A method, device, medium and product for identifying nitrogen nutrition surplus or deficiency of hydroponic vegetables in a fish-plant symbiotic system

Abstract

The application discloses a fish-vegetable symbiotic system water planting vegetable nitrogen nutrition gain-loss identification method, equipment, medium and product, and relates to the field of aquaculture. The method comprises the following steps: taking water planting vegetables in a fish-vegetable symbiotic system as a production group, and setting water planting vegetables with the same climate conditions and the best nutrient supply as a control group; collecting crown layer image data of the production group and the control group by using a depth camera; determining the nitrogen nutrition condition indexes of the vegetables in the production group and the control group based on a lightweight instance segmentation model according to the crown layer image data of the production group and the control group; taking the nitrogen nutrition condition indexes of the vegetables in the control group as a nitrogen nutrition gain-loss identification reference value; and judging the nitrogen nutrition gain-loss condition of the vegetables in the production group according to the nitrogen nutrition gain-loss identification reference value and the nitrogen nutrition condition indexes of the vegetables in the production group. The application can realize accurate identification of the nitrogen nutrition gain-loss condition of the vegetables that adapts to climate change.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of aquaculture, in particular to a method, device, medium and product for identifying nitrogen nutrition surplus or deficiency of hydroponic vegetables in a fish-vegetable symbiotic system. BACKGROUND

[0002] The fish-vegetable symbiotic system is a new type of green ecological and environmentally friendly zero-pollution farming mode. It effectively connects the nitrogen nutrition supply and demand relationship between aquaculture and hydroponic vegetables, achieving the goal of low water exchange rate, low environmental pollution, and low fertilizer addition. However, there are still many problems in the actual production process of the existing fish-vegetable symbiotic system: (1) Managers generally pay more attention to the economic value of fish and only consider vegetables as a water purification tool, ignoring the economic value of vegetables. (2) The life cycle of fish and vegetables is asynchronous, and the optimal growth of vegetables cannot be guaranteed. The nitrogen nutrition required for vegetable growth mainly comes from the tail water of aquaculture, and the richness of nitrogen nutrition in the tail water gradually increases with the increase of the aquaculture stage. The nitrogen demand of vegetables is small in the early growth stage, and large in the middle and late growth stages. Insufficient nitrogen supply will lead to slow growth of vegetables, and further lead to nitrogen deficiency of vegetables. (3) The nitrogen absorption capacity of vegetables is different under different seasons and different climate conditions, resulting in differences in external growth performance. (4) There are differences in the growth of individual plants, and the identification of nitrogen surplus or deficiency of a single point is one-sided. Therefore, the accurate identification of the nitrogen surplus or deficiency of vegetables in the fish-vegetable symbiotic system is the key to the subsequent optimal growth regulation of crops.

[0003] Currently, there is almost no research on the diagnosis of the nitrogen nutrition status of vegetables in the fish-vegetable symbiotic system. Most existing researches use spectral technology and visual technology to collect field or facility crop data and build nitrogen nutrition status diagnosis models, ignoring the influence of climate differences in different regions or different seasons on the nitrogen absorption capacity of crops. Although the nitrogen response data set of crops under different regional and seasonal conditions can be collected to build a model to ensure the applicability of the model, the data collection process is very time-consuming and requires a large amount of manpower and material resources. SUMMARY

[0004] The purpose of the present application is to provide a method, device, medium and product for identifying the nitrogen surplus or deficiency of hydroponic vegetables in a fish-vegetable symbiotic system, which can accurately identify the nitrogen surplus or deficiency of vegetables under climate change.

[0005] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0006] In a first aspect, the present application provides a method for identifying the nitrogen surplus or deficiency of hydroponic vegetables in a fish-vegetable symbiotic system, comprising:

[0007] The water-grown vegetables in the fish-plant symbiotic system are taken as a production group, and water-grown vegetables in the same climate conditions and with optimal nutrient supply as a control group;

[0008] A deep camera is used to collect canopy image data of the production group and the control group;

[0009] Based on a lightweight instance segmentation model, the nitrogen nutrition status indicators of the production group and the control group are determined according to the canopy image data of the production group and the control group; the nitrogen nutrition status indicators include canopy coverage mean value and plant height mean value; the lightweight instance segmentation model is constructed based on a YOLOv8 instance segmentation model;

[0010] The nitrogen nutrition status indicators of the control group are taken as nitrogen nutrition gain-loss identification reference values; the nitrogen nutrition gain-loss identification reference values include canopy coverage reference value and plant height reference value;

[0011] According to the nitrogen nutrition gain-loss identification reference values and the nitrogen nutrition status indicators of the production group, the nitrogen nutrition gain-loss status of the production group is determined.

[0012] Optionally, the lightweight instance segmentation model takes the YOLOv8 instance segmentation model as a reference model, replaces the backbone network of the reference model with a StarNet structure, replaces the batch normalization in the head network of the reference model with group normalization, and replaces five 3x3 convolution layers in the head network of the reference model with a DEConv structure.

[0013] Optionally, based on the lightweight instance segmentation model, the nitrogen nutrition status indicators of the production group and the control group are determined according to the canopy image data of the production group and the control group, including:

[0014] The canopy image data of the production group and the control group are respectively input into the lightweight instance segmentation model to obtain the target regions of different plant individuals within the corresponding current field of view;

[0015] The ratio of the number of pixels of the target regions of different plant individuals within the corresponding current field of view of the production group to the number of pixels of the canopy image data of the production group is calculated to obtain the group plant canopy coverage of the production group, and the group plant canopy coverage of the production group is divided by the number of plants in the production group to obtain the canopy coverage mean value of the production group;

[0016] The average value of the height values of the center points of the target regions of different plant individuals within the corresponding current field of view of the production group is calculated to obtain the plant height mean value of the production group;

[0017] The ratio of the number of pixels of the target region of each plant in the current view of the control group to the number of pixels of the canopy image data of the control group is calculated to obtain the canopy coverage of the control group, and the canopy coverage of the control group is divided by the number of plants in the control group to obtain the average canopy coverage of the control group.

[0018] The average height value of the center point of the target region of each plant in the current view of the control group is calculated to obtain the average plant height of the control group.

[0019] Optionally, the nitrogen nutrition status of the vegetables in the production group is determined according to the nitrogen nutrition identification reference value and the nitrogen nutrition status indicator of the vegetables in the production group, including:

[0020] If the average canopy coverage of the production group is less than the canopy coverage reference value, and the average plant height of the production group is less than the plant height reference value, the nitrogen nutrition status of the vegetables in the production group is nitrogen deficiency state;

[0021] If the average canopy coverage of the production group is equal to the canopy coverage reference value, and the average plant height of the production group is equal to the plant height reference value, the nitrogen nutrition status of the vegetables in the production group is nitrogen optimal state;

[0022] If the average canopy coverage of the production group is greater than the canopy coverage reference value, and the average plant height of the production group is greater than the plant height reference value, the nitrogen nutrition status of the vegetables in the production group is nitrogen excess state;

[0023] If the average canopy coverage of the production group is less than the canopy coverage reference value, and the average plant height of the production group is greater than the plant height reference value, or the average canopy coverage of the production group is greater than the canopy coverage reference value, and the average plant height of the production group is less than the plant height reference value, the vegetables in the production group grow abnormally or the data collection and calculation are abnormal, and the nitrogen nutrition status cannot be determined.

[0024] Optionally, the maximum allowed difference between the average canopy coverage of the production group and the canopy coverage reference value is 5%, when the absolute value of the difference between the average canopy coverage of the production group and the canopy coverage reference value is less than or equal to 5%, the average canopy coverage of the production group is considered to be equal to the canopy coverage reference value, otherwise the average canopy coverage of the production group is considered not to be equal to the canopy coverage reference value; the maximum allowed difference between the average plant height of the production group and the plant height reference value is 0.2 cm, when the absolute value of the difference between the average plant height of the production group and the plant height reference value is less than or equal to 0.2 cm, the average plant height of the production group is considered to be equal to the plant height reference value, otherwise the average plant height of the production group is considered not to be equal to the plant height reference value.

[0025] Optionally, the depth camera is arranged parallel to a plane of the culture tank of the fish-vegetable symbiotic system, and a height of the depth camera from the plane of the culture tank of the fish-vegetable symbiotic system is 50 cm.

[0026] Optionally, the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition balance identification method further comprises:

[0027] Different nitrogen treatment gradients are set according to a range of variation of nitrogen content in the tail water of the fish-vegetable symbiotic system.

[0028] The sample vegetables are transplanted into culture tanks under different nitrogen treatment gradients for cultivation, and a depth camera is used to collect canopy image data of the sample vegetables in a whole growth period.

[0029] The canopy image data of the sample vegetables in the whole growth period are input into a lightweight instance segmentation model to obtain target regions of different plant individuals in a current field of view of the sample vegetables.

[0030] Corresponding canopy coverage mean values and plant height mean values are determined according to the target regions of different plant individuals in the current field of view of the sample vegetables under different nitrogen treatment gradients.

[0031] A canopy coverage mean value variation curve and a plant height mean value variation curve under different nitrogen treatment gradients and different growth days are drawn, and a nitrogen nutrition balance identification reference line is determined; and the nitrogen treatment gradient of the control group is set based on the nitrogen nutrition balance identification reference line.

[0032] In a second aspect, the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition balance identification method.

[0033] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition balance identification method.

[0034] In a fourth aspect, the present application provides a computer program product comprising a computer program, wherein the computer program is executable by a processor to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition balance identification method.

[0035] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0036] The application provides a fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition surplus / deficiency identification method, device, medium and product. By additionally adding a group of water-grown vegetables as a control group under the same climate conditions as the production group and using the optimal nutrient supply, the control group can adapt to climate change and give the most standard nitrogen nutrition surplus / deficiency identification benchmark with the canopy coverage mean value and the plant height mean value as the discriminant indexes, so as to realize the accurate identification of the nitrogen nutrition surplus / deficiency status of the vegetables adapting to climate change. In addition, the construction of the lightweight instance segmentation model realizes the automatic collection of the canopy coverage and the plant height of the individual plants, which can improve the accuracy and efficiency of the vegetable nitrogen nutrition surplus / deficiency identification. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the application or the related art, the drawings needed to be used in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0038] Figure 1 A fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition surplus / deficiency identification method flowchart is provided for the application.

[0039] Figure 2 A canopy image data acquisition schematic diagram is provided for the application.

[0040] Figure 3 A lightweight improved YOLOv8 instance segmentation model network structure diagram is provided for the application.

[0041] Figure 4 A lightweight improved YOLOv8 instance segmentation model head network structure diagram is provided for the application.

[0042] Figure 5 A canopy coverage mean value change curve and a nitrogen nutrition surplus / deficiency identification benchmark line diagram of the canopy coverage mean value are provided for the application.

[0043] Figure 6 A plant height mean value change curve and a nitrogen nutrition surplus / deficiency identification benchmark line diagram of the plant height mean value are provided for the application.

[0044] Figure 7 A fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition surplus / deficiency identification method implementation diagram is provided for the application. DETAILED DESCRIPTION

[0045] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts are within the scope of the present application.

[0046] The present application aims at the problem of difficulty in identifying the nitrogen nutrition surplus or deficiency of vegetables in fish-vegetable symbiotic system under different climate conditions, caused by the imbalance between supply and demand of nitrogen in tail water of aquaculture under asynchronous life cycle of fish and vegetables, and different nitrogen absorption capacity of vegetables. The present application realizes accurate identification of the nitrogen nutrition status of vegetables in fish-vegetable symbiotic system by combining the steps of vegetable nitrogen surplus or deficiency test design, multi-modal data acquisition, lightweight instance segmentation model construction, core evaluation parameter determination, nitrogen nutrition surplus or deficiency identification strategy and system construction.

[0047] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0048] In an exemplary embodiment, the present application provides a method for identifying nitrogen nutrition surplus or deficiency of hydroponic vegetables in fish-vegetable symbiotic system. As shown in Figure 1 the method includes the following steps 1 to 5.

[0049] Step 1: The hydroponic vegetables in the fish-vegetable symbiotic system are taken as the production group, and the hydroponic vegetables under the same climate conditions and with the best nutrient supply as the control group.

[0050] Step 2: The crown layer image data of the production group and the control group are collected by using a depth camera.

[0051] The depth camera is parallel to the plane of the culture tank of the fish-vegetable symbiotic system, and the height from the plane of the culture tank of the fish-vegetable symbiotic system is 50 cm. Exemplarily, the depth camera is an Intel realsense D435i depth camera.

[0052] Step 3: Based on a lightweight instance segmentation model, the nitrogen nutrition status indicators of the production group and the control group are determined according to the crown layer image data of the production group and the control group. The nitrogen nutrition status indicators of the vegetables include the mean value of the crown layer coverage and the mean value of the plant height.

[0053] The lightweight instance segmentation model is constructed based on a YOLOv8 instance segmentation model. Specifically, the lightweight instance segmentation model takes the YOLOv8 instance segmentation model as a benchmark model, replaces the backbone network of the benchmark model with a StarNet structure, replaces the batch normalization in the head network of the benchmark model with group normalization, and replaces five 3*3 convolution layers in the head network of the benchmark model with a DEConv structure.

[0054] Step 3 specifically includes: inputting the canopy image data of the production group and the control group into the lightweight instance segmentation model respectively to obtain the target regions of different plant individuals in the corresponding current field of view; calculating the ratio of the number of pixels of the target regions of different plant individuals in the current field of view of the production group to the number of pixels of the canopy image data of the production group to obtain the canopy coverage of the production group, and dividing the canopy coverage of the production group by the number of plants in the production group to obtain the mean canopy coverage of the production group; calculating the average height value of the center points of the target regions of different plant individuals in the current field of view of the production group to obtain the mean plant height of the production group; calculating the ratio of the number of pixels of the target regions of different plant individuals in the current field of view of the control group to the number of pixels of the canopy image data of the control group to obtain the canopy coverage of the control group, and dividing the canopy coverage of the control group by the number of plants in the control group to obtain the mean canopy coverage of the control group; calculating the average height value of the center points of the target regions of different plant individuals in the current field of view of the control group to obtain the mean plant height of the control group.

[0055] Step 4: The nitrogen nutrition status indicators of the control group are used as the nitrogen nutrition gain-loss identification benchmark value. The nitrogen nutrition gain-loss identification benchmark value includes a canopy coverage benchmark value and a plant height benchmark value.

[0056] Step 5: According to the nitrogen nutrition gain-loss identification benchmark value and the nitrogen nutrition status indicators of the production group, the nitrogen nutrition gain-loss status of the production group is determined.

[0057] The step 5 specifically comprises: if the canopy coverage average value of the production group is less than the canopy coverage reference value, and the plant height average value of the production group is less than the plant height reference value, then the nitrogen nutrition status of the vegetables of the production group is in a nitrogen deficiency state; if the canopy coverage average value of the production group is equal to the canopy coverage reference value, and the plant height average value of the production group is equal to the plant height reference value, then the nitrogen nutrition status of the vegetables of the production group is in a nitrogen optimum state; if the canopy coverage average value of the production group is greater than the canopy coverage reference value, and the plant height average value of the production group is greater than the plant height reference value, then the nitrogen nutrition status of the vegetables of the production group is in a nitrogen excess state; if the canopy coverage average value of the production group is less than the canopy coverage reference value, and the plant height average value of the production group is greater than the plant height reference value, or the canopy coverage average value of the production group is greater than the canopy coverage reference value, and the plant height average value of the production group is less than the plant height reference value, then the vegetables of the production group are abnormal or the data collection and calculation are abnormal, and the nitrogen nutrition status cannot be determined.

[0058] The maximum allowed difference between the canopy coverage average value of the production group and the canopy coverage reference value is 5%, and when the absolute value of the difference between the canopy coverage average value of the production group and the canopy coverage reference value is less than or equal to 5%, the canopy coverage average value of the production group is considered to be equal to the canopy coverage reference value, otherwise, the canopy coverage average value of the production group is considered to be not equal to the canopy coverage reference value; the maximum allowed difference between the plant height average value of the production group and the plant height reference value is 0.2 cm, and when the absolute value of the difference between the plant height average value of the production group and the plant height reference value is less than or equal to 0.2 cm, the plant height average value of the production group is considered to be equal to the plant height reference value, otherwise, the plant height average value of the production group is considered to be not equal to the plant height reference value.

[0059] Further, the method further comprises: setting different nitrogen treatment gradients according to the nitrogen content variation range of the aquaponics system; transplanting the sample vegetables into the cultivation tanks under different nitrogen treatment gradients for cultivation, and collecting the canopy image data of the sample vegetables in the whole growth period by using a depth camera; inputting the canopy image data of the sample vegetables in the whole growth period into the lightweight instance segmentation model to obtain the target regions of different plant individuals in the current field of view of the sample vegetables; determining the corresponding canopy coverage average value and plant height average value according to the target regions of different plant individuals in the current field of view of the sample vegetables under different nitrogen treatment gradients; drawing the canopy coverage average value variation curve and the plant height average value variation curve under different nitrogen treatment gradients and different growth days, and determining a nitrogen nutrition identification reference line. The nitrogen treatment gradient of the control group is set based on the nitrogen nutrition identification reference line.

[0060] The following will be described in combination with Figures 2 to 7The specific implementation process of the above fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition surplus / deficiency identification method is described in detail, including vegetable nitrogen surplus / deficiency test design, multi-modal data acquisition, lightweight instance segmentation model construction, core evaluation parameter determination, nitrogen nutrition surplus / deficiency identification strategy and system construction.

[0061] The first step is to design a vegetable nitrogen surplus / deficiency test. The main purpose is to design different nitrogen treatment tests according to the nitrogen content range of the tail water in the fish-vegetable symbiotic system, which lays the foundation for the accurate collection of multi-modal data in the follow-up.

[0062] During the whole life cycle of fish-vegetable symbiotic system from fry to adult fish, the nitrogen content of the tail water after biochemical treatment is about 30-350 mg / L. In order to keep consistent with the actual farming scene, the present application takes Yamazaki salad nutrient solution formula as the basis, takes water-grown salad as the research object, and sets different nitrogen treatment gradients to analyze the nitrogen response ability of water-grown salad. The specific nitrogen treatment gradient is shown in Table 1 (where the nitrogen concentration of different nitrogen treatment gradients is an arithmetic sequence, the difference is 76.18375 mg / L, and the nitrogen concentration in the table is rounded to two decimal places), wherein N0, N1, N2, N3 and N4 belong to nitrogen deficiency treatment, N5 belongs to optimal nitrogen treatment, and N6, N7 and N8 belong to nitrogen excess treatment. Compared with the existing research on nitrogen treatment gradient, the nitrogen treatment gradient in the present application is more intensive, which can more accurately reflect the nitrogen response ability of the vegetable.

[0063] Table 1 Water-grown vegetable nitrogen treatment gradient table

[0064]

[0065] The second step is to collect multi-modal data, which is mainly to collect RGB images and depth images at the same time, so as to facilitate the acquisition of plant individual canopy coverage and plant height in the follow-up.

[0066] Based on the first step of vegetable nitrogen surplus / deficiency test design, after the vegetables are transplanted into the cultivation tank under different nitrogen treatments, an Intel realsense D435i depth camera is connected to a notebook computer, the camera plane is parallel to the cultivation tank plane and the height is kept at 50 cm. The canopy image data (including RGB images and depth images) of the vegetables under different nitrogen treatments are collected. During the data collection process, when the camera collects the current area, it is parallelly slid to the next area, and the plant samples in the front and back areas cannot be repeated. The specific data collection diagram is shown in Figure 2 .

[0067] The third step is to construct a lightweight instance segmentation model, which is mainly to realize the automatic collection of plant individual canopy coverage and plant height, and to improve the accuracy and efficiency of vegetable nitrogen nutrition surplus / deficiency identification.

[0068] Based on the second step of collecting the vegetable canopy image data under different nitrogen treatments, the data labeling software labelme is used to label the RGB image in the collected data. After all the data labeling is completed, the whole data set is divided into training set, validation set and test set. In order to ensure the prediction accuracy and inference speed of the model in the actual application process, the YOLOv8 instance segmentation model is taken as the benchmark model, and the main network and head network of the YOLOv8 instance segmentation model are improved to meet the lightweight deployment requirements of the actual production scene under the premise of ensuring the minimum loss of model accuracy. Among them, the StarNet structure is used to replace the main network of the original YOLOv8 instance segmentation model, which can make the model capture more rich and complex feature information without significantly increasing the computational complexity; the group normalization is used to replace the batch normalization in the head network of the original YOLOv8 instance segmentation model, the DEConv structure is used to replace the 5 3×3 convolution layers in the head network of the original YOLOv8 instance segmentation model, and the weight sharing is set for part of the convolution layers to minimize the loss of model performance under the condition of reducing the parameter amount. The network structure of the lightweight improved YOLOv8 instance segmentation model is shown in Figure 3 Figure 4

[0069] Fourth step: core evaluation parameter determination, mainly through analyzing the change curve of plant individual canopy coverage mean and plant height mean under different nitrogen treatments, and then determining the nitrogen nutrition gain and loss recognition reference line.

[0070] ​​Based on the lightweight instance segmentation model obtained in the third step, an Intel realsense D435i depth camera is connected to a notebook computer, the camera plane is kept parallel to the cultivation tank plane, and the height is kept at 50 cm. The canopy image data of the vegetables under different nitrogen treatments is collected during the whole growth period of the vegetables, and the pre-trained lightweight instance segmentation model is called to output the target area of different plant individuals in the current field of view by taking the obtained canopy image as the input. By calculating the ratio of the number of pixels of the vegetable plant area to the number of pixels of the entire input image, the canopy coverage of the group plant can be obtained. Then, by dividing the canopy coverage of the group plant by the number of plants, the average canopy coverage can be obtained. The Intel realsense D435i depth camera itself can obtain the height value in the vegetable plant area. In this application, the height value of the center point of each vegetable segmentation target area is obtained, and then the average plant height is obtained by averaging. Among them, the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N0 treatment are respectively Ci0 and Hi0; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N1 treatment are respectively Ci1 and Hi1; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N2 treatment are respectively Ci2 and Hi2; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N3 treatment are respectively Ci3 and Hi3; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N4 treatment are respectively Ci4 and Hi4; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N5 treatment are respectively Ci5 and Hi5; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N6 treatment are respectively Ci6 and Hi6; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N7 treatment are respectively Ci7 and Hi7; the average canopy coverage and the average plant height of different plant individuals in the current field of view under the N8 treatment are respectively Ci8 and Hi8. C represents the average canopy coverage, H represents the average plant height, i represents the growth day, and 0-8 represents different nitrogen treatments.

[0071] The average canopy coverage change curve and the average plant height change curve under different nitrogen treatments and different growth days are drawn, the differences in the average canopy coverage and the average plant height under different nitrogen treatments are analyzed, and the average canopy coverage change curve and the average plant height change curve under the best nitrogen treatment (N5) are selected as the nitrogen nutrition gain and loss identification reference line, which provides core evaluation parameter support basis for the accurate identification of the nitrogen nutrition gain and loss of vegetables in subsequent actual production. Specifically as Figure 5 and Figure 6As shown, the nitrogen nutrition surplus and deficiency identification reference line obtained by different nitrogen treatment experiments can illustrate the change rule of the average canopy coverage and the average plant height with the increase of the growth days under the conditions of nitrogen deficiency, nitrogen optimum and nitrogen excess, and the relationship between them, indicating that the average canopy coverage and the average plant height can reflect the nitrogen nutrition status of vegetables.

[0072] The fifth step is to establish a nitrogen nutrition surplus and deficiency identification strategy and system, which mainly compares the core evaluation indexes of the control group and the production group by setting the control group and the production group and using the nitrogen nutrition surplus and deficiency identification reference line, and then determines the current nitrogen nutrition surplus and deficiency status of the vegetables.

[0073] Based on the lightweight instance segmentation model constructed in the third step and the nitrogen nutrition surplus and deficiency identification reference line with the average canopy coverage and the average plant height as the discrimination indexes determined in the fourth step, in the actual production process of the fish-vegetable symbiotic system, in order to better adapt to the influence of climate change on the precise diagnosis of the nitrogen nutrition surplus and deficiency of vegetables, the present application proposes a best nutrition supply reference strategy based on climate self-adaptation. The strategy additionally increases a group of best nutrition supply control groups in actual production, and compares the core evaluation parameters (i.e. the nitrogen nutrition status indexes of vegetables, including the average canopy coverage and the average plant height) of the vegetables in different areas of the cultivation tank of the control group and the production group to judge the nitrogen nutrition surplus and deficiency status of the vegetables in the fish-vegetable symbiotic system under the current climate conditions. Since the growth rate of vegetables is not completely consistent under different geographical conditions and different climate conditions, but the relationship between nitrogen deficiency, nitrogen optimum and nitrogen excess embodied by the above-mentioned nitrogen nutrition surplus and deficiency identification reference line is determined. Therefore, in order to adapt to climate change and geographical change, a best nutrition supply control group needs to be separately set. At this time, the average canopy coverage and the average plant height obtained by the best nutrition supply control group with the increase of the growth days are the reference values on the nitrogen nutrition surplus and deficiency identification reference line. Among them, the average canopy coverage and the average plant height obtained by the control group (i.e. the canopy coverage reference value and the plant height reference value) are named as CC and CH, respectively, and the average canopy coverage and the average plant height obtained by the production group are named as PC and PH, respectively. The specific demonstration is shown in Figure 7 .

[0074] The specific vegetable nitrogen nutrition surplus and deficiency status identification criteria are as follows:

[0075] (1) If CC>PC and CH>PH in the data obtained at different periods, it indicates that the vegetables in the fish-vegetable symbiotic system at the current stage are in a state of nitrogen deficiency.

[0076] (2) If CC≈PC (|CC-PC|≤5%) and CH≈PH (|CH-PH|≤0.2 cm) in the data obtained at different periods, it indicates that the vegetables in the fish-vegetable symbiotic system at the current stage are in a state of nitrogen optimum.

[0077] (3) If CC< PC and CH< PH in the data collected at different time periods, it indicates that the vegetables in the current stage of the fish-vegetable symbiosis system are in a state of nitrogen excess.

[0078] (4) If CC> PC and CH< PH in the data collected at different time periods, it indicates that the vegetables in the current stage of the fish-vegetable symbiosis system grow abnormally (may be affected by continuous high temperature weather to cause bolting), and the nitrogen nutrition surplus or deficiency status cannot be judged.

[0079] (5) If CC< PC and CH> PH in the data collected at different time periods, it indicates that the vegetables in the current stage of the fish-vegetable symbiosis system grow abnormally (or the data collection and calculation are abnormal), and the nitrogen nutrition surplus or deficiency status cannot be judged.

[0080] Through the implementation of the above strategy, no matter which region or climate condition the fish-vegetable symbiosis system is in, the additional set of optimal nutrient supply control groups of the present application can always adapt to climate change to give the most standard nitrogen nutrition surplus or deficiency recognition benchmark with the canopy coverage mean value and the plant height mean value as the discriminant indicators, thereby realizing the accurate recognition of the nitrogen nutrition surplus or deficiency status of the vegetables adapting to climate change.

[0081] In an exemplary embodiment, the present application also provides a computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps in the above method embodiments when executing the computer program.

[0082] In an exemplary embodiment, the present application also provides a computer readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the above method embodiments.

[0083] In an exemplary embodiment, the present application also provides a computer program product comprising a computer program, which, when executed by a processor, implements the steps in the above method embodiments.

[0084] In the present application, all actions of obtaining signals, information or data are performed under the premise of complying with the corresponding data protection regulations and policies of the country where the device is located, and under the premise of obtaining authorization from the corresponding device owner. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant legal regulations.

[0085] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0086] The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a blockchain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0087] The technical features of the above embodiments can be combined arbitrarily. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.

[0088] The principles and implementation manners of the present application are described herein by using specific examples, and the above examples are only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will have changes. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A fish-plant symbiotic system water-grown vegetable nitrogen nutrition balance and deficiency identification method, characterized in that, The method comprises the following steps: Taking the hydroponic vegetables in the fish-vegetable symbiotic system as a production group, and setting a control group of hydroponic vegetables under the same climate conditions and with optimal nutrient supply as the production group; Collecting the crown layer image data of the production group and the control group by using a depth camera; Determining the nitrogen nutrition status indicators of the production group and the control group based on a lightweight instance segmentation model according to the crown layer image data of the production group and the control group; The nitrogen nutrition status indicators of the production group and the control group comprise a mean value of crown layer coverage and a mean value of plant height, and the lightweight instance segmentation model is constructed based on a YOLOv8 instance segmentation model, wherein the lightweight instance segmentation model takes the YOLOv8 instance segmentation model as a benchmark model, replaces the backbone network of the benchmark model with a StarNet structure, replaces the batch normalization in the head network of the benchmark model with group normalization, and replaces five 3x3 convolution layers in the head network of the benchmark model with a DEConv structure; Taking the nitrogen nutrition status indicators of the control group as a nitrogen nutrition gain-loss identification benchmark value, wherein the nitrogen nutrition gain-loss identification benchmark value comprises a crown layer coverage benchmark value and a plant height benchmark value; Judging the nitrogen nutrition gain-loss status of the production group according to the nitrogen nutrition gain-loss identification benchmark value and the nitrogen nutrition status indicators of the production group. 2.The fish-manaure system water-grown vegetable nitrogen nutrition balance identification method according to claim 1, characterized in that, Determining the nitrogen nutrition status indicators of the production group and the control group based on a lightweight instance segmentation model according to the crown layer image data of the production group and the control group comprises the following steps: Inputting the crown layer image data of the production group and the control group into the lightweight instance segmentation model respectively to obtain the target regions of different plant individuals in the current field of view corresponding to the production group and the control group; Calculating the ratio of the number of pixels of the target regions of different plant individuals in the current field of view corresponding to the production group to the number of pixels of the crown layer image data of the production group to obtain the group plant crown layer coverage of the production group, and dividing the group plant crown layer coverage of the production group by the number of plants of the production group to obtain the mean value of the crown layer coverage of the production group; Calculating the average value of the height values of the center points of the target regions of different plant individuals in the current field of view corresponding to the production group to obtain the mean value of the plant height of the production group; Calculating the ratio of the number of pixels of the target regions of different plant individuals in the current field of view corresponding to the control group to the number of pixels of the crown layer image data of the control group to obtain the group plant crown layer coverage of the control group, and dividing the group plant crown layer coverage of the control group by the number of plants of the control group to obtain the mean value of the crown layer coverage of the control group; Calculating the average value of the height values of the center points of the target regions of different plant individuals in the current field of view corresponding to the control group to obtain the mean value of the plant height of the control group. 3.The fish-manaure system water-grown vegetable nitrogen nutrition balance identification method according to claim 1, characterized in that, Judging the nitrogen nutrition gain-loss status of the production group according to the nitrogen nutrition gain-loss identification benchmark value and the nitrogen nutrition status indicators of the production group comprises the following steps: If the mean value of the crown layer coverage of the production group is less than the crown layer coverage benchmark value, and the mean value of the plant height of the production group is less than the plant height benchmark value, the nitrogen nutrition gain-loss status of the production group is nitrogen deficiency state; If the mean value of the crown layer coverage of the production group is equal to the crown layer coverage benchmark value, and the mean value of the plant height of the production group is equal to the plant height benchmark value, the nitrogen nutrition gain-loss status of the production group is nitrogen optimal state. If the average canopy coverage of the production group is greater than the reference value of the canopy coverage, and the average plant height of the production group is greater than the reference value of the plant height, the nitrogen nutrition status of the vegetables in the production group is in a nitrogen surplus state; If the average canopy coverage of the production group is less than the reference value of the canopy coverage, and the average plant height of the production group is greater than the reference value of the plant height, or the average canopy coverage of the production group is greater than the reference value of the canopy coverage, and the average plant height of the production group is less than the reference value of the plant height, the vegetables in the production group grow abnormally or the data collection and calculation are abnormal, and the nitrogen nutrition status cannot be determined.

4. The fish-microorganism symbiotic system hydroponic vegetable nitrogen nutrition balance and loss identification method according to claim 3, characterized in that, The maximum allowed difference between the average canopy coverage of the production group and the reference value of the canopy coverage is 5%. When the absolute value of the difference between the average canopy coverage of the production group and the reference value of the canopy coverage is less than or equal to 5%, the average canopy coverage of the production group is considered to be equal to the reference value of the canopy coverage, otherwise, the average canopy coverage of the production group is considered to be not equal to the reference value of the canopy coverage. The maximum allowed difference between the average plant height of the production group and the reference value of the plant height is 0.2 cm. When the absolute value of the difference between the average plant height of the production group and the reference value of the plant height is less than or equal to 0.2 cm, the average plant height of the production group is considered to be equal to the reference value of the plant height, otherwise, the average plant height of the production group is considered to be not equal to the reference value of the plant height.

5. The fish-microalgae symbiotic system hydroponic vegetable nitrogen nutrition balance and loss identification method according to claim 1, characterized in that, The depth camera is arranged parallel to the plane of the culture tank of the fish-vegetable symbiotic system, and the height from the plane of the culture tank of the fish-vegetable symbiotic system is 50 cm. 6.The fish-manaure system water-grown vegetable nitrogen nutrition balance and deficiency identification method according to claim 1, characterized in that, Further comprising: setting different nitrogen treatment gradients according to the range of nitrogen content change of the breeding tail water in the fish-vegetable symbiotic system; transplanting the sample vegetables into the cultivation tanks under different nitrogen treatment gradients for cultivation, and collecting the canopy image data of the sample vegetables in the whole growth period by using the depth camera; inputting the canopy image data of the sample vegetables in the whole growth period into the lightweight instance segmentation model to obtain the target regions of different plant individuals in the current field of view of the sample vegetables; determining the corresponding average canopy coverage and average plant height according to the target regions of different plant individuals in the current field of view of the sample vegetables under different nitrogen treatment gradients; drawing the average canopy coverage change curve and the average plant height change curve under different nitrogen treatment gradients and different growth days, and determining the nitrogen nutrition gain-loss identification reference line; the nitrogen treatment gradient of the control group is set based on the nitrogen nutrition gain-loss identification reference line.

7. A computer device comprising: The memory, the processor, and the computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition gain-loss identification method of any one of claims 1-6.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition gain-loss identification method of any one of claims 1-6.

9. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the fish-vegetable symbiotic system water-grown vegetable nitrogen nutrition gain-loss identification method of any one of claims 1-6.

Images