Low-cadmium-adsorption potato cultivation system

Abstract

The invention relates to the technical field of intelligent agriculture, in particular to a low-cadmium-adsorption potato cultivation system which comprises a genotype-phenotype prediction module used for receiving genotype data of potato germplasm resources, processing the genotype data by adopting a basic genotype-phenotype correlation model based on machine learning, and obtaining a genotype-phenotype prediction result; predicting and outputting cadmium adsorption potential rating data of the potato germplasm resources to a breeding and screening decision module; the breeding and screening decision module is used for outputting breeding scheme data containing optimal germplasm identification and cultivation parameters by adopting a multi-objective optimization algorithm; and the breeding and screening decision module also performs iterative optimization on the basic genotype-phenotype association model and / or a multi-objective optimization algorithm according to the cadmium adsorption capacity evaluation result data to form closed-loop feedback. According to the method, the mode of combining the basic model with online iterative optimization is adopted, rapid prediction of the cadmium adsorption potential of the potatoes and full-process automation and remote management of low-cadmium and high-yield targets are achieved, and the breeding efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural technology, specifically to a cadmium-low adsorption potato cultivation system. Background Technology

[0002] Cadmium (Cd), a highly toxic heavy metal that readily accumulates in organisms, poses a significant threat to potato growth, quality, safety, and human health. In cadmium-contaminated soil, potato plants can absorb cadmium ions through their roots and transport them to edible parts such as tubers for accumulation. Long-term consumption of potatoes with excessive cadmium levels can damage the kidneys and bones, and increase the risk of cancer. Furthermore, cadmium stress can inhibit potato photosynthesis, reducing biomass and yield. Therefore, developing potato varieties with low cadmium adsorption is one of the most economical and effective strategies to ensure the quality and safety of agricultural products and reduce the risk of dietary cadmium exposure at the source.

[0003] Currently, low-cadmium-adsorption potato cultivation systems mainly rely on two technical pathways: variety selection and gene regulation. Regarding variety selection, existing research often employs pot experiments combined with laboratory testing. For example, graphite furnace atomic absorption spectrophotometry was used to determine the cadmium content in different organs of 20 potato varieties, screening for low-accumulation varieties such as D47 and D157. However, this process requires digestion and testing of a large number of samples, making it cumbersome and time-consuming. In terms of detection equipment, although an online cadmium separation and enrichment device has been developed, capturing cadmium through electrothermal evaporation-dielectric barrier discharge to achieve on-site detection within half an hour to an hour, this device still needs to be coupled with an atomic spectrometer. Furthermore, to construct a cadmium adsorption capacity prediction model, massive amounts of sample data need to be accumulated for training. In the field of gene regulation, research has found that overexpression of the StWRKY6 gene can reduce cadmium accumulation in the edible parts of potatoes, but its cultivation effect still needs to be verified in conjunction with cadmium adsorption capacity testing. This makes the traditional techniques, where genotype analysis, environmental control, and detection evaluation are independent processes, unable to achieve data sharing and automated regulation. For example, greenhouse environmental parameters need to be set manually, and test data needs to be entered and analyzed manually, which is difficult to adapt to the needs of large-scale breeding. Therefore, current technology urgently needs a cadmium-low adsorption potato breeding system that can integrate genotype-phenotype prediction, environmental control, growth monitoring, and cadmium adsorption evaluation into a fully automated and remotely managed process. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a cadmium-low adsorption potato breeding system. By employing a basic model combined with online iterative optimization, it achieves rapid prediction of potato cadmium adsorption potential and fully automated and remote management of the low-cadmium, high-yield target, significantly improving breeding efficiency and accuracy.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows: a cadmium-low adsorption potato breeding system, comprising a genotype-phenotype prediction module, a breeding and screening decision module, an automated breeding execution module, and a cadmium adsorption evaluation module: The genotype-phenotype prediction module receives genotype data of potato germplasm resources from external databases or user input, processes it using a basic genotype-phenotype association model based on machine learning, and predicts and outputs the cadmium adsorption potential rating data of potato germplasm resources to the breeding and screening decision module. The breeding and screening decision module receives cadmium adsorption potential rating data, uses a multi-objective optimization algorithm to formulate and output breeding scheme data containing preferred germplasm identifiers and breeding parameters to the automated breeding execution module; The automated cultivation execution module is used to receive cultivation plan data, drive the controllable environment cultivation device according to the parameters of the cultivation plan data, execute automated cultivation operations, and collect and output cultivation environment data and plant growth data to the cadmium adsorption evaluation module in real time. The cadmium adsorption evaluation module is used to receive cultivation environment data and plant growth data, as well as cadmium content detection data of plant tissues from external detection devices. It processes the data using a preset comprehensive evaluation model of adsorption capacity, calculates and outputs the cadmium adsorption capacity evaluation results to the breeding and screening decision module. The breeding and screening decision module is also used to receive cadmium adsorption capacity evaluation results data, and to iteratively optimize the basic genotype-phenotype association model and / or multi-objective optimization algorithm based on the cadmium adsorption capacity evaluation results data to form a closed-loop feedback.

[0006] Furthermore, the genotype-phenotype prediction module includes: The data preprocessing unit receives genotype data of potato germplasm resources from external databases or user input, and outputs standardized genotype feature data to the model prediction unit by using data cleaning, normalization and feature dimensionality reduction processing methods. The model prediction unit receives genotype characteristic data, calls the pre-trained basic genotype-phenotype association model for processing, and outputs the cadmium adsorption potential rating data of potato germplasm resources to the breeding and screening decision module.

[0007] Furthermore, the genotype-phenotype prediction module also includes: The model optimization unit receives the cadmium adsorption capacity evaluation results and the corresponding genotype data, updates the basic genotype-phenotype association model using an online learning algorithm, and outputs the optimized model parameters to the model prediction unit.

[0008] Furthermore, the breeding and selection decision-making module includes: The germplasm selection unit receives cadmium adsorption potential rating data, selects candidate germplasm sets from the germplasm resource bank according to preset potential thresholds, and outputs the set to the scheme generation unit. The scheme generation unit receives the candidate germplasm set, uses a multi-objective optimization algorithm, with the goal of low cadmium adsorption and high yield, formulates and outputs the initial breeding scheme data to the scheme adjustment unit for the candidate germplasm set. The scheme adjustment unit is used to receive the initial breeding scheme data, integrate historical data from the cadmium adsorption capacity evaluation results, combine the real-time performance of candidate germplasm during the breeding process, calibrate and optimize the scheme, and output the final breeding scheme data to the automated breeding execution module. The dynamic screening unit is used to sort and re-screen candidate germplasms based on the evaluation results of cadmium adsorption capacity and growth data after each breeding and evaluation cycle, eliminate substandard germplasms, and feed back the identifier of the preferred germplasm to the germplasm initial selection unit to start the next breeding cycle.

[0009] Furthermore, the automated cultivation execution module includes: The environmental control unit is used to receive cultivation program data and aggregate it into a set of environmental parameters, generate control commands, and output them to the temperature, humidity, light, nutrient solution composition and concentration control devices in the controllable environment cultivation device. The growth monitoring unit receives raw environmental and image data from sensors installed in the controlled environment cultivation device. It uses image recognition and data fusion processing to output cultivation environment data and plant growth data to the cadmium adsorption evaluation module.

[0010] Furthermore, the growth monitoring unit includes: The image processing subunit is used to receive the original image data of the plant acquired by the multispectral camera, and use a deep learning image segmentation algorithm to extract and output growth index data including plant height, leaf area index and relative chlorophyll content estimated based on spectral reflectance to the data fusion subunit. The data fusion subunit is used to receive growth index data such as plant height, leaf area index and relative chlorophyll content, as well as real-time environmental data collected by sensors in the controlled environment cultivation device, perform spatiotemporal alignment and fusion processing, and output time-aligned cultivation environment data and plant growth data to the cadmium adsorption evaluation module.

[0011] Furthermore, the cadmium adsorption evaluation module includes: The data receiving unit is used to receive cultivation environment data and plant growth data, as well as plant tissue cadmium content detection data provided by external detection devices, perform verification and formatting, and output standardized cadmium content data to the comprehensive evaluation unit. The comprehensive evaluation unit receives standardized cadmium content data and plant growth data, calls the dynamically weighted adsorption capacity evaluation model, and outputs the cadmium adsorption capacity evaluation results to the breeding and screening decision module.

[0012] Furthermore, the adsorption capacity evaluation model is as follows: Adsorption capacity score = (1 - enrichment coefficient) × α + (1 - transport coefficient) × β + biomass coefficient × γ; α, β, and γ are weight coefficients, and α+β+γ=1. The weights can be dynamically adjusted according to the iterative evaluation results.

[0013] Furthermore, the system also includes: The user interaction and visualization module is used to receive user query and command data, as well as process and result data output by various modules of the system, perform visualization rendering, and output graphical interface data to the display device.

[0014] Furthermore, the system is deployed on a cloud platform. The automated cultivation execution module transmits data with remote cultivation devices and sensing equipment via IoT protocols, receives control commands, and sends back collected data.

[0015] The above approach has the following beneficial effects: 1. This solution effectively overcomes the problems of traditional potato cadmium adsorption performance testing, such as reliance on large amounts of prior data training, delayed model updates, and long breeding cycles, by constructing a closed-loop control mechanism of "prediction-decision-execution-evaluation-feedback". The system can quickly output prediction results based on only an initial basic model, and collects data in real time and dynamically optimizes the model during variety selection and identification, significantly improving the efficiency and accuracy of breeding low-cadmium adsorption varieties.

[0016] 2. This scheme achieves intelligent and automated control of the breeding process, and can simultaneously consider the breeding objectives of low cadmium adsorption and high yield based on a multi-objective optimization algorithm, thereby improving the synergy and practicality of the breeding scheme. By introducing a dynamically weighted adsorption capacity evaluation model, the system can flexibly adjust the weights of evaluation indicators, enhancing its adaptability to different cultivation conditions and breeding needs.

[0017] 3. This solution also supports cloud platform deployment and remote interaction, facilitating multi-terminal access and collaborative operation, greatly enhancing the system's scalability and practical value. Ultimately, this system not only achieves efficient and low-cost cadmium-low adsorption breeding in potatoes but also provides a referable technical framework for heavy metal resistance breeding in other crops.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the system framework of an embodiment of the cadmium-low adsorption potato cultivation system of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0022] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0023] The following detailed description illustrates the specific implementation method: Example

[0024] In existing research on potato variety breeding and cadmium pollution control, traditional methods often rely on pre-training complex predictive models with a large amount of historical data before assessing and screening for cadmium adsorption capacity. This approach is not only costly in terms of data acquisition and time-consuming in model training, but also struggles to adapt to the dynamic changes in potato genotype-phenotype relationships under different regions and cultivation conditions. This results in low detection efficiency, delayed response, and an inability to achieve rapid and accurate breeding of low-cadmium-adsorption varieties.

[0025] Based on the aforementioned problems, the inventors proposed a closed-loop control mechanism of "prediction-decision-execution-evaluation-feedback" to achieve dynamic evaluation and iterative optimization of the cadmium adsorption capacity of potatoes. This system can not only quickly output prediction results based on an initial basic model, but also continuously collect data and optimize the model during actual breeding, significantly improving the efficiency and accuracy of breeding low-cadmium adsorption varieties.

[0026] An agricultural research institute aims to cultivate a low-cadmium-adsorption potato variety suitable for planting in cadmium-contaminated soil. Traditional methods require collecting genotype and cadmium content data from thousands of samples, training a machine learning model, and then predicting the new germplasm. This process takes several months, and the prediction results are greatly affected by the quality of the initial data. Therefore, a low-cadmium-adsorption potato cultivation system is proposed, as shown in the attached figure. Figure 1 As shown, it includes a genotype-phenotype prediction module, a breeding and screening decision-making module, an automated breeding execution module, and a cadmium adsorption evaluation module: The genotype-phenotype prediction module receives genotype data from potato germplasm resources from external databases or user input, processes it using a basic genotype-phenotype association model based on machine learning, and predicts and outputs cadmium adsorption potential rating data of potato germplasm resources to the breeding and screening decision module. The genotype-phenotype prediction module includes: The data preprocessing unit receives genotype data from potato germplasm resources from external databases or user input. It employs data cleaning, normalization, and feature dimensionality reduction to output standardized genotype feature data to the model prediction unit. The specific processing steps are as follows: S11, Data Cleaning: Missing values ​​in the original genotype data are filled using mean imputation, that is, the mean genotype frequency of all samples at the SNP locus is used for replacement; outliers exceeding ±3 standard deviations are considered as detection noise and are removed.

[0027] Normalization: To eliminate the influence of different dimensions, a minimum-maximum normalization method is used to linearly map the data of each SNP locus to the [0, 1] interval. The calculation formula is as follows:

[0028] S12, where X is the original value of the data for each SNP site, and Xmin and Xmax are the minimum and maximum values ​​among all samples of that site, respectively; S13, Feature Dimensionality Reduction: To reduce data redundancy and highlight key features, Principal Component Analysis (PCA) is used. Specifically, the top N principal components with a cumulative variance contribution rate ≥ 95% are retained, thereby forming standardized genotype feature data with uniform dimensions and standard numerical ranges. The model prediction unit receives genotype feature data, processes it using a pre-trained basic genotype-phenotype association model, and outputs cadmium adsorption potential rating data of potato germplasm resources to the breeding and screening decision module. Specifically, the basic genotype-phenotype association model is a random forest classifier, and its pre-training process is as follows: It uses historically accumulated genotype feature data containing at least 500 potato germplasm resources and their corresponding cadmium adsorption phenotypic labels (divided into three levels: "high," "medium," and "low") as the training set. During training, the core parameter of the random forest is set as follows: the number of decision trees. Maximum depth of decision tree Minimum number of samples required for a leaf node After evaluating and optimizing the model performance through 10-fold cross-validation, the model parameters are saved for use by the prediction unit. This unit inputs the received feature data into the model and outputs cadmium adsorption potential rating data.

[0029] The model optimization unit receives the cadmium adsorption capacity evaluation results and corresponding genotype data from the cadmium adsorption evaluation module. It then uses an online learning algorithm to update the basic genotype-phenotype association model and outputs the optimized model parameters to the model prediction unit. Specifically, an incremental random forest algorithm is used to achieve online model updates: after each breeding and evaluation cycle, newly generated "genotype feature data - cadmium adsorption score" paired data is used as an incremental batch. The system calls the incremental data from the basic genotype-phenotype association model database and uses this incremental data to locally retrain some decision trees in the original random forest model, updating the splitting rules of the tree nodes. This achieves iterative optimization of the model parameters without retraining the entire model. The optimized model parameters are output to the model prediction unit in real time for the next round of prediction. First, researchers input the genotype data of 300 existing potato germplasm resources into the system. The genotype-phenotype prediction module is activated, and its data preprocessing unit cleans, normalizes, and performs principal component analysis (PCA) dimensionality reduction on the original genotype data, extracting the most representative feature data. Subsequently, the model prediction unit calls the pre-trained random forest model (basic genotype-phenotype association model) to output the cadmium adsorption potential rating of each germplasm (divided into three levels: high, medium, and low).

[0030] The breeding and screening decision module receives cadmium adsorption potential rating data, uses a multi-objective optimization algorithm to formulate and output breeding scheme data containing preferred germplasm identifiers and breeding parameters to the automated breeding execution module; the breeding and screening decision module also receives cadmium adsorption capacity evaluation results data, and iteratively optimizes the basic genotype-phenotype association model and / or multi-objective optimization algorithm based on the cadmium adsorption capacity evaluation results data to form a closed-loop feedback. The breeding and selection decision module includes: The germplasm selection unit receives cadmium adsorption potential rating data, selects candidate germplasm sets from the germplasm resource bank according to preset potential thresholds, and outputs the set to the scheme generation unit. The scheme generation unit receives the candidate germplasm set and, using a multi-objective optimization algorithm, aims at low cadmium adsorption and high yield to formulate and output initial breeding scheme data to the scheme adjustment unit. The specific steps of the multi-objective optimization algorithm are as follows: S21, Encoding and Population Initialization: S211, Input: Identifier of each germplasm in the candidate germplasm set and its corresponding cadmium adsorption potential rating data.

[0031] S212, coding method: real number coding is used, each individual (chromosome) represents a breeding program, which consists of a set of breeding parameters, specifically including: cadmium ion concentration in nutrient solution (unit: μM, range: 0.1-5.0), photoperiod (unit: hours / day, range: 12-16), and diurnal temperature range (unit: ℃, range: 8-12).

[0032] S213, Population initialization: Randomly generate an initial population within the allowable range of the parameters, and set the population size to 50.

[0033] S22, Fitness Assessment: S221 defines two objective functions to accommodate the dual-objective optimization of low cadmium adsorption and high yield: Objective function 1 (minimization): f1 = normalized cadmium adsorption potential score. This score comes directly from the output of the genotype-phenotype prediction module and has been mapped to the [0,1] interval. The higher the value, the greater the cadmium adsorption potential, so it needs to be minimized.

[0034] Objective function 2 (maximization): f2 = Yield potential index based on historical data. This index is obtained by querying the historical database, based on the germplasm identifier, to obtain the average tuber biomass of the germplasm or its parents under similar cultivation conditions, and normalized to the [0,1] interval.

[0035] S222: The algorithm calculates the value of each individual on two objective functions, performs fast non-dominated sorting, and determines the Pareto level of the individual.

[0036] S23, Selection, Crossover, and Mutation: S231, Selection: A binary tournament selection method is adopted, based on the individual's non-dominance level and crowding level.

[0037] S232, Crossover: Simulated binary crossover is used, and the crossover probability is set to 0.8.

[0038] S233, Mutation: Use polynomial mutation, with the mutation probability set to 0.1.

[0039] S234, Iteration: Set the number of generations to 100.

[0040] S24, Output: S241. After the algorithm finishes running, select the Pareto front with the highest non-dominant level from the final generation population.

[0041] S242, the system by default selects an individual from the frontier that performs best on the objective function f2, decodes it into a specific combination of breeding parameters, and outputs it as the initial breeding scheme data; The program adjustment unit receives initial breeding program data, integrates historical data from cadmium adsorption capacity evaluation results, and combines this with the real-time performance of candidate germplasm during the breeding process to calibrate and optimize the program. It then outputs the final breeding program data to the automated breeding execution module. The steps for program calibration and optimization are as follows: S31, Data Preparation: S311, Real-time performance data: Continuously receives the daily average growth rate of leaf area index uploaded by the growth monitoring unit.

[0042] S312, Historical Data: Retrieves the final cadmium adsorption capacity score of the current candidate germplasm set (or its parents) under different breeding parameters in past breeding cycles from the system database.

[0043] S32, calibration trigger mechanism: S321, the system sets a threshold for the daily average growth rate of leaf area index (e.g., 0.08).

[0044] S322, If the system detects that the growth rate of a candidate germplasm is lower than this threshold for three consecutive days, it will automatically trigger the calibration process of the breeding program for that germplasm.

[0045] S33, Calibration execution logic: S331, the system establishes a simple linear regression model: ;in, This represents the baseline prediction score when all breeding parameters are 0; This indicates the weight of the influence of cadmium concentration in the nutrient solution on the predicted score; cadmium concentration in the nutrient solution is an input variable, in μm, and is taken from the current breeding program or historical data. This represents the weight of the influence of light duration on the predicted score; light duration is an input variable, measured in hours per day, and is taken from the current breeding program or historical data. This linear regression model is trained using historical data obtained from S312.

[0046] S332, When calibration is triggered, the system applies this model to the current scheme: If the current real-time growth performance is significantly lower than expected, and the model indicates that reducing the cadmium concentration in the nutrient solution (e.g., reducing it by 0.5 μM) has an acceptable negative impact on the cadmium adsorption score (e.g., a predicted decrease of no more than 5%), then this adjustment should be implemented to prioritize growth.

[0047] If historical models indicate that a slight increase in light duration (e.g., 0.5 hours) under current parameters may simultaneously benefit growth and reduce cadmium adsorption, then this adjustment should be implemented.

[0048] S333, the calibration process is a fine-tuning based on rules and model predictions, and the adjustment of any parameter in a single step does not exceed 10% of the initial value.

[0049] S34, Output: Integrate the breeding parameters after the above logic calibration into the final breeding scheme data and output it; The dynamic screening unit is used to sort and re-screen candidate germplasms based on the evaluation results of cadmium adsorption capacity and growth data after each breeding and evaluation cycle, eliminate substandard germplasms, and feed back the identifier of the preferred germplasm to the germplasm initial selection unit to start the next breeding cycle.

[0050] The rating data is fed into the breeding and screening decision module. This module first uses a germplasm selection unit to screen candidate germplasm sets (e.g., 10 selected germplasm numbers) from all germplasm based on preset potential thresholds. Then, a scheme generation unit uses a multi-objective genetic algorithm, with "low cadmium adsorption" and "high yield" as optimization objectives, to formulate an initial breeding scheme for this candidate germplasm set, including corresponding cultivation parameters (such as nutrient solution Cd²⁺ concentration, pH value, and photoperiod). Further fine-tuning by the scheme adjustment unit, combined with historical evaluation results, forms the final breeding scheme, which is then output to the automated cultivation execution module. After each cultivation and evaluation cycle, the dynamic screening unit sorts and re-screens the candidate germplasm based on cadmium adsorption capacity evaluation results and growth data, eliminating inferior germplasm and feeding back the selected germplasm identifiers to the system to initiate the next progressive breeding cycle.

[0051] The automated cultivation execution module receives cultivation plan data, drives the controlled environment cultivation device according to the parameters of the cultivation plan data, executes automated cultivation operations, and collects and outputs cultivation environment data and plant growth data to the cadmium adsorption evaluation module in real time. The automated cultivation execution module includes: The environmental control unit is used to receive data from the cultivation program and aggregate it into a set of environmental parameters, generate control commands, and output them to the temperature, humidity, light, nutrient solution composition and concentration control devices in the controllable environment cultivation device.

[0052] Specifically, communication is established with a remote cultivation device via a signal connection, and instructions are generated and sent to the cultivation device: Temperature control: Set the target temperature as Ttarget (e.g., 22℃). The unit reads the real-time data Tcurrent from the temperature sensor, uses a PID control algorithm (proportional coefficient Kp=2.0, integral time Ti=300 seconds, derivative time Td=50 seconds) to calculate the control quantity, and adjusts the power of the heater or refrigerator through the PWM signal to maintain the temperature and humidity within ±2℃ of Ttarget.

[0053] Nutrient solution regulation: Based on the target Cd²⁺ concentration Ctarget in the plan, the unit controls a peristaltic pump to inject Cd²⁺-containing nutrient solution into the mother liquor tank at a specific speed. Simultaneously, a pH sensor monitors the pH value of the nutrient solution in real time. If the pH value exceeds the set range (e.g., 5.5-6.5), dilute acid or dilute alkali solution is automatically added for adjustment.

[0054] Light control: According to the light cycle set in the scheme, the unit controls the power supply of the LED supplementary light group through a relay switch to achieve precise timed switching, and the light intensity is steplessly controlled by adjusting the LED drive current.

[0055] The growth monitoring unit receives raw environmental and image data from sensors installed within the controlled environment cultivation device. It then uses image recognition and data fusion processing to output cultivation environment data and plant growth data to the cadmium adsorption evaluation module. The growth monitoring unit includes: The image processing subunit receives raw plant image data captured by a camera and uses a deep learning image segmentation algorithm to extract and output plant height and leaf area index growth index data to the data fusion subunit. Specifically, the image processing subunit receives raw plant image data periodically (e.g., daily at 10:00 AM) from a multispectral camera deployed above the cultivation area. Image segmentation is performed using a deep learning model based on the U-Net architecture. This model is pre-trained using 1000 potato plant images labeled with the LabelMe tool, with an input image size of 256x256 pixels and an output pixel-level plant segmentation mask. During image feature extraction: Plant height: Based on the segmented plant mask, calculate the vertical distance between the highest pixel and the pixel on the substrate surface, and convert it into the actual physical height (cm) by combining the camera calibration parameters.

[0056] Leaf area index: Calculates the total pixel area of ​​the plant mask and converts it according to the pre-calibrated relationship model between pixel area and actual leaf area.

[0057] The extracted plant height, leaf area index, and other growth index data are packaged and output to the data fusion subunit.

[0058] The data fusion subunit receives plant height and leaf area index growth index data, as well as real-time environmental data collected by sensors within the controlled environment cultivation device. It performs spatiotemporal alignment and fusion processing, outputting time-aligned cultivation environment data and plant growth data to the cadmium adsorption evaluation module. This is achieved by receiving growth index data from the image processing subunit and real-time environmental data (temperature, humidity, light intensity) from environmental sensors. A Kalman filter algorithm is used to perform spatiotemporal alignment and fusion of the multi-source data. The system maintains a state vector [plant height, leaf area index, temperature, humidity] and performs optimal estimations of image data (low-frequency but information-rich) and environmental sensor data (high-frequency but singular) every minute to eliminate instantaneous noise. The system then outputs time-aligned, consistent cultivation environment data and plant growth data in JSON format to the cadmium adsorption evaluation module.

[0059] After receiving the plan, the automated cultivation execution module drives the controllable environmental cultivation device (such as a smart greenhouse) to adjust temperature, humidity, light intensity, and nutrient solution composition and concentration. Simultaneously, the growth monitoring unit collects plant images and environmental data in real time using deployed multispectral cameras and environmental sensors. The image processing subunit uses a U-Net deep learning model to segment plant images and extract growth indicators such as plant height and leaf area index. The data fusion subunit performs spatiotemporal alignment and fusion of image data and environmental sensor data to form a time-consistent cultivation dataset, which is then transmitted to the cadmium adsorption evaluation module.

[0060] The cadmium adsorption evaluation module receives cultivation environment data and plant growth data, as well as cadmium content detection data from plant tissues obtained from external detection devices. It processes this data using a pre-set comprehensive evaluation model of adsorption capacity, calculates and outputs the cadmium adsorption capacity evaluation results to the breeding and screening decision module. The cadmium adsorption evaluation module includes: The data receiving unit receives cultivation environment data and plant growth data, as well as cadmium content detection data of plant tissues provided by external detection devices. It then verifies and formats the data, outputting standardized cadmium content data to the comprehensive evaluation unit. Specifically, the external detection device (such as an atomic absorption spectrometer) sends the detection data to the data receiving unit via an RS-232 serial port. The data receiving unit performs CRC32 verification on the received data to ensure the integrity of data transmission. Based on the standard curve used during detection, the cadmium content data is inversely calculated and standardized to mg / kg (dry weight), outputting standardized cadmium content data to the comprehensive evaluation unit.

[0061] The comprehensive evaluation unit receives standardized cadmium content data and plant growth data, calls the dynamically weighted adsorption capacity evaluation model, and outputs cadmium adsorption capacity evaluation results. The formula for the adsorption capacity evaluation model is: Adsorption capacity score = (1 - enrichment coefficient) × α + (1 - transport coefficient) × β + biomass coefficient × γ The enrichment coefficient is calculated as: tuber cadmium content / soil cadmium content; the soil cadmium content is a known value set during the initial cultivation stage.

[0062] Translocation coefficient = cadmium content in tubers / cadmium content in roots; cadmium content in roots and cadmium content in tubers were measured simultaneously.

[0063] Biomass coefficient = Fresh weight of tuber per plant / Maximum fresh weight of tuber per plant of all candidate germplasms in this round of breeding.

[0064] Where α, β, and γ are weighting coefficients, and α + β + γ = 1.

[0065] The weights can be dynamically adjusted based on the iterative evaluation results, and the cadmium adsorption capacity evaluation data is output to the breeding and screening decision module. The dynamic adjustment method of the weights is as follows: after each round of evaluation, the unit records the comprehensive score under the current weight and compares it with an "ideal target" based on actual cadmium content and yield (for example, setting the ideal state as cadmium content in tubers <= 0.2 mg / kg and biomass >= 200 g / plant). Fine-tuning is performed using the gradient descent method: the error between the comprehensive score and the ideal state is calculated, and then α, β, and γ are adjusted simultaneously in the direction of decreasing error with a learning rate η = 0.05. After adjustment, the weights are normalized to ensure that their sum is always 1. For example, if a germplasm with high biomass but also high cadmium adsorption is found to have an excessively high score in a certain iteration, the system will automatically increase the weights of α and β and decrease the weight of γ in the next round, thus making the evaluation criteria more inclined towards "low cadmium adsorption".

[0066] Once the potatoes reached the tuber formation stage, researchers collected samples for testing to obtain cadmium content data from the plant tissues, which was then input into the system. The data receiving unit of the cadmium adsorption evaluation module verified and standardized the test data, while the comprehensive evaluation unit invoked a dynamically weighted adsorption capacity evaluation model. In this model, α, β, and γ were weighting coefficients, initially set to 0.4, 0.3, and 0.3, respectively, and dynamically adjusted based on previous evaluation results. The evaluation results for this batch showed that three germplasms had significantly better cadmium adsorption scores than the others.

[0067] The evaluation results are fed back to the breeding and selection decision module, triggering the model optimization unit to update the basic genotype-phenotype association model using an online learning algorithm (such as incremental random forest). Simultaneously, the multi-objective optimization algorithm adjusts the weight settings based on the latest evaluation results, generating a more accurate breeding plan for the next round.

[0068] The system also includes: The user interaction and visualization module is used to receive user queries and command data, as well as process and result data output by various modules of the system, perform visualization rendering, and output graphical interface data to the display device. The system is deployed on a cloud platform. The automated cultivation execution module transmits data with remote cultivation devices and sensing devices through the Internet of Things protocol, receives control commands, and sends back collected data.

[0069] The user interaction and visualization module provides a graphical interface throughout, allowing researchers to query data from each module in real time, adjust system parameters, and trace and analyze the breeding process.

[0070] Compared to traditional methods, this system rapidly screens low-cadmium potential germplasm in the first round of breeding and continuously optimizes the model and scheme in subsequent iterations, significantly shortening the breeding cycle and reducing experimental costs. The system achieves model self-evolution through a closed-loop feedback mechanism, avoiding excessive dependence on the initial amount of data, making it particularly suitable for practical breeding scenarios with limited resources and variable environments.

[0071] In addition, the system is deployed on a cloud platform, supports remote access and control from multiple terminals, and achieves data interconnection with remote cultivation equipment through IoT protocols, thus possessing good scalability and practicality.

[0072] I. Experimental Design Background and Objectives Traditional cadmium-low adsorption potato breeding techniques suffer from problems such as independent genotype analysis, environmental control, and testing and evaluation processes, lack of data interoperability, and absence of automated regulation, leading to low breeding efficiency. This experiment compares the system of this invention with traditional methods to verify the advantages of the novel breeding system in terms of end-to-end automation, data interoperability, and breeding efficiency.

[0073] II. Experimental Grouping and Material Preparation Experimental group: The cadmium-low adsorption potato breeding system of the present invention was adopted, including a genotype-phenotype prediction module, a breeding and screening decision module, an automated breeding execution module, and a cadmium adsorption evaluation module. The system was deployed on a cloud platform and connected to a remote intelligent greenhouse.

[0074] Control group: The traditional pot experiment combined with laboratory testing methods was used, including artificial genotype screening, manual setting of greenhouse environmental parameters (temperature, humidity, light, etc.), manual collection of growth data and cadmium content detection.

[0075] Experimental materials: 300 potato germplasm resources (genotype data known) were selected and divided into two groups of 150 each. Each group included known low cadmium adsorption varieties (such as D47 and D157) and common varieties.

[0076] III. Experimental Procedure (I) Experimental Group Procedure Genotype-phenotype prediction The data preprocessing unit cleaned, normalized, and performed PCA dimensionality reduction on 150 germplasm genotype data.

[0077] The model prediction unit calls a pre-trained random forest model to output a cadmium adsorption potential rating (high / medium / low), and selects the top 30 potential germplasm samples.

[0078] Cultivation Program Development The breeding and screening decision module uses a multi-objective genetic algorithm to generate an initial breeding plan (including parameters such as cadmium concentration in nutrient solution, pH value, and photoperiod) with the goals of "low cadmium adsorption" and "high yield".

[0079] The scheme adjustment unit optimizes the scheme by combining historical evaluation data and outputs it to the automated cultivation execution module.

[0080] Automated cultivation and monitoring The environmental control unit drives the intelligent greenhouse to automatically adjust the temperature and humidity (22±2℃, humidity 60±5%), light intensity (16h photocycle, 300μmol / m²・s) and nutrient solution composition.

[0081] The growth monitoring unit collects plant images and environmental data in real time through multispectral cameras and sensors. The image processing subunit extracts indicators such as plant height and leaf area index, and the data fusion subunit generates a time-series dataset.

[0082] Cadmium adsorption evaluation and model iteration Samples were taken during the tuber formation period, and the cadmium content of the plant tissue was measured by an external detection device. The cadmium adsorption evaluation module calculated the adsorption capacity score through a dynamic weighted model (formula: adsorption capacity score = (1 - enrichment coefficient) × 0.4 + (1 - translocation coefficient) × 0.3 + biomass coefficient × 0.3).

[0083] The breeding and screening decision module iteratively optimizes the genotype-phenotype association model and multi-objective algorithm based on the evaluation results to generate the next round of plans.

[0084] (II) Control Group Procedure Genotype screening: Manually compare literature data to screen for possible low-cadmium varieties from 150 germplasm accessions, which takes 3 days.

[0085] Environmental control: Greenhouse parameters were manually set (same as the experimental group), and temperature and humidity were manually recorded daily with an error of ±3℃ / ±10%.

[0086] Growth data collection: Plant height and leaf area are measured manually every 7 days, and the data is recorded manually, which takes about 2 hours per instance.

[0087] Cadmium content detection: Samples were taken during the tuber formation period, and after manual digestion, the cadmium content was detected by graphite furnace atomic absorption spectrophotometry. The detection time for each sample was 4 hours, and the data was manually entered and analyzed.

[0088] IV. Detection Target Data Breeding cycle: The total time from genotype screening to obtaining low-cadmium varieties.

[0089] Labor costs: The total manual operation time (in hours) required throughout the experiment.

[0090] Data interoperability: The automated data transmission rate (%) of each link is 0% for traditional methods and 100% for this system.

[0091] Cadmium adsorption capacity: cadmium content in tubers (mg / kg), enrichment coefficient, translocation coefficient, and biomass (g / plant).

[0092] Model iteration efficiency: The improvement in model prediction accuracy after each round of training.

[0093] V. Experimental Results Table Table 1 Comparison of experimental results between the experimental group and the control group

[0094] VI. Discussion of Results Experimental data show that the cadmium-low adsorption potato cultivation system of the present invention is significantly superior to traditional methods in many aspects: 1. Significantly Improved Breeding Efficiency: Through systematic iterative optimization, the experimental group completed three rounds of screening and evaluation within 180 days, while traditional methods require 210 days for a single round of screening. Although the total cycle is similar, the system of this invention achieves multiple rounds of optimization within the same timeframe, significantly improving the density and efficiency of the breeding process. In terms of labor costs, the experimental group required only 120 hours, a 42.9% reduction compared to the control group's 210 hours, demonstrating the system's significant advantages in automated management and data integration.

[0095] 2. End-to-end data interoperability and automation: The system achieves automated data transmission and processing throughout the entire process, from genotype prediction, decision-making, environmental control to cadmium adsorption evaluation, with a transmission rate of 100%. In contrast, traditional methods rely on manual recording and input, which is not only inefficient but also carries the risk of data inconsistency and human error.

[0096] 3. Good balance between cadmium adsorption control and yield: The cadmium content in the tubers of the experimental group was 0.15 mg / kg, significantly lower than the 0.25 mg / kg in the control group, a reduction of 40%. The enrichment coefficient and translocation coefficient decreased by 36.2% and 37.5%, respectively, indicating that the system has a significant effect in inhibiting cadmium absorption and translocation. At the same time, the biomass of the experimental group increased by 13.6%, indicating that the system achieved the "low cadmium" target without sacrificing yield, demonstrating the effectiveness and practicality of the multi-objective optimization algorithm.

[0097] 4. The model possesses continuous optimization capabilities: Through an online learning mechanism, the system improved the model's prediction accuracy from 70% to 82% in three iterations, an increase of 17%. In contrast, traditional methods rely on static models and cannot be dynamically adjusted based on actual breeding data, with a fixed prediction accuracy of 65%. This demonstrates that the system of this invention possesses excellent adaptability and evolutionary capabilities, making it particularly suitable for complex and ever-changing breeding environments.

[0098] In summary, the system of this invention successfully solves the problems of independent process and lack of automation in traditional technologies, providing an efficient and precise technical solution for cadmium-low adsorption potato breeding, and also providing a framework that can be referenced for heavy metal resistance breeding of other crops.

[0099] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A cadmium-low adsorption potato cultivation system, characterized in that, The system comprises a genotype-phenotype prediction module, a breeding and screening decision module, an automated cultivation execution module, and a cadmium adsorption evaluation module. The genotype-phenotype prediction module is configured to receive genotype data of potato germplasm resources from an external database or user input, process the data using a machine learning-based basic genotype-phenotype correlation model, predict and output cadmium adsorption potential rating data of the potato germplasm resources to the breeding and screening decision module. The breeding and screening decision module is configured to receive the cadmium adsorption potential rating data, develop and output breeding program data including preferred germplasm identification and cultivation parameters to the automated cultivation execution module using a multi-objective optimization algorithm. The automated cultivation execution module is configured to receive the cultivation program data, drive a controllable environment cultivation device according to the parameters of the cultivation program data, perform automated cultivation operations, and collect and output cultivation environment data and plant growth data to the cadmium adsorption evaluation module in real time. The cadmium adsorption evaluation module is configured to receive the cultivation environment data and plant growth data, receive plant tissue cadmium content detection data from an external detection device, process the data using a preset adsorption capacity comprehensive evaluation model, calculate and output cadmium adsorption capacity evaluation result data to the breeding and screening decision module. The breeding and screening decision module is further configured to receive the cadmium adsorption capacity evaluation result data, and iteratively optimize the basic genotype-phenotype correlation model and / or the multi-objective optimization algorithm based on the cadmium adsorption capacity evaluation result data, forming a closed-loop feedback.

2. The cadmium low-adsorbing potato breeding system according to claim 1, characterized by The genotype-phenotype prediction module comprises: A data preprocessing unit configured to receive genotype data of potato germplasm resources from an external database or user input, and output standardized genotype feature data to the model prediction unit using data cleaning, normalization, and feature dimension reduction processing. A model prediction unit configured to receive the genotype feature data, call a pre-trained basic genotype-phenotype correlation model for processing, and output cadmium adsorption potential rating data of the potato germplasm resources to the breeding and screening decision module.

3. The cadmium low-adsorbing potato breeding system according to claim 2, characterized by The genotype-phenotype prediction module further comprises: A model optimization unit configured to receive cadmium adsorption capacity evaluation results and corresponding genotype data, update the basic genotype-phenotype correlation model using an online learning algorithm, and output optimized model parameters to the model prediction unit.

4. The cadmium low-adsorbing potato breeding system according to claim 3, characterized by The breeding and screening decision module comprises: A germplasm preliminary selection unit configured to receive the cadmium adsorption potential rating data, select a candidate germplasm set from a germplasm resource library based on a preset potential threshold, and output the set to a scheme generation unit. A scheme generation unit configured to receive the candidate germplasm set, develop and output initial breeding program data for the candidate germplasm set to a scheme adjustment unit using a multi-objective optimization algorithm targeting low cadmium adsorption and high yield. A scheme adjustment unit configured to receive the initial breeding program data, integrate historical data from cadmium adsorption capacity evaluation results, combine real-time performance of the candidate germplasm in the cultivation process, perform scheme calibration and optimization, and output final breeding program data to the automated cultivation execution module. A dynamic screening unit is configured to sort and re-screen the candidate germplasm according to the cadmium adsorption capacity evaluation results and growth data after each breeding and evaluation cycle, eliminate substandard germplasm, and feed back the preferred germplasm to the germplasm primary screening unit to start the next breeding cycle.

5. The cadmium low-adsorbing potato breeding system according to claim 4, characterized by The automatic breeding execution module comprises: An environment control unit is configured to receive breeding scheme data and aggregate the data into a set of environmental parameters, generate control instructions, and output the instructions to temperature and humidity, light, nutrient solution composition and concentration control equipment in the controlled environment cultivation device. A growth monitoring unit is configured to receive raw environmental and image data collected by sensors arranged in the controlled environment cultivation device, use image recognition and data fusion processing, and output cultivation environment data and plant growth data to the cadmium adsorption evaluation module.

6. The cadmium low-adsorbing potato breeding system according to claim 5, characterized by The growth monitoring unit comprises: An image processing subunit is configured to receive raw plant image data collected by a camera, use a deep learning image segmentation algorithm, extract and output plant height and leaf area index growth indicator data to a data fusion subunit; A data fusion subunit is configured to receive plant height and leaf area index growth indicator data, and real-time environmental data collected by sensors in the controlled environment cultivation device, perform spatio-temporal alignment and fusion processing, and output time-series aligned cultivation environment data and plant growth data to the cadmium adsorption evaluation module.

7. The cadmium low-adsorbing potato breeding system according to claim 6, characterized by The cadmium adsorption evaluation module comprises: A data receiving unit is configured to receive cultivation environment data and plant growth data, receive plant tissue cadmium content detection data provided by an external detection device, perform verification and formatting, and output standardized cadmium content data to a comprehensive evaluation unit; A comprehensive evaluation unit is configured to receive standardized cadmium content data and plant growth data, call a dynamic weighted adsorption capacity evaluation model, and output cadmium adsorption capacity evaluation result data to a breeding and screening decision module.

8. The cadmium low-adsorbing potato breeding system according to claim 7, characterized by The adsorption capacity evaluation model is: Adsorption capacity score = (1 - enrichment coefficient) × α + (1 - transport coefficient) × β + biomass coefficient × γ; wherein α, β, and γ are weight coefficients, and α + β + γ = 1, and the weights can be dynamically adjusted according to the iteration evaluation results.

9. The cadmium low-adsorbing potato breeding system according to claim 8, characterized by The system further comprises: A user interaction and visualization module is configured to receive user query and instruction data, receive process data and result data output by each module of the system, perform visualization rendering, and output graphical interface data to a display device.

10. The cadmium low-adsorbing potato breeding system according to claim 9, characterized by The system is deployed on a cloud platform, and the automatic breeding execution module transmits data to remote cultivation devices and sensor equipment through an Internet of Things protocol, receives control instructions, and transmits collected data.

Images