Method and device for measuring nitrogen content of rice leaves based on deep learning

Abstract

The invention relates to the technical field of agriculture, in particular to a method and device for measuring the nitrogen content of rice leaves based on deep learning, and the method comprises the following steps: S1, sample collection: collecting a plurality of rice as a sample, and soaking a sample root system in a nutrient solution; s2, sample transfer: adjusting transfer parameters according to the current environment, and transferring the sample to a laboratory; recording the activity index of the sample in the transfer process; s3, sample treatment: cleaning the sample transferred to the laboratory with clear water, and cutting off withered and damaged rice leaves; s4, sample detection: detecting the processed rice leaves by using a spectrograph to obtain spectral data of the rice leaves; s5, intelligent analysis: analyzing the spectral data of the rice leaves based on a deep learning model to obtain the current nitrogen content of the rice leaves; and compensating the current nitrogen content of the rice leaves based on the compensation model according to the transport parameters and the activity indexes. The sample transfer environment can be optimized, and the detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of agricultural technology, and in particular to a method and device for measuring the nitrogen content of rice leaves based on deep learning. Background Art

[0002] Nitrogen is the nutrient element most closely related to crop photosynthesis, yield and quality, and is also the mineral element with the largest demand and application amount for crops. As one of the world's major food crops, the nitrogen nutrition assessment of rice is of great significance for monitoring rice growth and achieving precise field management. Therefore, it is an urgent need for modern agricultural production to quickly and accurately obtain the nitrogen status of crops and achieve precise and efficient fertilization of farmland.

[0003] In the prior art, the diagnosis method of nitrogen content in rice leaves mainly involves spectral detection of plant tissues in the laboratory, and intelligent analysis of the nitrogen content in rice leaves using a deep learning model based on the detected spectral data, thereby quickly obtaining the nitrogen content in rice leaves. However, in the prior art, since the samples need to be collected from the field and then transported to the laboratory, the activity and nitrogen content of the samples change during the transportation process, and the prior art does not adequately protect the samples during the transportation process, which further reduces the accuracy of the sample detection results.

[0004] In summary, how to solve the problem that the existing technology does not provide sufficient protection for samples during transportation, thereby reducing the accuracy of sample detection results has become a difficult problem that needs to be solved urgently in the field. Therefore, it is necessary to propose a method and device for measuring the nitrogen content of rice leaves based on deep learning, which can enhance the protection of samples during transportation and improve the accuracy of detection results. Summary of the invention

[0005] To solve the above problems, the present invention provides a method and device for measuring the nitrogen content of rice leaves based on deep learning. Through the design of the outer shell and the top cover, the sample can be sealed and transported, and the transportation environment of the sample can be adjusted at the same time, so that the activity of the sample is protected and the change of its carbon content is reduced; and through the design of the deep learning model and the compensation model, the actual carbon content of the sample after compensation can be effectively calculated to improve the detection accuracy.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: A method for measuring the nitrogen content of rice leaves based on deep learning comprises the following steps:

[0007] S1, sample collection: collect a number of rice plants as samples and immerse the sample roots in nutrient solution.

[0008] S2, sample transportation: adjust the transportation parameters of the transportation environment according to the current environment, and transport the samples to the laboratory; record the activity indicators of the samples during the transportation process.

[0009] S3, sample processing: After being transported to the laboratory, the samples were washed with clean water and the dry and damaged rice leaves were removed.

[0010] S4, sample detection: using a spectrometer to perform spectral detection on the treated rice leaves to obtain spectral data of the rice leaves.

[0011] S5, intelligent analysis: Analyze the spectral data of rice leaves based on the deep learning model to obtain the current nitrogen content of the rice leaves; compensate the current nitrogen content of the rice leaves according to the transport parameters and activity indicators based on the compensation model to obtain the actual nitrogen content of the rice leaves.

[0012] Furthermore, in S2, the transport parameters include temperature, humidity and light of the transport environment; the activity indicators include photosynthetic rate, transpiration rate and stomatal conductance during sample transport; in S5, the deep learning module is trained based on the spectral data of several rice leaves; and the compensation model is trained based on the linear change data of several transport parameters and activity indicators.

[0013] The photosynthetic rate calculation formula is as follows:

[0014] V O2 =(Vx-Vt) / M (1).

[0015] Among them, V O2 represents the oxygen generation rate, Vx represents the oxygen release amount, Vt represents the initial oxygen concentration), and M represents the leaf area.

[0016] The transpiration rate calculation formula is as follows:

[0017] Tr = Ox / (Mt×T) (2).

[0018] Among them, Tr represents transpiration rate, Ox represents transpiration water loss, Mt represents unit leaf area, and T represents time.

[0019] The formula for calculating stomatal conductance is as follows:

[0020] gs = E / (1.6×Vpd) (2).

[0021] Where gs is the stomatal conductance, E is the transpiration rate, Vpd is the water vapor pressure deficit at the leaf-air interface, and 1.6 is an empirical constant.

[0022] The technical principles of the above scheme are as follows:

[0023] A number of rice plants are collected as samples, and the roots of all the samples are immersed in a nutrient solution. At the same time, the temperature, humidity and light of the transportation environment are adjusted according to the appropriate temperature, humidity and light of the samples.

[0024] The photosynthetic rate, transpiration rate and stomatal conductance of the samples during transport were recorded in real time during the transport process.

[0025] After the samples were transported to the laboratory, they were washed with clean water and the dried and damaged rice leaves were removed. After the treatment, the spectrometer was used to detect the spectrum of the rice leaves to obtain the spectrum data of the rice leaves.

[0026] The deep learning model is used to analyze the spectral data of rice leaves to obtain the current nitrogen content of the rice leaves. Then, based on the compensation model, the current nitrogen content of the rice leaves is compensated according to the temperature, humidity and light of the transportation environment and the photosynthetic rate, transpiration rate and stomatal conductance of the rice during transportation to obtain the actual nitrogen content of the rice leaves.

[0027] The above scheme has the following beneficial effects:

[0028] 1. The present invention analyzes the spectral data of rice leaves through a deep learning model, and can quickly and accurately estimate the nitrogen content of rice leaves. Compared with the prior art, the present invention further considers the temperature, humidity and light during the transportation process as well as the photosynthetic rate, transpiration rate and stomatal conductance of rice through a compensation model, and further corrects the estimated nitrogen content, thereby improving the accuracy of detection.

[0029] 2. During the transportation process, the present invention adjusts the temperature, humidity and light of the transportation environment to a range suitable for rice, thereby improving the activity of rice during the transportation process and reducing the loss of its carbon content, thereby optimizing the transportation environment of rice, improving the quality of samples transported to the laboratory, and further improving the accuracy of its carbon content detection results.

[0030] 3. The present invention conducts a large amount of training on the deep learning model and the compensation model, so that the present invention can analyze the sample intelligently and automatically, thereby quickly and accurately obtaining the carbon content analysis results of the sample. Compared with manual analysis, the present invention has higher analysis efficiency.

[0031] Furthermore, a device for measuring the nitrogen content of rice leaves based on deep learning includes a controller, an outer shell and a top cover, wherein a plurality of clamping blocks are slidably fitted on the outer shell; a plurality of rotating rods are rotatably fitted on the outer wall of the outer shell, and the rotating rods all penetrate the side wall of the outer shell, and the rotating rods are respectively fixedly connected with an external gear and an internal gear at one end outside the outer shell and one end inside the outer shell, and the external gears are meshed with racks, and the upper parts of the racks are fixedly connected to the top cover, and a plurality of limiting holes are opened on the top cover, and a plurality of limiting rods are fixedly connected to the top of the outer shell, and the limiting rods are slidably fitted with the limiting holes adjacent to them; a placement groove is opened in the middle of the outer shell; a nutrient box is fixedly connected to the bottom of the outer shell; and nutrient solution is stored in the nutrient box.

[0032] A toothed disc is rotatably mounted inside the outer shell, and the internal gears are meshed with the toothed disc. A spiral convex strip is fixedly connected to the top of the toothed disc, and a concave strip is fixedly connected to the bottom of the clamping block, and the concave strips are slidably matched with the convex strips. A driving member is fixedly connected inside the outer shell, and the output shaft of the driving member is coaxially fixedly connected to the adjacent internal gear. The controller is used to control the opening and closing of the driving member.

[0033] Beneficial effects: During the transportation process, the operator can place the collected samples into the placement tank and the sample roots into the nutrient box.

[0034] When putting the plants in, use clean water to clean the surface of the leaves, and rotate the plants during the process so that the plants can be cleaned evenly, and wipe the leaves with a sponge; this can effectively avoid the problem that if there are hydrophobic attachments on the surface of the leaves during the clean water cleaning process, it will be difficult for clean water to clean them. The key is to use clean water to clean the leaves and wipe the leaves with a sponge to improve the cleaning effect of the plants and improve the consistency of subsequent data.

[0035] The operator controls the output shaft of the driving member to rotate clockwise through the controller. The driving member drives the adjacent internal gear to rotate, the internal gear drives the toothed disc to rotate, and the toothed disc drives the remaining internal gears to rotate; the internal gear drives the rotating rod to rotate, the rotating rod drives the outer gear to rotate, and the outer gear drives the rack meshing with it to move upward.

[0036] When the rack moves upward, the rack will drive the top cover to move upward, so that the outer shell and the top cover move away from each other; at the same time, due to the rotation of the toothed disc, the convex strip will also rotate accordingly, and since the convex strip is spiral, the concave strip and the convex strip slide together, so due to the limiting effect of the convex strip shape and the thrust of the rotation of the convex strip, the concave strip will drive the clamping block to move outside the outer shell, thereby leaving the placement slot; at this time, the operator can put the collected sample into the placement slot, so that the sample root system is placed in the nutrient box.

[0037] After the sample is placed, the operator controls the output shaft of the driving member to rotate counterclockwise through the controller. At this time, the clamping blocks move toward the sample together to clamp the sample. At the same time, the outer shell and the top cover move closer to each other, so that the device is closed as a whole, maintaining a closed environment, thereby protecting the sample.

[0038] Furthermore, one end of the clamping block close to the placement groove is fixedly connected with an air bag.

[0039] Beneficial effects: During the clamping process, the airbag can adapt to samples of different volumes and magnitudes, making it easier for operators to transport large quantities of samples and samples of different root diameters. It can also avoid mechanical damage to the plants during the clamping process and improve the quality of the samples.

[0040] Furthermore, the top cover is hollow inside and the inner wall is fixedly connected with a protective layer, and the top cover is respectively fixedly connected with a photosynthetic rate meter, a transpiration rate meter, a plurality of heating wires and a plurality of lighting lamps.

[0041] Beneficial effects: During the transportation process, the protective layer can greatly reduce the mechanical damage caused by the collision between the sample and the inside of the device during transportation, and improve the quality of the sample. The photosynthetic rate meter and the transpiration rate meter are used to measure the photosynthetic rate and transpiration rate of the sample respectively, the heating wire is used to adjust the temperature during transportation, and the lighting lamp is used to adjust the light during transportation.

[0042] Furthermore, a one-way valve is fixedly connected to the side wall of the nutrient box, and the controller is used to control the opening and closing of the one-way valve.

[0043] Beneficial effects: During the transportation process, the operator can open the one-way valve through the controller to allow air to enter the nutrient box to optimize the transportation environment of the sample.

[0044] Furthermore, the clamping blocks are all in a stepped shape.

[0045] Beneficial effects: The stepped design can reduce the material consumption and weight of the clamping block, thereby reducing the cost and shipping weight of the device while maintaining the clamping area of ​​the clamping block.

[0046] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of a method for measuring nitrogen content in rice leaves based on deep learning in an embodiment of the present invention.

[0048] Figure 2 This is an axonometric diagram of a device for measuring nitrogen content in rice leaves based on deep learning in an embodiment of the present invention.

[0049] Figure 3 It is an axonometric view of the top cover in an embodiment of the present invention.

[0050] Figure 4 It is a front view of the toothed disc in the embodiment of the present invention.

[0051] Figure 5 1 is a top view of a toothed disc in an embodiment of the present invention.

[0052] Figure 6 It is a front view of the clamping block in the embodiment of the present invention.

[0053] The figure marks in the drawings of the specification include: 1, shell; 2, clamping block; 201, air bag; 202, concave strip; 3, rotating rod; 301, internal gear; 4, external gear; 5, rack; 6, top cover; 7, limit rod; 8, toothed disc; 801, convex strip; 9, nutrition box. DETAILED DESCRIPTION

[0054] The following is further described in detail through specific implementation methods:

[0055] Embodiment 1:

[0056] As attached Figure 1 As shown: A method for measuring nitrogen content in rice leaves based on deep learning, comprising the following steps:

[0057] S1, sample collection: collect a number of rice plants as samples and immerse the sample roots in nutrient solution.

[0058] S2, sample transportation: adjust the transportation parameters of the transportation environment according to the current environment, and transport the samples to the laboratory; record the activity indicators of the samples during the transportation process.

[0059] S3, sample processing: After being transported to the laboratory, the samples were washed with clean water and the dry and damaged rice leaves were removed.

[0060] S4, sample detection: using a spectrometer to perform spectral detection on the treated rice leaves to obtain spectral data of the rice leaves.

[0061] S5, intelligent analysis: Analyze the spectral data of rice leaves based on the deep learning model to obtain the current nitrogen content of rice leaves. The deep learning module is trained based on the spectral data of several rice leaves. Based on the compensation model, the current nitrogen content of rice leaves is compensated according to the transport parameters and activity indicators to obtain the actual nitrogen content of rice leaves; the compensation model is trained based on the linear change data of several transport parameters and activity indicators. Through a large amount of training of deep learning models and compensation models, the analysis of samples can be intelligent and automated, so as to quickly and accurately obtain the carbon content analysis results of samples. Compared with manual analysis, this method has higher analysis efficiency.

[0062] Among them, the transport parameters include the temperature, humidity and light of the transport environment, and the activity indicators include the photosynthetic rate, transpiration rate and stomatal conductance during sample transport.

[0063] The photosynthetic rate calculation formula is as follows:

[0064] V O2 =(Vx-Vt) / M (1).

[0065] Among them, V O2represents the oxygen generation rate, Vx represents the oxygen release amount, Vt represents the initial oxygen concentration), and M represents the leaf area.

[0066] The transpiration rate calculation formula is as follows:

[0067] Tr = Ox / (Mt×T) (2).

[0068] Among them, Tr represents transpiration rate, Ox represents transpiration water loss, Mt represents unit leaf area, and T represents time.

[0069] The formula for calculating stomatal conductance is as follows:

[0070] gs = E / (1.6×Vpd) (2).

[0071] Where gs is the stomatal conductance, E is the transpiration rate, Vpd is the water vapor pressure deficit at the leaf-air interface, and 1.6 is an empirical constant.

[0072] The specific implementation process is as follows: First, the operator collects several rice samples in the field, soaks the roots of all samples in nutrient solution, and adjusts the temperature, humidity and light of the transportation environment according to the appropriate temperature, humidity and light of the samples, thereby increasing the activity of the rice during transportation and reducing the loss of its carbon content, thereby optimizing the transportation environment of the rice, improving the quality of samples transported to the laboratory, and further improving the accuracy of the carbon content detection results.

[0073] During the transportation process, the operator records the photosynthetic rate and transpiration rate of the samples in real time, and then uses the formula to calculate the photosynthetic rate, transpiration rate and stomatal conductance of the samples during the transportation process, which can provide good assistance for the subsequent deep learning model and compensation model nitrogen content analysis and improve the accuracy of the analysis.

[0074] When the samples were transported to the laboratory, the temperature, humidity and light of the laboratory were also adjusted to the appropriate range for the samples to maintain the consistency of the environment; afterwards, the operators washed the samples with clean water and removed the dry and damaged rice leaves in the samples. After the processing was completed, the operators used a spectrometer to perform spectral detection on the rice leaves and obtained the spectral data of the rice leaves.

[0075] Finally, the deep learning model was used to analyze the spectral data of rice leaves to obtain the current nitrogen content of the rice leaves. Then, based on the compensation model, the current nitrogen content of the rice leaves was compensated according to the temperature, humidity and light of the transportation environment and the photosynthetic rate, transpiration rate and stomatal conductance of the rice during transportation to obtain the actual nitrogen content of the rice leaves.

[0076] The present invention analyzes the spectral data of rice leaves through a deep learning model, and can quickly and accurately estimate the nitrogen content of rice leaves. Compared with the prior art, the present invention further considers the temperature, humidity and light during the transportation process as well as the photosynthetic rate, transpiration rate and stomatal conductance of rice through a compensation model, and further corrects the estimated nitrogen content, thereby improving the accuracy of detection.

[0077] Embodiment 2:

[0078] As attached Figure 2-Figure 6 As shown: Different from the above-mentioned embodiment, the device for measuring the nitrogen content of rice leaves based on deep learning includes a controller, a shell 1 and a top cover 6; Figure 2 As shown, a plurality of clamping blocks 2 are slidably fitted on the outer shell 1; a plurality of rotating rods 3 are rotatably fitted on the outer wall of the outer shell 1, and the rotating rods 3 all penetrate the side wall of the outer shell 1, and the rotating rods 3 are respectively bolted to an external gear 4 and an internal gear 301 at one end outside the outer shell 1 and at one end inside the outer shell 1, and the external gear 4 is meshed with a rack 5.

[0079] like Figure 3 As shown, the upper part of the rack 5 is fixedly connected with the top cover 6 by bolts, a plurality of limit holes are opened on the top cover 6, a plurality of limit rods 7 are welded on the top of the outer shell 1, and the limit rods 7 are slidably matched with the limit holes adjacent thereto; a placement groove is opened in the middle of the outer shell 1; the bottom of the outer shell 1 is fixedly connected with a nutrient box 9 by bolts; nutrient solution is stored in the nutrient box 9.

[0080] like Figure 4 and Figure 5 As shown, a toothed disc 8 is rotatably mounted in the housing 1, and the internal gears 301 are meshed with the toothed disc 8. A spiral convex strip 801 is welded on the top of the toothed disc 8; Figure 6 As shown, the bottom of the clamping block 2 is welded with concave strips 202, and the concave strips 202 are slidably matched with the convex strips 801; a driving member is fixedly connected with bolts in the outer shell 1. In this embodiment, the driving member is a servo motor, and the output shaft of the servo motor is fixedly connected with the adjacent internal gear 301 by coaxial bolts; the controller is used to control the opening and closing of the servo motor.

[0081] The specific implementation process is as follows: During the transportation process, the operator can put the collected samples into the placement tank and put the sample roots into the nutrient box 9.

[0082] by Figure 3 and Figure 4 For example, after the sample collection is completed, the operator controls the servo motor output shaft to rotate clockwise through the controller, and the servo motor drives the adjacent internal gear 301 to rotate. Since the internal gear 301 is meshed with the toothed disc 8, the internal gear 301 will drive the toothed disc 8 to rotate, and the toothed disc 8 will drive the remaining internal gears 301 to rotate.

[0083] Since the inner gear 301 and the rotating rod 3 are fixedly connected by bolts, the inner gear 301 will drive the rotating rod 3 to rotate, and the rotating rod 3 will drive the outer gear 4 to rotate. Since the outer gear 4 is meshed with the adjacent rack 5, the outer gear 4 will drive the meshed rack 5 to move upward.

[0084] by Figure 2 and Figure 3 For example, when all the racks 5 move upward together, since the racks 5 and the top cover 6 are fixedly connected by bolts, the racks 5 will drive the top cover 6 to move upward, so that the housing 1 and the top cover 6 move away from each other.

[0085] At the same time, Figure 5 and Figure 6 For example, due to the rotation of the toothed disc 8, the convex strip 801 will also rotate accordingly. Since the convex strip 801 is spiral, the concave strip 202 slides with the convex strip 801. Therefore, the shapes of the convex strip 801 and the concave strip 202 fit each other and produce a limiting effect. At the same time, with the thrust of the rotation of the convex strip 801, the concave strip 202 will drive the clamping block 2 to move outside the shell 1, thereby leaving the placement groove; at this time, the operator can put the collected sample into the placement groove, so that the sample root system is placed in the nutrient box 9.

[0086] After the sample is placed, the operator controls the servo motor output shaft to rotate counterclockwise through the controller. At this time, the clamping blocks 2 move toward the sample together to clamp the sample, so that the outer shell 1 and the top cover 6 are close to each other, so that the device is closed as a whole, maintaining a closed environment, thereby protecting the sample.

[0087] by Figure 2 and Figure 3 For example, during this process, the limit rod 7 will slide in the limit hole. Due to the limiting effect of the limit rod 7, the top cover 6 will not rotate, so that the rack 5 and the outer gear 4 adjacent to it remain engaged; and due to the self-locking ability of the servo motor, the outer gear 4 will only rotate when the controller is started, so that the distance between the top cover 6 and the outer shell 1 can only be increased by the controller for regulation, and will not be affected by gravity, thereby ensuring the stable operation of the device.

[0088] Embodiment 3:

[0089] As attached Figure 1 and Figure 6 As shown: Different from the above-mentioned embodiment, an air bag 201 is fixedly bonded to one end of the clamping block 2 close to the placement groove.

[0090] The specific implementation process is as follows: during the clamping process, the airbag 201 can adapt to samples of different volumes and magnitudes, making it easier for operators to transport a large number of samples and samples of different root diameters. It can also avoid mechanical damage to the plants during the clamping process, thus improving the quality of the samples.

[0091] Embodiment 4:

[0092] As attached Figure 3 As shown: Different from the above embodiment, the top cover 6 is hollow inside and a protective layer is fixedly bonded to the inner wall, and a photosynthetic rate meter, a transpiration rate meter, a plurality of heating wires and a plurality of lighting lamps are fixedly connected to the top cover 6 by screws.

[0093] The specific implementation process is as follows: During the transportation process, the protective layer can greatly reduce the mechanical damage caused by the collision between the sample and the inside of the device during transportation, and improve the quality of the sample. The photosynthetic rate meter and the transpiration rate meter are used to measure the photosynthetic rate and transpiration rate of the sample respectively, the heating wire is used to adjust the temperature during transportation, and the lighting lamp is used to adjust the light during transportation.

[0094] Embodiment 5:

[0095] As attached Figure 2 As shown: the side wall of the nutrient box 9 is fixedly connected with a one-way valve by bolts, and the controller is used to control the opening and closing of the one-way valve.

[0096] The specific implementation process is as follows: During the transportation process, the operator can open the one-way valve through the controller to allow air to enter the nutrient box 9, thereby optimizing the transportation environment of the sample.

[0097] Embodiment 6:

[0098] As attached Figure 6 As shown: the clamping blocks 2 are all stepped.

[0099] The specific implementation process is as follows: the stepped design can reduce the consumption of materials for making the clamping block 2, while reducing the weight of the clamping block 2, thereby reducing the cost and shipping weight of the device, and while reducing the volume and gravity of the clamping block 2, it will not affect the clamping area of ​​the clamping block 2, thereby ensuring the stability of the clamping block 2 when clamping the sample.

[0100] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. A method for measuring nitrogen content in rice leaves based on deep learning, characterized in that: The following steps are involved: S1, sample collection: collect a number of rice plants as samples and immerse the sample roots in nutrient solution; S2, sample transportation: adjust the transportation parameters of the transportation environment according to the current environment, and transport the samples to the laboratory; record the activity indicators of the samples during the transportation process; S3, sample processing: After being transported to the laboratory, the samples were washed with clean water and the dried and damaged rice leaves were removed; S4, sample detection: using a spectrometer to perform spectral detection on the treated rice leaves to obtain spectral data of the rice leaves; S5, intelligent analysis: Analyze the spectral data of rice leaves based on the deep learning model to obtain the current nitrogen content of rice leaves; Based on the compensation model, the current nitrogen content of rice leaves is compensated according to the transport parameters and activity indicators to obtain the actual nitrogen content of rice leaves.

2. The method for measuring nitrogen content in rice leaves based on deep learning according to claim 1, characterized in that: In S2, the transport parameters include the temperature, humidity and light of the transport environment; the activity indicators include the photosynthetic rate, transpiration rate and stomatal conductance during the sample transport process; in S5, the deep learning module is trained based on the spectral data of several rice leaves; the compensation model is trained based on the linear change data of several transport parameters and activity indicators; The photosynthetic rate calculation formula is as follows: In O2 =(Vx-Vt) / M (1); Among them, V O2 represents the oxygen generation rate, Vx represents the oxygen release amount, Vt represents the initial oxygen concentration), and M represents the leaf area; The transpiration rate calculation formula is as follows: Tr = Ox / (Mt×T) (2); Among them, Tr represents transpiration rate, Ox represents transpiration water loss, Mt represents unit leaf area, and T represents time; The formula for calculating stomatal conductance is as follows: gs = E / (1.6×Vpd) (2); Where gs is the stomatal conductance, E is the transpiration rate, Vpd is the water vapor pressure deficit at the leaf-air interface, and 1.6 is an empirical constant.

3. A device for measuring nitrogen content in rice leaves based on deep learning, operating according to the method for measuring nitrogen content in rice leaves based on deep learning according to any one of claims 1-2, characterized in that: It comprises a controller, a housing (1) and a top cover (6), wherein a plurality of clamping blocks (2) are slidably fitted on the housing (1); The outer wall of the outer shell (1) is rotatably matched with a plurality of rotating rods (3), the rotating rods (3) all penetrate the side wall of the outer shell (1), the rotating rods (3) are respectively fixedly connected with an external gear (4) and an internal gear (301) at one end outside the outer shell (1) and at one end inside the outer shell (1), the external gears (4) are meshed with racks (5), and the upper parts of the racks (5) are fixedly connected with the top cover (6); A plurality of limiting holes are formed on the top cover (6), and a plurality of limiting rods (7) are fixedly connected to the top of the housing (1), and the limiting rods (7) are all slidably matched with the limiting holes adjacent thereto; A placement groove is provided in the middle of the outer shell (1); a nutrient box (9) is fixedly connected to the bottom of the outer shell (1); and nutrient solution is stored in the nutrient box (9); A toothed disc (8) is rotatably engaged in the housing (1), the internal gear (301) is meshed with the toothed disc (8), a spiral convex strip (801) is fixedly connected to the top of the toothed disc (8), and a concave strip (202) is fixedly connected to the bottom of the clamping block (2), and the concave strip (202) is slidably engaged with the convex strip (801); A driving member is fixedly connected inside the housing (1), and an output shaft of the driving member is coaxially fixedly connected to an adjacent internal gear (301); and a controller is used to control the opening and closing of the driving member.

4. The device for measuring nitrogen content in rice leaves based on deep learning according to claim 3, characterized in that: The clamping block (2) is fixedly connected to an air bag (201) at one end close to the placement groove.

5. The device for measuring nitrogen content in rice leaves based on deep learning according to claim 4, characterized in that: The top cover (6) is hollow inside and has a protective layer fixedly connected to the inner wall; A photosynthetic rate measuring instrument, a transpiration rate measuring instrument, a plurality of heating wires and a plurality of lighting lamps are respectively fixedly connected inside the top cover (6).

6. The device for measuring nitrogen content in rice leaves based on deep learning according to claim 5, characterized in that: A one-way valve is fixedly connected to the side wall of the nutrient box (9), and the controller is used to control the opening and closing of the one-way valve.

7. The device for measuring nitrogen content in rice leaves based on deep learning according to claim 6, characterized in that: The clamping blocks (2) are all in a stepped shape.

Images