09-27
Filing date: 2021-09-27
Publication date: 2021-12-31

Abstract

The invention belongs to the technical field of vegetable detection, and provides a method for detecting the nitrogen nutrition condition of greenhouse tomatoes based on a digital camera in order to discuss the feasibility of nondestructive diagnosis of water nitrogen treatment of greenhouse tomatoes by a digital image technology. The method comprises the steps of obtaining color characteristic values of tomato leaf images of tomatoes treated by different water nitrogen, carrying out correlation analysis on the color characteristic values and plant nitrogen nutrition indexes, screening out nitrogen nutrition diagnosis indexes and diagnosis models of tomato canopy color parameters by combining detection of traditional nitrogen nutrition indexes, establishing a fitting model for inverting the tomato nitrogen nutrition indexes, determining nutrition diagnosis indexes and providing fertilization recommendation. The nitrogen nutrition index is obtained by rapidly inverting the digital index G/(R + G + B) of greenhouse tomato canopy images, and the plant nitrogen nutrition state is accurately diagnosed in time. The research can provide better theoretical reference for the real-time nitrogen nutrition diagnosis and the optimized nitrogen management of the greenhouse tomatoes.

Description

Method for detecting nitrogen nutrition status of greenhouse tomatoes based on digital camera

Technical Field

The invention belongs to the technical field of vegetable detection, and particularly relates to a method for detecting the nitrogen nutrition condition of greenhouse tomatoes based on a digital camera.

Background

Tomatoes are one of the most common vegetables cultivated all over the world and are widely cultivated all over the north of China. Meanwhile, the tomatoes are also one of the main characteristic economic vegetables for facility cultivation in northern areas of China. The nitrogen is used as a key factor for limiting nutrients and influencing yield of tomatoes, and excessive application of a farmland nitrogen fertilizer can not only reduce the yield of vegetables, but also reduce the agricultural economic benefit, and can also cause accumulation of nitrate nitrogen in soil, cause secondary salinization of the soil and hinder sustainable development of agriculture (begonia and millet calmness, 2014). Therefore, timely diagnosis, accurate fertilization and fertilization according to needs in the actual greenhouse vegetable cultivation process are of great significance to ecology, environment and economy.

Traditional nitrogen nutrition diagnosis mainly measures plant total nitrogen, and the methods are accurate in judgment and high in reliability, but the required sample size is large, plant sampling and laboratory chemical analysis are required, the requirements of monitoring crops at any time and field topdressing are difficult to meet, and the methods are difficult to popularize in vast rural areas (Lemaire et al, 1997; Jialiang, etc., 2009).

In recent years, modern instrument nitrogen diagnosis technology can be used for quickly, conveniently and nondestructively performing crop nitrogen diagnosis, such as hyperspectral remote sensing diagnosis, SPAD instrument diagnosis, digital image and computer vision diagnosis technology and the like. The high-light remote sensing technology is characterized in that a spectrum is divided in detail in a certain spectrum area, and further spectrum information of more bands is obtained, compared with multi-band remote sensing, the high-light remote sensing spectrum resolution is higher, Maderia research shows that the chlorophyll content in rice leaves and the spectrum characteristics have correlation, and the nitrogen nutrition condition of plants can be detected through the spectrum (Madeire et al, 2000), but the method has the defects that data acquisition needs to be repeatedly converted, influence factors are more, instruments are expensive and the like; portable chlorophyllates are widely used in rapid diagnostic studies of nitrogen in crops, but the use of chlorophyllates requires multi-point measurements to reduce variability (Blackmer and Schepers, 1995), and the variety and growth period of the crop have a large impact on the results of the measurements (zhao manxing et al, 2005), and the chlorophyllate fails to respond to chlorophyll content when the substance is near or above the optimum nitrogen application (Liu et al, 2010); the digital image diagnosis technology carries out plant nitrogen nutrition diagnosis and nitrogen fertilizer recommendation by acquiring red, green, blue and other image characteristic values of plant canopies, has the characteristics of simplicity and no damage (Jialiang, 2003), and has better development space along with the improvement of science and technology and the reduction of the price of a digital camera (Jialiang, etc., 2009; Borhan et al, 2004).

The leaves are important nutritional organs of crops, and the color of the leaves has strong sensitivity to the loss of fertilizers (Scharf and Lory, 2002) and moisture (Hendrawan and Murase, 2011), so the technology of performing nitrogen nondestructive testing by using digital images of the leaves is widely used for evaluating and researching the nitrogen nutritional condition of the crops. Numerous studies indicate that significant or extremely significant correlation exists between the color parameters of crop canopy digital images and indicators such as chlorophyll meter readings (Xiaoyan wave, etc., 2008), NDVI (Rouse et al, 1974; Jialiang, 2003), overground biomass (Xiaoyan wave, etc., 2008), total nitrogen absorption (Zhangzhou, etc., 2010), and the like. Zhang Zhou et al (Zhang Zhou et al, 2010, 2011) studied the correlation between color parameters of images of canopy of winter wheat and summer corn obtained based on a digital camera and traditional nitrogen nutrition indexes, and screened out green light standardized values and blue light standardized values as sensitive color parameters for diagnosing the nitrogen nutrition states of the green light standardized values and the blue light standardized values. The Wanjuan and the like (2008) and the Wuwei Quanquan and the like (2015) both consider that the characteristic value of the color of the leaves has better correlation with the nitrogen content of the leaves, and have feasibility for predicting the nitrogen nutrition status of crops. Lee et al (2013) consider that indexes such as red (R), green (G), blue (B) and the like of rice canopy images are significantly related to overground biomass, leaf area index and overground nitrogen accumulation of plants. Jialiang et al (2003) adopt low-altitude photography on the ground and near ground to obtain digital images of wheat and corn, extract various spectral indexes and plant nitrogen nutrition indexes, and perform correlation analysis to find that the red light standardized value and the nitrogen index are in a negative correlation relationship, and the green light standardized value and the blue light standardized value are in a positive correlation relationship.

Xugui power et al (2002) utilize digital image technology to identify whether the tomato seedling stage leaves are deficient, find that the blue light component standard deviation in the GRB color model can distinguish normal tomato leaves from potassium deficient tomato leaves, and the correlation coefficient R/G of red light component and green light component can separate normal leaves from nitrogen deficient leaves and potassium deficient leaves. Huchunhua et al (2004) utilize digital image technology to perform tomato deficiency diagnosis research, color feature analysis is performed on color images of nitrogen deficiency, magnesium deficiency, iron deficiency and normal tomato leaves in GRB and HSI color models respectively, and it is found that standard deviations of G/R and G/B can well distinguish magnesium deficiency, nitrogen deficiency and normal tomato leaves, and correlation coefficients of G/R and G/B can distinguish magnesium deficiency, nitrogen deficiency and normal tomato leaves. The research of plum blossom (2007) finds that the correlation between the G/B and R/B characteristics of leaf images and the nitrogen content of leaves is remarkable in the whole growth period of tomatoes, and the leaf images can be used as a reliable basis for diagnosing the nitrogen nutrition level of the tomatoes. Wu Xuei et al (2004) found that the difference between the mean value of the G component and the mean value of the H component between normal tomato leaves and nitrogen-deficient tomato leaves was large. Studies in the tengyun et al (2012) found that the color characteristics of tomato leaves at maturity were significantly correlated with total nitrogen content.

At present, research on tomato nitrogen nutrition status based on digital image technology (Liyan, 2007; Wuxue Mei, 2004; Dingyongjun et al, 2012) is available, however, the relationship between tomato digital image characteristic values and crop nitrogen nutrition indexes is different, most of research focuses on nitrogen nutrition diagnosis and recommended fertilizing amount under a single moisture condition, and research on facility tomatoes nutrition diagnosis by using digital camera digital image technology under different moisture conditions is less.

Disclosure of Invention

The invention provides a method for detecting the nitrogen nutrition condition of greenhouse tomatoes based on a digital camera, aiming at discussing the feasibility of nondestructive diagnosis of greenhouse tomato water nitrogen treatment by a digital image technology. The color characteristic value of the tomato leaf image is obtained through different water nitrogen treatment tests for 2 years, correlation analysis is carried out on the color characteristic value and plant nitrogen nutrition indexes, a suitable tomato nitrogen level rapid diagnosis parameter is screened out by combining detection of a traditional nitrogen nutrition index, a fitting model for inverting the tomato nitrogen nutrition index is established, the nutrition diagnosis index is established, fertilization recommendation is provided, and a theoretical basis is established for nondestructive diagnosis of the tomato nitrogen level.

The invention is realized by the following technical scheme: a method for detecting nitrogen nutrition status of greenhouse tomatoes based on a digital camera comprises the steps of obtaining color characteristic values of tomato leaf images of tomatoes treated by different water nitrogen, carrying out correlation analysis on the color characteristic values and plant nitrogen nutrition indexes, screening out nitrogen nutrition diagnosis indexes and diagnosis models of tomato canopy color parameters by combining detection of traditional nitrogen nutrition indexes, establishing a fitting model for inverting the tomato nitrogen nutrition indexes, determining nutrition diagnosis indexes and providing fertilization recommendation.

The different water nitrogen treatments of the tomatoes are as follows: the water treatment is set to 3 levels, and the water filling amount is respectively full water filling treatment in the whole growth period, 25% of water deficiency in the whole growth period and 50% of water deficiency in the whole growth period; the nitrogen treatment was set at 4 levels with nitrogen application rates of 0, 150, 300 and 450 kg/hm, respectively²。

Specifically, green light standardization value G/(G + R + B) and nitrogen nutrition index are utilizedNNIThe correlation relationship between:NNI=81.59(G/(G+R+B))²65.51(G/(G + R + B)) +13.99, when 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNIif the nitrogen nutrition index is more than 1, the nitrogen nutrition index is obtained through inversion by obtaining the greenhouse tomato canopy image digital index G/(R + G + B) and the plant nitrogen nutrition state is accurately diagnosed in time.

The specific method comprises the following steps:

(1) observation item and data acquisition:

detecting the water content of the soil: sampling at a position 20 cm away from the plants by using a soil drill, respectively measuring the soil moisture content of each wide row, each narrow row and the soil moisture content among the plants, measuring for 1 time before and after irrigation, measuring for 1 time every 15 cm from the surface layer to the depth of 60 cm, and taking the average value during calculation;

measurement of irrigation quantity: the irrigation starts from 15 d after the field planting, the upper limit of irrigation is field water capacity after the full irrigation treatment (θ _FC) 90% of it, it is pouredAmount of waterI(mm) is:I=10(0.9θ _FC−θ _i)Z _r(ii) a In the formula:θ _iis the water content of the soil before irrigation, cm³/cm³；Zr is the planned wetting layer depth of 60 cm;

tomato biomass determination: sampling every 20 days after seedling setting, and taking 3 plants each time; weighing fresh tomato stems, leaves, fruits and roots in each sampling, drying at 105 ℃ for 15 min for de-enzyming, drying at 72 ℃ to constant weight, and calculating the biomass;

and (3) measuring the nitrogen content of the plant: pulverizing the treated dry materials, sieving, and adding H₂SO₄-H₂O₂Measuring the total nitrogen content of each organ by a digestion method and a Kjeldahl azotometer, and calculating the total nitrogen content of the plant; cumulative nitrogen amount of each organ kg/hm²= organ nitrogen content% × organ biomass kg/hm²Adding the nitrogen accumulation of all organs to obtain the nitrogen accumulation of the overground part plant; plant nitrogen content (%) = plant nitrogen accumulation kg/hm²Plant biomass kg/hm²；

(2) Nitrogen nutrition indexNNIAnd (3) calculating: critical nitrogen concentration dilution curve model and nitrogen nutrition indexNNIThe equation of (1) is:N _c=a·DW^-b；NNI=N _t/N _c(ii) a In the formula:N _cis the critical nitrogen concentration value, g/(100 g); the parameter a is that the biomass of the overground part of the crop is 10³ kg/hm²Critical nitrogen concentration of the plant; DW is the maximum value of biomass on the overground part of the crop, 10³ kg/hm²(ii) a b is a statistical parameter for determining the slope of the critical nitrogen concentration dilution curve;NNIis the nitrogen nutrition index;N _tg/(100 g) as an observed value of the nitrogen concentration of the aboveground biomass;

canopy image processing and color parameter analysis: the digital image of the canopy of the tomato is obtained by shooting with a camera, clear weather without clouds at 12:00-14:00 at noon is selected, the tomato canopy is vertically downward and overlooked along the light, an included angle of 30-60 degrees is formed between the tomato canopy and the canopy, each test area is shot, the shooting height is 30cm away from the canopy, a forced exposure mode is set, the resolution of the camera is 2084 multiplied by 1536 pixels, the tomato canopy is obtained simultaneously with biomass in 30, 50, 70, 90, 110, 130 and 150 days after transplanting, and each test area shoots 2-3 frames at each time and is stored in a JPEG format;

importing the picture into a computer, acquiring a color G, R, B characteristic value of the blade by utilizing Photo-shop and Matlab software, and averaging; eliminating interference factors, only leaving color pixels of normal tomato leaves, then calculating the standardized value of G, R, B color parameters of each pixel, introducing visible light atmospheric impedance vegetation indexes [ (G-R)/(R + G + B), VARI ], and finally carrying out average value statistics of the color parameters of the whole image and screening color parameter indexes for crop nitrogen nutrition diagnosis;

(3) respectively selecting a linear equation and a quadratic equation as an evaluation prediction equation model, wherein regression equations of the G/(G + R + B) characteristic value, the nitrogen concentration, the SPAD value, the plant biomass and the nitrogen nutrition index are respectively y (G/100G) =14.51x-3.95, y = -111.90x +86.51 and y (t/hm)²) = 74.04x +33.97 and y =81.59x²-65.51x+13.99；

When in useNNI=1, nitrogen nutrition status is most suitable; when in useNNI>1, indicated as nitrogen over-nutrition; when in useNNI<1, manifesting as nitrogen nutrient deficiency;

(4) establishing the characteristic value of G/G + R + B and the nitrogen nutrition index of the tomato canopyNNIThe relationship between:NNI=81.59(G/(G+R+B))²65.51(G/(G + R + B)) +13.99, determining the equation for the linear monitoring of the nitrogen nutrition with the independent variable of the characteristic values of the tomato canopy G/G + R + B, when 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNImore than 1, nitrogen is in excess.

According to the invention, according to the test data of different water nitrogen treatments of greenhouse tomatoes in 2 years, a method for diagnosing the nitrogen nutrition of greenhouse tomatoes by using a digital camera is researched, the relation between the digital index of an image analysis canopy image and the nitrogen nutrition condition is preliminarily established, the nitrogen nutrition diagnosis index and the diagnosis model of the color parameter of the canopy of tomatoes are screened and provided, and the result shows that: tomato canopy green value (G), red value (R), blue value (B), green standardized value (G/(G + R + B)), blue standardized value (B/(G + R + B)) and leaf SPAD value, plant nitrogen concentrationHas a significant or extremely significant positive correlation with biomass; green light value (G), red light value (R), blue light value (B), green light normalized value (G/(G + R + B)), red light normalized value (R/(G + R + B)), visible light atmosphere resistance vegetation index (VARI), leaf SPAD value and nitrogen nutrition index ((G + R + B)), and the likeNNI) The plant biomass is in a significant or extremely significant negative correlation relationship; compared with other canopy digital image parameters, the green light standardization value (G/(G + R + B)) can better reflect the nitrogen nutrition status of the tomato.

Therefore, the standardized value of tomato canopy green light in 2017 and 2018 (G/(G + R + B)) and the SPAD value, the plant nitrogen concentration and the nitrogen nutrition index (NNI) And biomass, fitting the relationships of y = -111.90x +86.51, y (g/100g) =14.51x-3.95 and y =81.59x²65.51x +13.99 and y (t/hm)²) = -74.04x + 33.97. The fitting equation is tested by using 2018-year 2019 tests, the determination coefficients between the actual value and the predicted value are respectively 0.62, 0.68, 0.65 and 0.82, the absolute errors are respectively 7.73, 0.35, 0.10 and 1.18, the root mean square error is respectively 9.36, 0.47, 0.14 and 1.87, and the accuracy of the prediction result is good, so that the method can be used for testing the SPAD value, the plant nitrogen concentration and the nitrogen nutrition index of the tomato leaves (SPAD value, the plant nitrogen concentration and the nitrogen nutrition index: (1)NNI) And biomass were accurately estimated. Will be provided withNNI=1 as standard for appropriate nitrogen application using green normalisation value G/(G + R + B) and nitrogen nutrition index (G: =1 as standard for appropriate nitrogen gas is no: 1 by green standardized with green light and bothNNI) The correlation relationship between the two is that when G/(G + R + B) < 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNImore than 1, nitrogen nutrition is excessive; therefore, the nitrogen nutrition index is obtained by quickly inverting the digital index G/(R + G + B) of the greenhouse tomato canopy image, and the plant nitrogen nutrition state is accurately diagnosed in time. Can provide better theoretical reference for the real-time nitrogen nutrition diagnosis and the optimized nitrogen management of greenhouse tomatoes.

Drawings

FIG. 1 is a graph showing the dilution curve of the nitrogen concentration of the aboveground biomass of tomatoes treated with different water contents;

FIG. 2 shows the dynamic change of nitrogen nutrition index of tomatoes treated by different water nitrogen in 2017-2019;

FIG. 3 is a regression analysis of the standard value (G/(G + R + B)) of green light of tomato canopy and nitrogen nutrition index in 2017 and 2018;

FIG. 4 is the correlation between the predicted value and the measured value of the tomato nitrogen nutrition index using G/(G + R + B) as the independent variable.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments; 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.

In order to investigate the feasibility of the digital image technology for carrying out nondestructive diagnosis on the greenhouse tomato water nitrogen treatment, the following experiments are carried out:

1. materials and methods

Summary of test area: the test was carried out in 2018-2019 in a greenhouse (39 ℃ 42 'N, 113 ℃ 41' E) in muddy county of Shanxi province. The test greenhouse is a traditional non-heating natural ventilation greenhouse, the main body of the test greenhouse is of a steel frame structure and is covered by a plastic film, the directions of north and south (length, width and height, 55 m, 6.5 m, 4.6 m) and south are respectively provided with a 1 m wide ventilation opening at the top and the bottom of the greenhouse, a manual opening and closing device is arranged, and the temperature in the greenhouse is adjusted by opening or closing the ventilation openings. The tomatoes are planted in the greenhouse in the north-south direction. In 0-60 cm of soil in a greenhouse, 23% of sticky particles (less than 2 microns), 54% of powder particles (2-20 microns), 23% of sand particles (more than or equal to 20-2000 microns), 3.44% of organic matter mass fraction and 1.43 g/cm of volume weight³Saturated water content theta_SATIs 0.43 cm³/cm³Water capacity in the field theta_FCIs 0.36cm³/cm³Withering water content theta_WPIs 0.16 cm³/cm³。

2. And (3) experimental design: 2 factors were designed: moisture and nitrogen. 3 irrigation levels: full water filling treatment in the whole growth period (W)₁) 25% of water deficiency in the whole growth period (W)₂) 50 percent of water deficiency in the whole growth period (W)₃) (ii) a 4 nitrogen levels, N respectively₀（0 kg/hm²）、N₁₅₀（150 kg/hm²）、N₃₀₀（300 kg/hm²) And N₄₅₀（450 kg/hm²) Each test treatment was repeated 3 times for a total of 36 cells, each cell having an area of 6 m × 2.2 m =13.2 m²The cells are completely randomly arranged and are isolated by a plastic film with the buried depth of 60 cm.

Adopting drip irrigation mode under film, irrigation begins from 15 d after field planting, irrigation period is about 4 d, and setting full irrigation treatment (W)₁) The upper limit of irrigation is 90% of the field water-holding rate. According to the study of the rocky tiger (2013), the full irrigation treatment (W) is calculated₁) Amount of water to be poured in each growth periodIOther water treatment is only performed on the irrigation quantity, and the irrigation time and the irrigation times are equal to W₁The treatment is the same; high nitrogen levels (N) according to the Gekko Swinhonis (2013) study₃₀₀) Set at 300 Kg/hm²(in terms of N), N₀、N₁₅₀And N₄₅₀The treatment was only a 50% reduction or increase in nitrogen application. The nitrogen fertilizer is urea (containing 46% of nitrogen by mass), 40% of the nitrogen fertilizer is applied before planting, and the rest 60% of the nitrogen fertilizer is dissolved in water after planting (DAYS after transfer, DAT) for about 60 days, 80 days and 100 days respectively and is evenly applied along with drip irrigation.

The variety for the test is tomato 'zidary', the tomato growth period is divided into a seedling stage (2017-08-10-10-10, 2018-08-20-10-15), a flowering stage (2017-10-11-12-15, 2018-10-16-12-18) and a mature stage (2017-12-16-2018-02-05, 2018-12-19-2019-02-10). The planting mode is a typical local ridging and film mulching cultivation mode, the ridge height is 20 cm, the ridge width is 80 cm, the tomato seedlings are planted on two sides of the ridge according to single hole and single plant, the wide row spacing is 60 cm, the cross-ditch narrow row spacing is 40 cm, the plant spacing is 40 cm, and the planting density is 5 plants/m². The same amount of phosphate fertilizer 200 kg/hm is evenly applied in the greenhouse before field planting²Calculated by P and 300 kg/hm of potash fertilizer²(in K).

And (3) irrigating planting water for 20 mm during planting, not irrigating water within 14 d after planting, and performing test treatment after the tomato seedlings survive. And (3) paving a mulching film with the width of 1.2 m and the thickness of 0.005 mm along the north-south direction of the greenhouse on the planting day, suspending the tomatoes on an iron wire above the greenhouse by using a thin rope after the tomatoes bloom, carrying out artificial pollination for 1 time every 3 d, and simultaneously carrying out pesticide spraying and other agricultural management. In the whole growth period, 4 spikes of each tomato are left for pinching, the tomato is picked for 1 time every 2-3 days after being ripe, and other farming management is carried out according to local conventions.

3. Observation item and method

A. And (3) soil moisture content: sampling at a position 20 cm away from the plants by using a soil auger, respectively measuring the water content of the soil in wide rows, narrow rows and among the plants in each cell, measuring for 1 time before and after irrigation, measuring for 1 time every 15 cm from the surface layer to the depth of 60 cm, and taking the average value during calculation.

B. The irrigation quantity is as follows: the irrigation starts from 15 d after the field planting, the upper limit of irrigation is field water capacity after the full irrigation treatment (θ _FC) 90% of water, the amount of water pouredI(mm) is:I=10(0.9θ _FC−θ _i)Z _r(ii) a In the formula:θ _iis the water content of the soil before irrigation, cm³/cm³；Zr is the planned wetting layer depth, cm, and the invention takes 60 cm.

C. Tomato biomass: destructive sampling is carried out every 20 days or so after seedling setting, and 3 plants are taken every time. Weighing fresh tomato stems, leaves, fruits and roots in each sampling, drying at 105 ℃ for 15 min for deactivation of enzymes, drying at 72 ℃ to constant weight, and calculating the biomass.

D. And (3) measuring the nitrogen content of the plant: pulverizing the treated dry materials, sieving, and adding H₂SO₄-H₂O₂The total nitrogen content of each organ is measured by a digestion method and a Kjeldahl apparatus (FOSS 2300 type), and the total nitrogen content of the plant is calculated. Cumulative amount of nitrogen in each organ (kg/hm)²) = organ nitrogen content (%). times organ biomass (kg/hm)²) And adding the nitrogen accumulation of all organs to obtain the nitrogen accumulation of the overground part plants. Plant nitrogen content (%) = plant nitrogen accumulation (kg/hm)²) Plant Biomass (kg/hm)²）。

E. Nitrogen nutrition index: a critical nitrogen concentration dilution curve model and a nitrogen nutrition index (nitrogen nutrition index,NNI) The equation of (1) is: (1)N _c=a·DW^-b；（2）NNI=N _t/N _c(ii) a In the formula:N _cis the critical nitrogen concentration value, g/(100 g); the parameter a is that the biomass of the overground part of the crop is 10³ kg/hm²Critical nitrogen concentration of the plant; DW is the maximum value of biomass on the overground part of the crop, 10³ kg/hm²(ii) a b is a statistical parameter for determining the slope of the critical nitrogen concentration dilution curve;NNIis the nitrogen nutrition index;N _tthe measured value of the nitrogen concentration of the aboveground biomass is g/(100 g).

F. Canopy image processing and color parameter analysis: the digital image of the canopy of the tomato is obtained by taking a picture with a camera, clear weather without clouds at 12:00-14:00 noon is selected, the tomato canopy is taken in a direct-light vertical downward and direct-light overlooking (forming an included angle of 30-60 degrees with the canopy) mode, the picture taking height is 30cm away from the canopy, a forced exposure mode is set, the resolution of the camera is 2084 multiplied by 1536 pixels, the tomato canopy is obtained simultaneously with biomass about 30, 50, 70, 90, 110, 130 and 150 days after transplanting, and each test cell takes 2-3 pictures each time and stores the pictures in a JPEG format. The pictures were imported into a computer and the color G, R, B values were obtained using Photo-shop, Matlab software and averaged (Ahamd and Reid, 1996; Ballard and Brown, 1982). By eliminating interference factors, only color pixels of normal tomato leaves are left, then the standardized value of G, R, B color parameters of each pixel is calculated again, visible light atmospheric impedance vegetation indexes [ (G-R)/(R + G + B), VARI ] are introduced, and finally the average value statistics of the color parameters of the whole image is carried out for screening the color parameter indexes of crop nitrogen nutrition diagnosis.

4. Results and analysis

A. Dynamic change of nitrogen content of tomato plants treated by different water nitrogen

The process of variation of nitrogen content in tomato plants under different water nitrogen treatment is shown in table 1. As can be seen from Table 1, the nitrogen content of the tomato plants gradually decreases with time, and reaches a maximum value of 1.76-3.55 g/100g after planting for about 30 days; when the tomato pulls seedlings (DAT =150 d) reaches a minimum value of 1.01-2.00 g/100 g. The irrigation quantity is W₁And W₂The nitrogen content of tomato plants shows a remarkable trend along with the increase of nitrogen application amount, N₃₀₀And N₄₅₀Treatment significantly higher than N₀And N₁₅₀Is processed and N is₃₀₀And N₄₅₀The nitrogen content of treated plants has no obvious difference; the irrigation quantity is W₃The nitrogen content of tomato plants shows a remarkable trend along with the increase of nitrogen application amount, N₁₅₀、N₃₀₀And N₄₅₀Treatment significantly higher than N₀Is processed and N is₁₅₀、N₃₀₀And N₄₅₀The nitrogen content of the treated plants has no obvious difference. The result shows that the nitrogen content of tomato plants can be obviously increased by properly increasing the nitrogen application amount, and the irrigation amount is W₁And W₂While the nitrogen application amount is increased to 300 kg/hm²The nitrogen content of the plants is not obviously increased when the plant is continuously increased; the irrigation quantity is W₃While the nitrogen application rate is increased to 150 kg/hm²The nitrogen content of the plants is not obviously increased when the plant nitrogen content is continuously increased.

TABLE 12017 Nitrogen concentration in 2019 plants treated with different Water Nitrogen

B. SPAD value dynamic change of tomato leaves treated by different water nitrogen

Table 2 shows the change trend of SPAD values of tomato leaves treated by different nitrogen elements with time in 2017-2019. As can be seen from the table 2, the SPAD value of the tomato leaves is in a trend of increasing and then decreasing along with the increase of time, the SPAD value of each treated tomato leaf reaches the maximum value which is 33.49-64.50 respectively when the DAT =110 d, then the tomato leaves begin to age, the SPAD value of the tomato leaves begins to decrease, and the SPAD value of the tomato leaves during seedling pulling (DAT =150 d) is reduced to 28.90-53.30. The result shows that the SPAD value of the tomato leaves can be obviously increased by properly increasing the nitrogen application amount, and the water irrigation amount is W₁And W₂While the nitrogen application amount is increased to 300 kg/hm²When the SPAD value of the blade is continuously increased, the SPAD value of the blade is not obviously increased; the irrigation quantity is W₃While the nitrogen application rate is increased to 150 kg/hm²There was no significant increase in the SPAD value of the blade with continued increase.

TABLE 22017 SPAD values of different water nitrogen-treated folium Psidii Guajavae in 2019

C. Dynamic change of tomato biomass by different water nitrogen treatments

Table 3 shows the time-dependent changes of tomato biomass in 2017 and 2019 in different water nitrogen treatments. As can be seen from table 3, the biomass of the tomato plants gradually increased with time, and at the time of planting (DAT =30 d), the biomass was minimal, 0.36 × 10³～0.89×10³ kg/hm²(ii) a The maximum value is 5.48 multiplied by 10 when the tomato pulls seedlings (about DAT =150 d)³～11.64×10³ kg/hm². When the irrigation quantity is W₁When the tomato grows in the early stage (DAT is less than or equal to 70 d), the biomass of tomato plants tends to increase along with the increase of nitrogen application amount, and N is₃₀₀And N₄₅₀Treatment significantly higher than N₀And N₁₅₀Is processed and N is₃₀₀And N₄₅₀The biomass of the treated tomato plants has no obvious difference; in the middle and later growth period of the tomatoes (DAT > 70 d), the biomass of tomato plants is in a trend of remarkably increasing along with the increase of nitrogen application amount; when the irrigation quantity is W₂When the nitrogen application amount is increased, the biomass of tomato plants shows a remarkable increasing trend along with the increase of the nitrogen application amount, N₃₀₀And N₄₅₀Treatment significantly higher than N₀And N₁₅₀Is processed and N is₃₀₀And N₄₅₀The biomass of the treated tomato plants has no obvious difference; when the irrigation quantity is W₃When the nitrogen application amount is increased, the biomass of tomato plants shows a remarkable increasing trend along with the increase of the nitrogen application amount, N₁₅₀、N₃₀₀And N₄₅₀Treatment significantly higher than N₀Is processed and N is₁₅₀、N₃₀₀And N₄₅₀The biomass of the treated plants has no significant difference. The result shows that the proper increase of nitrogen application amount can obviously increase the biomass of tomato plants, and the irrigation amount is W₁When the tomato grows in the early stage (DAT is less than or equal to 70 d), the nitrogen application amount is increased to 300 kg/hm²When the plant biomass is continuously increased, the plant biomass is not obviously increasedThe increase of nitrogen application amount in the middle and later growth period (DAT > 70 d) of the eggplant can obviously increase the biomass of tomato plants; the irrigation quantity is W₂While the nitrogen application amount is increased to 300 kg/hm²When the biomass of the plant is continuously increased, the biomass of the plant is not obviously increased; the irrigation quantity is W₃While the nitrogen application rate is increased to 150 kg/hm²When the biomass of the plant is increased continuously, the biomass of the plant is not increased obviously.

TABLE 32017 + 2019 tomato biomass treated with different water nitrogen

D. Greenhouse tomatoNNINitrogen nutrition diagnostics

Determination of critical nitrogen concentration dilution model constants: the aboveground biomass of the tomato treated by different water nitrogen in 2017 and 2019 and the corresponding nitrogen concentration are calculated to obtain the critical nitrogen concentration (Justes et al, 1994) of each sampling day, and a tomato critical nitrogen dilution curve is established according to the aboveground biomass and the corresponding critical nitrogen concentration, as shown in figure 1. The coefficient of determination (R) of the critical nitrogen dilution curve of greenhouse tomatoes can be seen in FIG. 1²) The degree of fitting is 0.90-0.98, and the degree of fitting reaches an extremely significant level, which shows that the model can better reflect the relationship between the critical nitrogen concentration of the greenhouse tomatoes and the overground biomass.

Nitrogen nutrition index of tomato treated with different water nitrogen (NNI) Dynamic change: FIG. 2 is the dynamic change of nitrogen nutrient index of tomato under different water nitrogen treatment calculated according to the formula (3). From FIG. 2, it can be seen that the different water nitrogen treatmentsNNIThe nitrogen dosage is increased along with the increase of the nitrogen dosage, and the value range is 0.60 to 1.34. The tomatoes begin to bloom 70 days after field planting, the vegetative growth and reproductive growth are vigorous, the demand of plants on nitrogen is large, and different treatment roomsNNIThe difference is increased, and at the moment, the nitrogen fertilizer begins to be applied to meet the nitrogen requirement of plants. W₁At the time of treatment, N₀Treatment and N₁₅₀Treating the whole growth periodNNIAll are less than 1, and nitrogen cannot meet the requirements of plants; n is a radical of₃₀₀Treatment after 50 days of permanent plantingNNIThe nitrogen application amount is more suitable when the fluctuation is less than or near 1; n is a radical of₄₅₀ Treatment ofAfter 50 days of permanent plantingNNIAll are more than 1, and the nitrogen exceeds the requirement of plants. W₂At the time of treatment, N₀Treatment and N₁₅₀Treating the whole growth periodNNIAll are less than 1, and nitrogen cannot meet the requirements of plants; n is a radical of₃₀₀Treatment and N₄₅₀Treatment after 50 days of permanent plantingNNIAll are more than 1, and the nitrogen exceeds the requirement of plants. W₃At the time of treatment, N₀Treating the whole growth periodNNIAll are less than 1, and nitrogen cannot meet the requirements of plants; n is a radical of₁₅₀Treatment after 50 days of permanent plantingNNIThe nitrogen application amount is more suitable when the fluctuation is less than or near 1; n is a radical of₃₀₀Treatment and N₄₅₀Treatment after 50 days of permanent plantingNNIAll are more than 1, and the nitrogen exceeds the requirement of plants. The results show that W₁At the time of treatment N₀、N₁₅₀And N₄₅₀Since insufficient or too much nitrogen inhibits plant growth, the optimum nitrogen treatment should be N₃₀₀；W₂At the time of treatment N₀And N₄₅₀The treatment is performed because the plant growth is inhibited by insufficient nitrogen or too much nitrogen, and the optimum nitrogen treatment is N₁₅₀And N₃₀₀To (c) to (d); w₃At the time of treatment N₀、N₃₀₀And N₄₅₀Since insufficient or too much nitrogen inhibits plant growth, the optimum nitrogen treatment should be N₁₅₀。

E. Correlation analysis of tomato canopy digital image index and plant nitrogen nutrition parameter

Digital image index change analysis of tomato canopy under different water nitrogen treatment: the change process of digital image indexes (G, R and B) of tomato canopy under different water nitrogen treatment is shown in tables 4-6. As can be seen from tables 4-6, the tomato canopy digital image indicators (G, R and B) both have a decreasing trend with time, and the canopy digital image indicators have the characteristic that G > R > B. As can be seen from Table 4, the G characteristic value for different nitrogen treatments is represented by N when DAT < 90 d₀＞N₁₅₀＞N₃₀₀＞N₄₅₀The rule shows that the G characteristic value shows a trend of remarkably reducing along with the increase of the nitrogen application amount before the permanent planting for 90 days; DAT > 90 d, N₀Process G eigenvalues significantly higher than N₁₅₀、N₃₀₀And N₄₅₀Process G eigenvalues, and N₁₅₀、N₃₀₀And N₄₅₀The G characteristic value has no obvious difference after treatment, which indicates that the G characteristic value can be obviously reduced by properly increasing the nitrogen application amount after the permanent planting for 90 days, and when the nitrogen application amount is increased to 150 kg/hm²There was no significant change in the G characteristic value as the increase continued. As can be seen from Table 5, the R characteristic values for different nitrogen treatments are shown as N when DAT < 70d₀＞N₁₅₀＞N₃₀₀＞N₄₅₀The rule shows that the R characteristic value shows a trend of remarkably reducing along with the increase of nitrogen application amount before the permanent planting for 70 d; DAT > 70d, N₀Treatment and N₁₅₀Treatment of characteristic values of R significantly higher than N₃₀₀Treatment and N₄₅₀Processing R characteristic values, and N₃₀₀And N₄₅₀The R characteristic value is not obviously different after treatment, which shows that the R characteristic value can be obviously reduced by properly increasing the nitrogen application amount after the permanent planting for 70d, and when the nitrogen application amount is increased to 300 kg/hm²The R characteristic value did not change significantly as the increase continued. As can be seen from Table 6, the B characteristic values for different nitrogen treatments are represented by N when DAT < 110 d₀＞N₁₅₀＞N₃₀₀＞N₄₅₀The rule shows that before the planting for 110 d, the characteristic value of B shows a trend of remarkably reducing along with the increase of the nitrogen application amount; when DAT is greater than or equal to 110 d, N₀Treatment and N₁₅₀Treatment of B eigenvalues significantly higher than N₃₀₀Treatment and N₄₅₀Processing B eigenvalues, and N₃₀₀And N₄₅₀The characteristic value of B has no obvious difference after treatment, which indicates that the characteristic value of B can be obviously reduced by properly increasing the nitrogen application amount after the planting for 110 d, and when the nitrogen application amount is increased to 300 kg/hm²The B characteristic value has no significant change when the B characteristic value is continuously increased. The results show that the increase of nitrogen application amount when DAT is less than 70d can obviously reduce the digital image indexes (G, R and B) of the tomato canopy, and the increase of nitrogen application amount when DAT is more than 110 d can obviously reduce the digital image indexes (G, R and B) of the tomato canopy, and the increase of nitrogen application amount is more than 300 kg/hm²There was no significant decrease in tomato canopy digital image indicators (G, R and B) with continued increase.

Table 42017 and 2019 digital image G characteristic values of different water nitrogen treatment tomato canopy digital images

Table 52017-2019 digital image R characteristic values of different water nitrogen treatment tomato canopy digital images in different years

Characteristic values of digital images B of tomato canopy subjected to different water nitrogen treatments in table 62017-2019

Respectively comparing the color characteristic values (G, R, B, G/(G + R + B), R/(G + R + B), B/(G + R + B) and VARI) of the tomato canopy treated by different water nitrogen in 2017 and 2019 with the plant nitrogen concentration, SPAD value, biomass and nitrogen nutrition index (SPAD value, biomass and nitrogen nutrition indexNNI) Pearson correlation analysis was performed (table 7). The result shows that G, R and B characteristic values are in extremely obvious positive correlation with the nitrogen content of the plant, and the correlation coefficient is 0.693-0.965; G. the characteristic values of R and B are extremely obviously and negatively correlated with the biomass of the plants, and the correlation coefficient is-0.858 to-0.979; the characteristic values of G and R are extremely obviously and negatively correlated with the SPAD value, and the correlation coefficient is-0.732 to-0.842; G. r and B characteristic values and the index of nutrient nutrition: (NNI) There is no significant relationship. G/(G + R + B) characteristic value, SPAD value and nitrogen nutrition index (A)NNI) The biomass of the plant is obviously inversely correlated with the biomass of the plant, and the correlation coefficient is-0.515 to-0.935; the G/(G + R + B) characteristic value is in positive correlation with the plant nitrogen concentration, and the correlation coefficient is 0.875-0.904; R/(G + R + B) eigenvalue and nitrogen nutrient index (G + R + B)NNI) The correlation coefficient is-0.542 to-0.627; the characteristic value of B/(G + R + B) is in positive correlation with the SPAD value and the plant biomass, and the correlation coefficient is 0.750-0.912; the VARI is obviously inversely related to plant biomass, and the correlation coefficient is-0.576 to-0.825.

The correlations of the characteristic values of G/(G + R + B) of different water nitrogen treatments with the plant nitrogen concentration, the leaf SPAD value, the biomass and the nitrogen nutrition index reach obvious levels, and the correlations are obviously superior to other canopy digital image indexes (Table 7). Therefore, the G/(G + R + B) characteristic value is selected as a parameter index for carrying out nitrogen nutrition on the tomatoes by the digital camera, and the representativeness and the stability are high.

TABLE 7 correlation coefficient between digital image index and plant nitrogen index of tomato treated differently

In order to further explore the feasibility of applying the characteristic value of the canopy image digitization index G/(G + R + B) to greenhouse tomato nitrogen nutrition assessment prediction, the invention analyzes the correlation between the characteristic value of G/(G + R + B) and tomato plant nitrogen nutrition parameters in 2017 + 2018 by respectively adopting a linear equation function, a logarithmic equation function, a quadratic equation function, a power equation function and an exponential equation function (Table 7). The result shows that the G/(G + R + B) characteristic value is optimally related to the nitrogen concentration, the SPAD value and the plant biomass by a linear equation, and the G/(G + R + B) characteristic value is related to the nitrogen nutrition index (A)NNI) The correlation performance is optimal by quadratic equation, so that the invention selects a linear equation and a quadratic equation as the optimal estimation prediction equation model (figure 3) respectively, wherein the regression equations of the G/(G + R + B) characteristic value, the nitrogen concentration, the SPAD value, the plant biomass and the nitrogen nutrition index are respectively y (G/100G) =14.51x-3.95, y = -111.90x +86.51 and y (t/hm)²) = 74.04x +33.97 and y =81.59x²-65.51x+13.99。

TABLE 8 Linear and non-linear regression analysis of tomato leaf color eigenvalues G/(G + R + B) and nitrogen nutrient index

In order to verify the adaptability and the accuracy of the greenhouse tomato nitrogen nutrition equation model, the results of 2018-year experiment treatment are used for verification, the plant nitrogen concentration, the SPAD value, the biomass and the predicted value of the nitrogen nutrition index calculated according to the model are compared with the measured value, and the absolute error (MAE), the Root Mean Square Error (RMSE) and the decision coefficient (R) are used for determining²) The accuracy of the model was investigated comprehensively by 3 indices (fig. 4). The results show that (G/(G + R + B)) and plant nitrogen concentration, SPAD value, biomass and nitrogen nutrition are utilizedThe estimation model established between indexes has the decision coefficient R between the measured value and the predicted value reaching the most significant level²0.68, 0.62, 0.82 and 0.65 respectively, the absolute error MAE is 0.35, 7.73, 1.18 and 0.10 respectively, the root mean square error RMSE is 0.47, 9.36, 1.87 and 0.14 respectively, the prediction result has better precision and can be used for model prediction.

Nitrogen nutrition index: (NNI) As the ratio of the actual plant nitrogen content to the critical nitrogen content, the nutrient status of the nitrogen in the crop can be visually reflected, and the nutrient status of the nitrogen in the crop can be directly reflectedNNI=1, nitrogen nutrition status is most suitable; when in useNNI>1, indicated as nitrogen over-nutrition; when in useNNI<1, manifested as nitrogen nutrient deficiency. The test is carried out by establishing the characteristic value of G/G + R + B of the tomato canopy and the nitrogen nutrition indexNNIRelation between (A) and (B)NNI=81.59(G/(G+R+B))²65.51(G/(G + R + B)) + 13.99), determining the equation for the linear monitoring of the nitrogen nutrition with the independent variable of the characteristic values of the tomato canopy G/G + R + B, when 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNImore than 1, nitrogen nutrition is excessive; therefore, the nitrogen nutrition index is obtained by quickly inverting the digital image information of the canopy of the greenhouse tomato, and the nitrogen nutrition condition of the greenhouse tomato is monitored.

Greenhouse tomatoes are crops with large nitrogen demand, and the yield of the tomatoes is influenced when the nitrogen application amount is too large or too small, so that the nitrogen nutrition condition can be diagnosed in time, and the accurate nitrogen application is very important. Research shows that with the increase of nitrogen application amount, the nitrogen concentration, the SPAD value of leaves, the biomass and the nitrogen nutrition index of tomato plants tend to increase, and have a tendency that the nitrogen concentration, the SPAD value, the biomass and the nitrogen nutrition index of the tomato plants do not change obviously after the nitrogen application amount is increased to a certain degree (Zhao Huatian and the like, 2011), the promotion effect of properly increasing the nitrogen concentration of soil on the biomass of tomatoes is obvious, and a large proportion of nitrogen is easy to accumulate, which is similar to the research of Wangbui and the like (2003) and Turkish and the like (2018).

The plant leaf is one of the most sensitive parts to the nitrogen nutrition condition, the quantity of nitrogen can cause the color, texture and other characteristics of the leaf to change, and the color of the plant leaf can be changed by computer vision rather than by human eyesSensitivity and reliability are added, especially in the early stage of the deficiency of the nitrogen, so that the change of the nitrogen deficiency on the color of plant leaves needs to be evaluated by using a computer image processing technology, and the nitrogen stress condition of the crops needs to be analyzed from the quantitative point of view (Xugui power et al, 2002). It can be seen from Table 1 that the tomato canopy digital image indices (G, R and B) both have a decreasing trend with time, with the characteristic G > R > B. This is consistent with the results of the studies of Ding Yong Jun et al (2012), where the green eigenvalue of the digital image of the tomato canopy is the largest, the red order and the blue color is the smallest. Greenhouse tomato canopy digital image indexes (G, R and B) have similar tendency with tomato nitrogen nutrition indexes, the tomato canopy digital image indexes (G, R and B) can be obviously reduced by increasing the nitrogen application amount in the early and middle stages (DAT < 70 d) of the tomato, mainly because the tomato grows vigorously in the early stage and has self-regulation capability when stressed by nitrogen, and slight nitrogen stress can cause the color change of the plant canopy; the nitrogen application amount is increased to 300 kg/hm at the later stage of the tomato (DAT > 110 d)²The digital image indexes (G, R and B) of the tomato canopy are not changed significantly when the increase is continued, which is mainly because the plant bodies lose the self-regulation ability when being stressed seriously at the later stage of the crop (DAT > 110 d), and the color change amplitude of the leaves is small.

There is a significant or very significant correlation between the crop canopy digital image color parameters and plant nitrogen concentration (zhangyuan et al, 2010; juan et al, 2008), chlorophyll meter reading (xiaozhen wave et al, 2008), NDVI (Rouse et al, 1974; jialian, 2003), biomass (xiaozhen wave et al, 2008) and nitrogen nutrition index (wei quan et al, 2015). The research is carried out by analyzing the color characteristic values (G, R, B, G/(G + R + B), R/(G + R + B), B/(G + R + B) and VARI) of the canopy of the greenhouse tomato and the nitrogen nutrition parameters (plant nitrogen concentration, SPAD value, biomass and nitrogen nutrition index ((B)NNI) A relation between the tomato canopy color characteristic values G, R, B and G/(G + R + B) and the plant nitrogen concentration are in a significant or extremely significant positive correlation; G. r, B, G/(G + R + B) and VARI characteristic values are in extremely obvious negative correlation with plant biomass, and B/(G + R + B) characteristic values are in obvious positive correlation with the plant biomass, which is consistent with the research result of Jia et al (2004); wang et al (2)008) Researches also prove that the green light-red light ratio (G/R) of the bamboo leaf chip is well correlated with the reading of a chlorophyll meter, and the researches also show that G, R, G/(G + R + B) characteristic values are extremely obviously and negatively correlated with SPAD values, and B/(G + R + B) characteristic values are obviously and positively correlated with the SPAD values and are consistent with the research results of Scharf and Lory (Scharf and Lory, 2002); the study of Weiquan congruent (2015) shows that the red light standardized value R/(G + R + B) and the nitrogen nutrition index of the winter rape canopy image digital index reach significant or extremely significant levels, and the study also shows that G/(G + R + B), R/(G + R + B) characteristic values and the nitrogen nutrition index (A (B) (G + R + B))NNI) Exhibit a significant negative correlation.

Different crops reflect different color parameters of nitrogen nutrition indexes, the research of Zhangzhou and the like (2010, 2011) is based on the correlation between color parameters of images of canopy layers of winter wheat and summer corn obtained by a digital camera and the traditional nitrogen nutrition indexes, and green light standardized values and blue light standardized values are screened out and are respectively used as sensitive color parameters for diagnosing the nitrogen nutrition state of the crops; the nitrogen nutrition status of crops is diagnosed by using digital image technology, namely the Wangliujun and the like (2010) and the Wangzuan and the like (2008), and G/(R + G + B) is considered to be a better nitrogen nutrition diagnosis index of the crops, and the research shows that compared with other digital indexes of images of tomato canopy, G/(R + G + B) is compared with the nitrogen concentration, the SPAD value, the biomass and the nitrogen nutrition index of plants (the value is the ratio of the concentration to the concentration of plant nitrogen to the value of SPAD), and the value is the ratio of biomass to the nitrogen nutrition index of (the value is the ratio of the concentration of plant nitrogen to the value of plant nitrogen to the index of plant nitrogen to the value of plant nitrogen to the plantNNI) All reach a significant or extremely significant level, and can be recommended as the optimal color parameter for predicting the nitrogen nutrition of the tomatoes by applying a digital image technology.

Nitrogen nutrition index: (NNI) As the ratio of the actual plant nitrogen content to the critical nitrogen content, the nutrient status of the nitrogen in the crop body can be intuitively reflected, but a large amount of field sampling and laboratory analysis are needed, the operation is complex, time and labor are wasted, and the requirement of real-time topdressing diagnosis in the crop growth period is difficult to adapt. Many studies have shown the feasibility of digital image analysis techniques to estimate crop nitrogen nutrition indices by obtaining red, green, blue and other image feature values of plant canopy (shashasha et al, 2019; zhangzhuanzhou et al, 2010, 2011). The green light standardization value G/(G + R + B) and the nitrogen nutrition index are established and verified by 2017-year experiment dataNNIThe relationship between the two or more of them,and determining a nitrogen nutrition monitoring equation with G/(G + R + B) as an independent variable, so that a nitrogen nutrition index can be obtained through rapid inversion of the digital image information of the greenhouse tomato canopy, the nitrogen nutrition condition of the tomato can be monitored, and reference are provided for application and development of a digital camera in crop nitrogen nutrition diagnosis in a greenhouse environment.

According to the invention, according to the test data of different water nitrogen treatments of greenhouse tomatoes in 2 years, a method for diagnosing the nitrogen nutrition of greenhouse tomatoes by using a digital camera is researched, the relation between the digital index of an image analysis canopy image and the nitrogen nutrition condition is preliminarily established, the nitrogen nutrition diagnosis index and the diagnosis model of the color parameter of the canopy of tomatoes are screened and provided, and the result shows that:

1) the nitrogen concentration, the SPAD value of leaves, the biomass, the nitrogen nutrition index and the canopy digital image index (G, R and B) of tomato plants under different water nitrogen treatment trends to be consistent, mainly represented by increasing with the increase of nitrogen application amount and having no significant change after increasing to a certain degree.

2) Compared with other canopy image digital indexes, the tomato canopy green light standardized value G/(R + G + B) and plant nitrogen nutrition parameters reach significant or extremely significant levels. By establishing a nitrogen nutrition assessment prediction equation model based on a tomato green light standardized value (G/(R + G + B)) and obtaining a better prediction result, the nitrogen concentration, the leaf SPAD value, the biomass and the nitrogen nutrition index of a greenhouse tomato plant can be accurately predicted by taking G/(R + G + B) as an optimal color parameter index (the formula (A) (B))NNI）。

3) Using green light normalized value G/(G + R + B) and nitrogen nutrition index (G:)NNI) Correlation between (A) and (B) ((B))NNI=81.59(G/(G+R+B))²65.51(G/(G + R + B)) + 13.99), when 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNImore than 1, nitrogen nutrition is excessive; therefore, the nitrogen nutrition index is obtained by quickly inverting the digital index G/(R + G + B) of the greenhouse tomato canopy image, and the plant nitrogen nutrition state is accurately diagnosed in time.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.

Reference documents:

Ahamd I S, Reid J F. Evaluation of color representations for maize image[J]. Journal of Agricultural Engineering Research, 1996, 63(3): 185-196.

Ballard D H, Brown C M. Computer vision[M]. London: Prentice-Hall, Inc, 1982: 313-437.

Blackmer T M, Schepers J S. Use of a chlorophyll meter to monitor nitrogen status and schedule fertigation for corn[J]. Journal of Production Agriculture，1995, 8(1): 56-60.

Borhan M S, Panigrahi S, Lorenzen J H, et al. Multispectral and color imaging techniques for nitrate and chlorophyll determination of potato leaves in a controlled environment[J]. Transactions-American Society of Agricultural Engineers, 2004, 47(2): 599-608.

Hendrawan Y, Murase H, Determining an ANN pre-treatment algorithm to predict water content of moss using RGB intensities[J]. Engineering in Agriculture, Environment and Food, 2011, 4(4): 95-105.

Jia L L, Chen X P, Zhang F S, BürkertA, Römheld V. Use of digital camera to assess nitrogen status of winter wheat in the Northern China Plain[J]. Journal of Plant Nutrition, 2004, 27(3): 441-450.

Justes E, Mary B, Meynard J M. Determination of a critical nitrogen dilution curve for winter wheat crops. Annals of Botany, 1994, 74(4): 397－407.

Lee K J, Lee B W. Estimation of rice growth and nitrogen nutrition status using color digital camera image analysis[J]. European Journal of Agronomy, 2013, 48(3): 57-65.

Lemaire G, Gastal F, Salette J. N uptake and distributionin plant canopies: Diagnosis of the nitrogen status in crops[J]. Berlin: Spriger. 1997.

Lemaire G, Jeuffroy M H, Gastal F. Diagnosis tool for plant and crop N status in vegetative stage: Theory and practices for crop N management. European Journal of Agronomy, 2008, 28(4): 614-624.

Lin F, Qiu L, Dang J S, Shi Y, Chen L S, Wang K. Investigation of SPAD meter-based indices for estimating rice nitrogen status[J]. Computers and Electronics in Agriculture, 2010, 71: 60-65.

Madeire A C, Mendonca A, Ferreire M E, et al. Relationship between spectroradiometric and chlorophyll measurements in green beans[J]. Communications in Soil Science & Plant Analysis, 2000, 7(2/3): 695－704.

Rouse J W, Has R H, Schell J A, et al. Monitorning vegetation systems in the great plains with ERTS[C]. Third ERTS Symposium, NASA SP-351, Vol. 1. Washington DC: NASA, 1974: 309-317.

Scharf P C, Lory J A, Calibrating corn color from aerial photographs to predict sidedress nitrogen need[J]. Agronomy Journal, 2002, 94(3): 397-404.

Wang F, Haruhiko Y, Yasuomi I. Measuring leaf necrosis and chlorosis of bamboo induced by typhoon 0613 with RGB image analysis[J]. Journal of Forestry Research, 2008, 19(3): 225-230.

ding Yong Jun, Li Xun Zan, Sun Red, etc. tomato nutrient diagnosis model based on multispectral image technology [ J ] agricultural engineering report, 2012, 28(8): 175-.

Ding Yongjun, Li Minzan, Sun Hong, et al. Diagnosis model of tomato nutrient content based on multispectral images[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012, 28(8): 175－180. (in Chinese with English abstract)

Application of computer image processing in identification of color characteristics of lacked leaves [ J ] computer measurement and control, 2004, 12 (9): 859-.

Hu C H, Li P P. Application of Computer Image Processing in Recognition of Color Feature of Nutrient Deficiency Leaves[J]. Computer measurement and control, 2004, 12(9): 859-862. (in Chinese with English abstract)

Jialiang, Vanming, Zhang Fu Lock, etc. the digital camera is used for the diagnosis of nitrogen nutrition in rice, spectroscopy and spectrum analysis, 2009, 29(8): 2176-.

Jia L L, Fan M S, Zhang F S, et al. Nitrogen status diagnosis of rice by using a digital camera. Spectroscopy and Spectral Analysis, 2009, 29(8): 2176-2179. (in Chinese with English abstract)

Jialiang, application of digital image technology and soil plant testing for winter wheat nitrogen nutrition diagnosis [ D ]. Beijing, university of Chinese agriculture, 2003.

Jia L L. To detect nitrogen status of winter wheat by using color digital camera[D]. Beijing: China Agricultural University, 2003. (in Chinese with English abstract)

The present situation, the problems and the trend of the application of the nitrogenous fertilizer in farmland in China [ J ] plant nutrition and fertilizer bulletin 2014, 20(4): 783-.

Ju X T, Gu B J. Status-quo, problem and trend of nitrogen fertilization in China[J]. Journal of Plant Nutrition and Fertilizer, 2014, 20(4): 783-795. (in Chinese with English abstract)

Plum blossom, research on a rapid diagnosis method of nitrogen nutrition of processed tomatoes [ D ]. Shihe-Shihe university, 2007.

Li Y. Study on quick diagnosing processing tomato nitrogen nutrition states techniques[D]. Shihezi: Shihezi University, 2007. (in Chinese with English abstract)

Response study of drip irrigation of tomatoes under greenhouse film to water-nitrogen coupling [ D ]. yangling: northwest university of agriculture and forestry, 2013.

Shi X H. Response of Greenhouse Tomato with Drip Irrigation under Plastic Mulch to Water-nitrogen Coupling[D]. Yangling: Northwest A&F University, 2013. (in Chinese with English abstract)

Wang Juan, Lei Yong Wen, Zhang Yong Shuai, etc. the research on the nitrogen nutrition diagnosis of cotton by applying the digital image analysis technology [ J ]. the ecological agriculture bulletin in China, 2008, 16(1): 145-.

Wang J, Lei Y W, Zhang Y S, et al. Diagnosis of cotton N status using digital image analysis technique[J]. Chinese Journal of Eco-Agriculture, 2008, 16(1): 145-149. (in Chinese with English abstract)

Wanglijun, Chengyu, application research of digital image technology in strawberry nitrogen nutrition diagnosis [ J ]. proceedings of south China agricultural university, 2010, 31(2): 19-21.

Wang L J, Xing Y. Study on strawberry of N nutrition diagnosis using digital image processing technique[J]. Journal of South China Agriculture University, 2010, 31(2): 19−21. (in Chinese with English abstract)

Wangyan, Miguo Hua, Chenfanjun, etc. the genotype difference of the absorption of nitrogen in corn and the correlation with the root system morphology [ J ] ecology bulletin, 2003, 23(2): 297 + 302.

Wang Y, Mi G H, Chen F J, et al. Genotypic differences in nitrogen uptake by maize inbred itsrelation to root morphology[J]. Acta Ecologica Sinica, 2003, 23(2): 297-302. (in Chinese with English abstract)

Weiquan, Lilan, Tao, etc. Dongbai nitrogen nutrition diagnosis based on digital image technology [ J ] China agricultural science 2015, 48(19): 3877-3886.

Wei Q Q, Li L T, Ren T, et al. Diagnosing nitrogen nutrition status of winter rapeseed via digital image processing technique[J]. Scientia Agricultura Sinica, 2015, 48(19): 3877-3886. (in Chinese with English abstract)

Wu Xue Mei, identification research on deficiency of nutrient elements of tomatoes based on computer vision [ D ]. Zhenjiang, Jiangsu university, 2004.

Wu X M. Research on Recognition of Nutrient Element Deficiency in Tomato Based on Computer Vision[D]. Zhengjiang: Jiangsu University, 2004. (in Chinese with English abstract)

Xiasa, Zhang Smart, Li Jiazhen, etc. Nitrogen diagnosis and recommended fertilization research based on a digital image of winter wheat canopy obtained by a mobile phone camera [ J ] China ecological agriculture news, 2019, 26(4): 538) 546.

Xia S S, Zhang C, Li J Z, et al. Study on nitrogen nutrition diagnosis and fertilization recommendation of winter wheat using canopy digital images from cellphone camera[J]. Chinese Journal of Eco-Agriculture, 2019, 26(4): 538-546. (in Chinese with English abstract)

Xiaoyan wave, Jialiang, Chenxinping, etc. the digital image analysis technology is applied to nitrogen nutrition diagnosis in the elongation stage of winter wheat [ J ] agricultural information science, 2008, 24(8): 448-453.

Xiao Y B, Jia L L, Chen X P, et al. N status diagnosis of winter wheat by using digital image analysis technology[J]. Chinese Agricultural Science Bulletin, 2008, 24(8): 448−453. (in Chinese with English abstract)

Study on color feature extraction of Xugui power, Erhanping, Lemna, color image of ischemic leaf [ J ] agricultural engineering report 2002, 18(4): 150-.

Xu G L, Mao H P, Li P P. Extracting Color Features of Leaf Color Images[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2002, 18(4): 150-154. (in Chinese with English abstract)

Zhang Wen, Houdayu, Zhang Yu Ming, etc. the application of digital image diagnosis technology in the nitrogen nutrition diagnosis of winter wheat [ J ]. the report of ecological agriculture in China 2011, 19(5): 1168-.

Zhang L Z, Hou X Y, Zhang Y M, et al. Diagnosis of the state of N in wheat via digital image processing technique[J]. Chinese Journal of Eco-Agriculture, 2011, 19(5): 1168-1174. (in Chinese with English abstract)

Zhang Zhou, Wang Dynasty, Zhang Yu Ming, etc. the application of digital image technology in the diagnosis of nitrogen element in summer corn [ J ]. the report of ecological agriculture in China 2010, 18(6): 1340-1344.

Zhang L Z, Wang D W, Zhang Y M, et al. Diagnosis of N nutrient status of corn using digital image processing technique[J]. Chinese Journal of Eco-Agriculture, 2010, 18(6):1340-1344. (in Chinese with English abstract)

Zhao Huatian, Wang Rui Fang, xu Yun Feng, etc. screening and evaluating the low nitrogen tolerant genotype of wheat in seedling stage [ J ] Proc of ecological agriculture in China 2011, 19(5): 1199-1204.

Zhao H T, Wang R F, Xu Y F, et al. Screening and evaluating low nitrogen tolerant wheat genotypeat seedling stage[J]. Chinese Journal of Eco-Agriculture, 2011, 19(5): 1199-1204. (in Chinese with English abstract)

Zhao Manxing, Zhou Jian, Zhai, et al, chlorophyll diagnosis of nitrogen nutrition of winter wheat varieties in dry land plant nutrition and fertilizer bulletin, 2005, 11(4): 461-466.

Zhao M X，Zhou J B，Zhai B N，et al. Chlorophyll diagnoses of N nutrition for different winter wheat cultivars on dry land Plant Nutrition and Fertilizer Science, 2005, 11(4): 461-466. (in Chinese with English abstract)

Zhanqiang, Gongyu, Chenren Campsis, etc. the research on the diagnosis of nitrogen nutrition of tobacco based on digital image technology [ J ] academic newspaper of Jiangxi university of agriculture, 2018, 39(6): 1104-1111.

Zhong S R, Gong S Y, Chen R X, et al. Diagnosis of Nitrogen Nutrition Status of Tobacco based on Digital Image Processing Technique[J]. Acta Agriculturae Universitatis Jiangxiensis, 2018, 39(6): 1104-1111. (in Chinese with English abstract)。

Claims

1. A method for detecting the nitrogen nutrition status of greenhouse tomatoes based on a digital camera is characterized by comprising the following steps: the method comprises the steps of obtaining color characteristic values of tomato leaf images of tomatoes treated by different water nitrogen, carrying out correlation analysis on the color characteristic values and plant nitrogen nutrition indexes, screening out nitrogen nutrition diagnosis indexes and diagnosis models of tomato canopy color parameters by combining detection of traditional nitrogen nutrition indexes, establishing a fitting model for inverting the tomato nitrogen nutrition indexes, determining nutrition diagnosis indexes and providing fertilization recommendation.

2. The method for detecting the nitrogen nutrition status of greenhouse tomatoes based on a digital camera as claimed in claim 1, wherein the method comprises the following steps: specifically, green light standardization value G/(G + R + B) and nitrogen nutrition index are utilizedNNIThe correlation relationship between:NNI=81.59(G/(G+R+B))²65.51(G/(G + R + B)) +13.99, when 0.36 < G/(G + R + B) < 0.45,NNIless than 1, nitrogen nutrient deficiency; when G/(G + R + B) < 0.36 or G/(G + R + B) > 0.45,NNIif the nitrogen nutrition index is more than 1, the nitrogen nutrition index is obtained through inversion by obtaining the greenhouse tomato canopy image digital index G/(R + G + B) and the plant nitrogen nutrition state is accurately diagnosed in time.

3. The method for detecting the nitrogen nutrition status of greenhouse tomatoes based on a digital camera as claimed in claim 1, wherein the method comprises the following steps: the different water nitrogen treatments of the tomatoes are as follows: the water treatment is set to 3 levels, and the water filling amount is respectively full water filling treatment in the whole growth period, 25% of water deficiency in the whole growth period and 50% of water deficiency in the whole growth period; the nitrogen treatment was set at 4 levels with nitrogen application rates of 0, 150, 300 and 450 kg/hm, respectively²。

4. The method for detecting nitrogen nutrition status of greenhouse tomatoes based on a digital camera as claimed in any one of claims 1 to 3, wherein the method comprises the following steps: the method is characterized in that: the specific method comprises the following steps:

(1) observation item and data acquisition:

measurement of irrigation quantity: the irrigation starts from 15 d after the field planting, the upper limit of irrigation is field water capacity after the full irrigation treatment (θ _FC) 90% of water, the amount of water pouredI(mm) is:I=10(0.9θ _FC−θ _i)Z _r(ii) a In the formula:θ _ito be irrigatedWater content of soil before water, cm³/cm³；Zr is the planned wetting layer depth of 60 cm;

(2) Nitrogen nutrition indexNNIAnd (3) calculating: critical nitrogen concentration dilution curve model and nitrogen nutrition indexNNIThe equation of (1) is:N _c=a·DW^-b；NNI=N _t/N _c(ii) a In the formula:N _cis the critical nitrogen concentration value, g/(100 g); the parameter a is that the biomass of the overground part of the crop is 10³kg/hm²Critical nitrogen concentration of the plant; DW is the maximum value of biomass on the overground part of the crop, 10³ kg/hm²(ii) a b is a statistical parameter for determining the slope of the critical nitrogen concentration dilution curve;NNIis the nitrogen nutrition index;N _tg/(100 g) as an observed value of the nitrogen concentration of the aboveground biomass;