Multi-index prediction management platform for grain quality analysis

Abstract

The invention discloses a multi-index prediction management platform for grain quality analysis, and relates to the technical field of grain quality analysis, the multi-index prediction management platform comprises a grain detection module and a prediction evaluation module, the prediction evaluation module is connected with the grain detection module through signal transmission, and the prediction evaluation module is connected with the grain detection module through signal transmission. And the prediction and evaluation module performs analysis and prediction and evaluation by receiving the information collected by the grain detection module, and transmits a result to the database. By designing the prediction evaluation module, the function of accurately predicting the grain quality change trend is achieved, the problems that a single model is insufficient in prediction accuracy, poor in adaptability and insufficient in information utilization are solved, deviation and limitation existing in the single prediction model can be avoided, multiple factors influencing the grain quality are comprehensively considered, and the prediction accuracy is improved. The accuracy, stability and reliability of prediction are improved, and the utilization of computing resources and time cost is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of grain quality analysis, and in particular to a multi-index prediction management platform for grain quality analysis. Background Art

[0002] With the continuous growth of population, the demand for food continues to rise. How to scientifically manage and utilize existing food resources has become an urgent problem to be solved. Multi-indicator prediction management platforms such as the index food platform have emerged. These platforms integrate a variety of data and analysis tools to optimize food production, circulation and consumption, help managers at all levels make accurate decisions, and use big data technology to collect food production and consumption data from all over the country in real time, comprehensively analyze the supply and demand of food, and provide strong support for the scientific management of the food industry. In addition, with the in-depth research in fields such as molecular biology and genetics, and the rapid development of bioinformatics, big data analysis and artificial intelligence technology, seed analysis systems for specific food crops have also emerged. These systems provide strong support for the selection and breeding of new varieties suitable for specific ecological conditions by deeply analyzing the genomes of food crops and combining the influence of environmental factors. At the same time, they provide more accurate means for multi-indicator prediction of food quality. In the process of grain quality analysis, farmers have too few ways to understand grain quality measurement information, which leads farmers to weigh yield when choosing planting varieties, and lack the understanding of measuring with quality indicators. Predicting grain quality is a complex and multidimensional task, which is affected by many factors. It is difficult to fully consider all relevant factors through a single model, resulting in low confidence in the prediction results.

[0003] Patent document CN118566149B discloses a grain quality monitoring and inspection method and system based on data analysis. The above patent realizes rapid and accurate inspection of grain quality, thereby improving the level of grain quality assurance.

[0004] In summary, the above patent obtains coarse impurity samples, fine impurity samples and clean grain samples by sorting impurities on grain samples to be inspected; moisture content of grain samples to be inspected is measured based on coarse impurity samples, fine impurity samples and clean grain samples to be inspected; protein spectrum analysis and protein content detection are performed on grain samples to be inspected to obtain spectral absorption protein content data of grain samples to be inspected; imperfect evaluation analysis and quality grade detection are performed on grain samples to be inspected to obtain quality grade results of grain samples to be inspected, and there is still room for optimization in the collection of information on grain production process and prediction of grain quality; To this end, this application proposes a multi-indicator prediction management platform for grain quality analysis that can collect complete information on the grain production process and accurately predict the changing trends of grain quality. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-index prediction management platform for grain quality analysis to solve the technical problems of insufficient prediction accuracy of a single model and incomplete information collection proposed in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: a multi-index prediction management platform for grain quality analysis, the multi-index prediction management platform comprising a grain detection module and a prediction evaluation module, the prediction evaluation module being connected to the grain detection module via signal transmission; The prediction and evaluation module receives the information collected by the grain detection module, performs analysis and prediction and evaluation, and transmits the results to the database; The food detection module is responsible for collecting data information on food production, processing, storage and circulation, and transmitting it to the database; The prediction and evaluation module includes: XGBoost algorithm, random forest algorithm, SVM algorithm, BP neural network algorithm, Stacking algorithm and Boosting algorithm; The Bagging algorithm constructs a base learner by extracting a training sample set using the Bootstraping method from the prediction results of the XGBoost algorithm, the random forest algorithm, the SVM algorithm, and the BP neural network algorithm, inputs the base learner generated by the Bagging algorithm and the grain information collected by the grain detection module into the meta-learner for secondary learning, constructs a Stacking integration model, trains and evaluates the Stacking integration model, and uses the trained Stacking integration model to predict and evaluate grain quality.

[0007] Preferably, the grain detection module includes: a grain planting monitoring component, a grain processing detection component, a grain storage monitoring component and a grain circulation detection component; The grain planting monitoring component includes: a seedling monitoring unit, a soil moisture monitoring unit, an insect monitoring unit and a disaster monitoring unit. The grain planting monitoring component is connected to the database through signal transmission; The seedling monitoring unit monitors the growth and development of grain crops during planting, including the height, number of leaves, leaf area, number of tillers, root development, leaf color and leaf density of grain crops; The soil moisture monitoring unit monitors the moisture content, temperature, humidity, pH, and nitrogen, phosphorus, and potassium content of grain-growing soil; The pest monitoring unit monitors the pest species, pest numbers, and occurrence time and location, and uses pest monitoring lights and sex traps to automatically trap pests; The disaster monitoring unit monitors natural disasters by monitoring temperature, humidity, wind speed and rainfall through meteorological sensors.

[0008] Preferably, the prediction and evaluation module is connected to the intelligent decision-making module through signal transmission, and the intelligent decision-making module combines the prediction results of the prediction and evaluation module with market demand, consumer preference and policy orientation to provide managers with decision support on planting structure, irrigation and fertilization scheme, harvesting and storage conditions.

[0009] Preferably, the multi-index prediction management platform further includes a shared management module, which is a touch-controlled terminal for managers to understand the food quality prediction results and control the food quality.

[0010] Preferably, the XGBoost algorithm analyzes the temperature, humidity and chemical content in the granary, constructs a prediction model to predict the quality of grain, and outputs the prediction result.

[0011] Preferably, the random forest algorithm classifies the collected food information, constructs a decision tree, and integrates the output results of the decision tree as the prediction result.

[0012] Preferably, the SVM algorithm predicts changes in grain quality by using parameters in the grain storage process through training an SVM model, and outputs a prediction result, and the BP neural network algorithm predicts grain quality by constructing a prediction model, and outputs a prediction result.

[0013] Preferably, the food processing detection component comprises: a physical detection unit, a chemical detection unit and a microbial detection unit, and the food production detection component is connected to the database through signal transmission; The physical detection unit includes a visual sensor, a moisture meter, a temperature and humidity meter, and a vibration screening machine, and the physical detection unit is connected to the database and the prediction and evaluation module through signal transmission; The visual sensor collects information on the color, shape, size, glossiness and integrity of the grain, the moisture meter detects the moisture content of the grain, the temperature and humidity meter detects the temperature and humidity of the grain, and the vibration screening machine detects the size distribution of grain particles and transmits the information to the database to evaluate the processing and edible quality of the grain. The chemical detection unit includes a food heavy metal detector, a gas chromatograph and a liquid chromatograph, and the chemical detection unit is connected to the database and the prediction and evaluation unit through signal transmission; Grain heavy metal detector detects harmful heavy metals in grain, gas chromatograph detects pesticides with high volatility and high thermal stability, and liquid chromatograph detects pesticides with high polarity and low thermal stability; The microbial detection unit includes a mycotoxin detector and an enzyme marker, and the microbial detection unit is connected with the database and the prediction and evaluation module through signal transmission; The mycotoxin detector detects the mycotoxins in the grain, and the enzyme labeler detects the non-edible chemicals, abused food additives, heavy metals, nutritional enhancers and chemical residues in the grain.

[0014] Preferably, the grain storage monitoring component includes: a temperature and humidity monitoring unit, a moisture monitoring unit, a pest density monitoring unit and a grain quality monitoring unit; The temperature and humidity monitoring unit monitors the temperature and humidity in the granary. The temperature range of the granary is 15℃~25℃, and the humidity range is 55%~65%; The moisture monitoring unit monitors the moisture content of grain stored in the granary; The pest density monitoring unit monitors the pest density in the granary; The grain quality monitoring unit monitors the color, smell, fatty acid esters, taste scores and food safety indicators of grain in the granary.

[0015] Preferably, the grain circulation detection component comprises: a packaging quality detection unit and a simulated transportation environment unit; The packaging quality inspection unit includes a texture analyzer, an intelligent electronic tensile tester, an impact tester, a rubbing resistance tester, a thickness tester, and a friction coefficient tester. The packaging quality inspection unit is connected to the database through signal transmission; The texture analyzer is used to test the hardness, brittleness, stickiness and breaking strength of food packaging materials. The intelligent electronic tensile testing machine is used to test the tensile strength, peel strength and heat-sealing strength of packaging materials. The impact tester is used to test the impact resistance of packaging materials. The rubbing resistance tester is used to test the performance of packaging when rubbing, bending, twisting and squeezing. The thickness tester is used to test the thickness uniformity of packaging materials. The friction coefficient tester is used to evaluate the slippery performance of the inside and outside of packaging materials. The simulated transport environment unit includes a packaged goods vibration tester, a vacuum sealing tester and a temperature and humidity conditioning box, and the simulated transport environment unit is connected to the database through signal transmission; The packaged goods vibration testing machine is used to evaluate the strength of the packaging under sinusoidal vibration and resonance and its ability to protect grain. The vacuum sealing tester is used to test the sealing performance of the packaging. The temperature and humidity conditioning box is used to test the protection performance of the packaging on grain quality under different temperature and humidity conditions.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention realizes the function of accurately predicting the trend of grain quality changes by designing a prediction and evaluation module, solves the problems of insufficient prediction accuracy, poor adaptability and insufficient information utilization of a single model, avoids the deviation and limitations of a single prediction model, comprehensively considers various factors affecting grain quality, improves the accuracy, stability and reliability of prediction, and optimizes the utilization of computing resources and time costs; 2. The present invention realizes the function of collecting complete information on the grain production process by designing a grain detection module, solves the problem of being unable to accurately locate the factors affecting the quality due to incomplete information collection when there is a problem with the grain quality, can accurately locate the links affecting the grain quality, improve the quality of the grain, reduce the cost of later processing, and improve the economic benefits of the grain; 3. The present invention realizes the function of providing scientific guidance for the grain production process by designing an intelligent decision-making module, solves the problems of unreasonable resource allocation, high labor cost and deviation in behavioral decision-making, can provide and analyze real-time data of the grain production process, provide decision-making support for producers and managers, optimize resource allocation, and improve production efficiency; 4. The present invention realizes the function of transparent management by designing a shared management module, solves the problems of information asymmetry, insufficient resource optimization allocation and low emergency response efficiency, can reasonably arrange production plans, avoid waste of resources, improve the accuracy and reliability of prediction, and ensure the stability, reliability and safety of food supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the composition of the multi-index prediction management platform of the present invention; Figure 2 This is a schematic diagram of the composition of the grain detection module of the present invention; Figure 3 It is a schematic diagram of the composition of the prediction and evaluation module of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1: Please refer to Figure 1 and Figure 3, a multi-index prediction management platform for grain quality analysis, the multi-index prediction management platform includes a grain detection module and a prediction evaluation module, the prediction evaluation module is connected to the grain detection module through signal transmission; The prediction and evaluation module receives the information collected by the grain detection module, performs analysis and prediction and evaluation, and transmits the results to the database; The food detection module is responsible for collecting data information on food production, processing, storage and circulation, and transmitting it to the database; The prediction and evaluation module includes: XGBoost algorithm, random forest algorithm, SVM algorithm, BP neural network algorithm, Stacking algorithm and Boosting algorithm; The Bagging algorithm constructs a base learner by extracting a training sample set using the Bootstraping method from the prediction results of the XGBoost algorithm, the random forest algorithm, the SVM algorithm, and the BP neural network algorithm, inputs the base learner generated by the Bagging algorithm and the grain information collected by the grain detection module into the meta-learner for secondary learning, constructs a Stacking integration model, trains and evaluates the Stacking integration model, and uses the trained Stacking integration model to predict and evaluate grain quality; The XGBoost algorithm analyzes the temperature, humidity and chemical content in the granary, builds a prediction model to predict the grain quality, and outputs the prediction results; The random forest algorithm classifies the collected food information, constructs a decision tree, and integrates the output results of the decision tree as the prediction result; The SVM algorithm predicts changes in grain quality by using parameters in the grain storage process through training the SVM model and outputs the prediction results. The BP neural network algorithm predicts grain quality by building a prediction model and outputs the prediction results. Furthermore, the grain detection module collects information about grain during its planting, processing, storage and circulation. The XGBoost algorithm analyzes the temperature, humidity and chemical content in the granary, builds a prediction model to predict grain quality, and outputs the prediction results. The random forest algorithm classifies the collected grain information, builds a decision tree, integrates the output results of the decision tree, automatically selects features, and obtains the final prediction by voting or averaging. The SVM algorithm uses various parameters in the grain storage process, including temperature, humidity, time, etc. to build and train the SVM model. The parameters in the grain storage process are used to predict changes in grain quality and output the prediction results. The BP neural network algorithm predicts grain quality by building prediction models, such as time series quality prediction models and temperature and humidity-quality prediction models, to predict grain storage quality indicators and output the prediction results. The agging algorithm constructs a base learner by extracting training sample sets using the Bootstraping method from the prediction results of the XGBoost algorithm, random forest algorithm, SVM algorithm and BP neural network algorithm, inputs the base learner generated by the Bagging algorithm and the grain information collected by the grain detection module into the meta-learner for secondary learning, constructs a Stacking integration model, trains and evaluates the Stacking integration model, and uses the trained Stacking integration model to predict and evaluate grain quality, thereby realizing the function of accurately predicting the changing trend of grain quality, solving the problems of insufficient prediction accuracy, poor adaptability and insufficient information utilization of a single model, avoiding the deviations and limitations of a single prediction model, comprehensively considering various factors affecting grain quality, improving the accuracy, stability and reliability of the prediction, and optimizing the utilization of computing resources and time costs.

[0020] Example 2: Please refer to Figure 1 and Figure 2 , a multi-index prediction management platform for grain quality analysis, the grain detection module includes: grain planting monitoring component, grain processing detection component, grain storage monitoring component and grain circulation detection component; The grain planting monitoring component includes: a seedling monitoring unit, a soil moisture monitoring unit, an insect monitoring unit and a disaster monitoring unit. The grain planting monitoring component is connected to the database through signal transmission; The seedling monitoring unit monitors the growth and development of grain crops during planting, including the height, number of leaves, leaf area, number of tillers, root development, leaf color and leaf density of grain crops; The soil moisture monitoring unit monitors the moisture content, temperature, humidity, pH, and nitrogen, phosphorus, and potassium content of grain-growing soil; The pest monitoring unit monitors the pest species, pest numbers, and occurrence time and location, and uses pest monitoring lights and sex traps to automatically trap pests; The disaster monitoring unit monitors natural disasters by monitoring temperature, humidity, wind speed and rainfall through meteorological sensors; Furthermore, the seedling monitoring unit uses high-definition cameras and multi-spectral sensors to capture growth images of grain seedlings, including daytime images and nighttime infrared images, to ensure all-weather uninterrupted monitoring and real-time recording of the growth process of grain crops, such as plant height, number of leaves and color changes. The sensor network in the seedling monitoring unit includes multiple sensors such as temperature, humidity, light intensity, soil moisture and carbon dioxide concentration, which monitor the parameters of the growth environment of grain crops in real time, helping farmers understand the soil moisture conditions and the growth environment of crops so that they can carry out irrigation and fertilization when needed. The soil moisture monitoring device is composed of soil temperature and humidity sensors, soil conductivity sensors and soil pH sensors, etc., to monitor the soil environment for the growth of grain crops. Farmers can intuitively understand the soil information during the grain planting stage through the shared management module without having to judge through experience. The insect monitoring unit uses the biological characteristics of pests such as phototropism, height tendency and climbing ability to monitor the growth of grain crops. Features: set specific light sources or baits to attract and trap pests, automatically process the trapped pests through far-infrared technology or other non-contact heating methods to ensure the mortality rate and completeness rate, and be equipped with high-definition cameras to collect information on the processed pest images, generate insect situation reports and early warning information, and the disaster monitoring unit collects and analyzes various disaster-related data through front-end data acquisition equipment, monitors meteorological elements such as temperature, humidity, wind speed and rainfall through meteorological sensors, and combines the early warning model for natural disasters to notify farmers in advance of possible disasters, so that farmers have enough time to take protective measures, realizing the functions of intelligent and efficient grain production, solving the problems of abnormal growth and development of grain crops, waste of water resources, pest outbreaks and large losses from natural disasters, ensuring the healthy growth of grain crops, increasing grain yield and quality, improving the utilization efficiency of water resources, improving the efficiency and effectiveness of pest control, reducing disaster losses, and reducing production costs.

[0021] Example 3: Please refer to Figure 1 and Figure 2 , a multi-index prediction management platform for food quality analysis, the food processing detection component includes: a physical detection unit, a chemical detection unit and a microbial detection unit, and the food production detection component is connected to the database through signal transmission; The physical detection unit includes a visual sensor, a moisture meter, a temperature and humidity meter, and a vibration screening machine, and the physical detection unit is connected to the database and the prediction and evaluation module through signal transmission; The visual sensor collects information on the color, shape, size, glossiness and integrity of the grain, the moisture meter detects the moisture content of the grain, the temperature and humidity meter detects the temperature and humidity of the grain, and the vibration screening machine detects the size distribution of grain particles and transmits the information to the database to evaluate the processing and edible quality of the grain. The chemical detection unit includes a food heavy metal detector, a gas chromatograph and a liquid chromatograph, and the chemical detection unit is connected to the database and the prediction and evaluation unit through signal transmission; Grain heavy metal detector detects harmful heavy metals in grain, gas chromatograph detects pesticides with high volatility and high thermal stability, and liquid chromatograph detects pesticides with high polarity and low thermal stability; The microbial detection unit includes a mycotoxin detector and an enzyme marker, and the microbial detection unit is connected with the database and the prediction and evaluation module through signal transmission; The mycotoxin detector detects mycotoxins in grains, and the ELISA reader detects non-edible chemicals, abused food additives, heavy metals, nutritional enhancers and chemical residues in grains; Furthermore, during the grain processing process, visual sensors are used to collect information on the color, shape, size, glossiness and integrity of the grain, moisture meters are used to detect the moisture content of the grain, temperature and humidity meters are used to detect the temperature and humidity of the grain, vibration screening machines are used to detect the size distribution of grain particles and transmit the information to a database to evaluate the processing and edible quality of the grain, grain heavy metal detectors are used to detect harmful heavy metals in the grain, gas chromatographs are used to detect pesticides with high volatility and high thermal stability, liquid chromatographs are used to detect pesticides with large polarity and low thermal stability, mycotoxins detectors are used to detect mycotoxins in the grain, and ELISA instruments are used to detect non-edible chemicals, abused food additives, heavy metals, nutritional enhancers and chemical residues in the grain, thereby ensuring the quality, safety and nutrition of the grain, solving the problems of inaccurate collection of grain quality information and non-compliance with edible standards, preventing defective products from entering the market, improving the personal safety of consumers, and improving the reliability and credibility of grain quality.

[0022] Example 4: Please refer to Figure 1 and Figure 2 , a multi-index prediction management platform for grain quality analysis, wherein the grain storage monitoring component includes: a temperature and humidity monitoring unit, a moisture monitoring unit, a pest density monitoring unit and a grain quality monitoring unit; The temperature and humidity monitoring unit monitors the temperature and humidity in the granary. The temperature range of the granary is 15℃~25℃, and the humidity range is 55%~65%; The moisture monitoring unit monitors the moisture content of grain stored in the granary; The pest density monitoring unit monitors the pest density in the granary; The grain quality monitoring unit monitors the grain color, smell, fatty acid esters, taste scores and food safety indicators in the granary; The grain circulation detection component comprises: a packaging quality detection unit and a simulated transportation environment unit; The packaging quality inspection unit includes a texture analyzer, an intelligent electronic tensile tester, an impact tester, a rubbing resistance tester, a thickness tester, and a friction coefficient tester. The packaging quality inspection unit is connected to the database through signal transmission; The texture analyzer is used to test the hardness, brittleness, stickiness and breaking strength of food packaging materials. The intelligent electronic tensile testing machine is used to test the tensile strength, peel strength and heat-sealing strength of packaging materials. The impact tester is used to test the impact resistance of packaging materials. The rubbing resistance tester is used to test the performance of packaging when rubbing, bending, twisting and squeezing. The thickness tester is used to test the thickness uniformity of packaging materials. The friction coefficient tester is used to evaluate the slippery performance of the inside and outside of packaging materials. The simulated transport environment unit includes a packaged goods vibration tester, a vacuum sealing tester and a temperature and humidity conditioning box, and the simulated transport environment unit is connected to the database through signal transmission; The packaged goods vibration tester is used to evaluate the strength of the package under sinusoidal vibration and resonance and its ability to protect grain. The vacuum seal tester is used to test the sealing performance of the package. The temperature and humidity conditioning box is used to test the protection performance of the package on grain quality under different temperature and humidity conditions. Furthermore, during the grain storage process, the internal environment of the granary and the grain need to be tested. The temperature range in the granary is 15℃~25℃, and the humidity range is 55%~65%. Different types of grains have different moisture contents during storage. Among cereals, the storage moisture content of indica rice is below 13%, hard rice is below 14%, wheat and barley are below 12.5%, corn is stored at 12%~18%, beans are stored at 10%~13%, flour is stored at 13%~15%, and peanuts are stored at 15%. The storage moisture content of fruit is below 10%, the storage moisture content of peanut kernel is below 9%, the storage moisture content of sesame is below 8%, and the storage moisture content of coix seed is 12%. At the same time, there will be differences due to factors such as region, climate conditions, storage facilities and technical level. During the grain storage process, the moisture content of grain needs to be tested regularly. The grain storage monitoring component monitors the temperature and humidity of the granary, the moisture content of the grain, the pest density in the granary and the quality of the grain in the granary. The grain circulation detection component monitors the quality of grain during circulation and transportation. The tests mainly include the testing of food packaging and the testing of changes in grain quality during simulated transportation. The texture analyzer tests the hardness, brittleness, stickiness and breaking strength of food packaging materials. The intelligent electronic tensile testing machine is used to test the tensile strength, peel strength and heat sealing strength of packaging materials. The impact tester is used to test the impact resistance of packaging materials. The rubbing resistance tester is used to test the performance of packaging when rubbing, bending, twisting and squeezing. The thickness tester is used to test the thickness uniformity of packaging materials. The friction coefficient tester is used to evaluate the smoothness of the inside and outside of the packaging materials. The packaged goods vibration testing machine is used to evaluate the strength of the packaging under sinusoidal vibration and resonance and its ability to protect grain. The vacuum sealing tester is used to test the sealing performance of the packaging. The temperature and humidity conditioning box is used to test the protection performance of the packaging for grain quality under different temperature and humidity conditions, realizing the function of ensuring food safety and quality stability, solving the problem that the conditions during grain storage and transportation are difficult to understand and control and reduce the quality of grain, which can reduce the damage of grain during transportation, improve the economic benefits of grain, and ensure the quality of grain.

[0023] Example 5: Please refer to Figure 1 and Figure 3 , a multi-index prediction management platform for grain quality analysis, the multi-index prediction management platform includes a grain detection module and a prediction evaluation module, the prediction evaluation module is connected to the grain detection module through signal transmission; The prediction and evaluation module receives the information collected by the grain detection module, performs analysis and prediction and evaluation, and transmits the results to the database; The food detection module is responsible for collecting data information on food production, processing, storage and circulation, and transmitting it to the database; The prediction and evaluation module includes: XGBoost algorithm, random forest algorithm, SVM algorithm, BP neural network algorithm, Stacking algorithm and Boosting algorithm; The Bagging algorithm uses the Bootstraping method to extract a training sample set from the grain information collected by the grain detection module to construct a base learner, inputs the base learner generated by the Bagging algorithm and the grain information collected by the grain detection module into the meta learner for secondary learning, constructs a Stacking integration model, trains and evaluates the Stacking integration model, and uses the trained Stacking integration model to predict and evaluate grain quality; The prediction and evaluation module is connected to the intelligent decision-making module through signal transmission. The intelligent decision-making module combines the prediction results of the prediction and evaluation module with market demand, consumer preferences and policy orientation to provide managers with decision support on planting structure, irrigation and fertilization schemes, harvesting and storage conditions; The multi-index prediction management platform also includes a shared management module, which is a touch-controlled terminal for managers to understand the grain quality prediction results and control the grain quality; Furthermore, during the grain planting stage, farmers' judgments on the growth of grain crops mostly come from their own experience. They visually observe the leaf color, growth height, flowering and fruiting of crops as indicators of growth conditions. The results of visual observation are affected by factors such as farmers' experience, knowledge background and environmental conditions during observation, resulting in certain deviations in the final observation and judgment results. In addition, many diseases and insect pests do not have obvious symptoms in the early stages, or are hidden in the back of leaves and roots of crops that are difficult to observe directly. At this time, it is difficult to detect them in time by naked eye observation, and they are often not discovered until they affect the growth of crops. Ultimately, the growth of grain crops is seriously affected, resulting in a reduction in grain production or even a total loss of harvest. The intelligent decision-making module is based on the information collected by the grain detection module, including multi-dimensional information such as soil moisture, light intensity, temperature, crop growth conditions, meteorological data and soil information, combined with the prediction results of the prediction and evaluation module on grain quality, to generate the best grain planting plan, including sowing time, planting density, irrigation plan, fertilization plan, harvest time and storage conditions. Decision support, grain The food detection module displays indicators such as soil moisture, temperature and nutrient content collected during the planting stage of grain crops to farmers through a touch terminal, and analyzes them through the intelligent decision-making module to help farmers understand the impact of soil on grain quality. When problems occur with grain crops, farmers can use the touch terminal to understand the area where the problem occurs, the type of problem and the solution to the problem, and can solve the problem of grain crops as soon as possible. During grain processing, storage and circulation, managers can understand the status of grain in real time through the touch terminal, without the need for frequent manual inspection of grain. The analysis results and decision support given by the intelligent decision-making module can provide good guarantee for grain quality, realize the function of providing scientific guidance for the grain production process to improve grain quality, solve the problems of unreasonable resource allocation, high labor costs, behavioral decision-making deviations, information asymmetry and insufficient resource optimization allocation, and can provide real-time data of the grain production process and analyze it, provide decision support for producers and managers, optimize resource allocation, improve production efficiency, and ensure the stability, reliability and safety of grain supply.

[0024] Working principle: The seedling monitoring unit uses high-definition cameras and multi-spectral sensors to capture growth images of grain seedlings, including daytime images and nighttime infrared images, to ensure all-weather uninterrupted monitoring and real-time recording of the growth process of grain crops, such as plant height, number of leaves and color changes. The sensor network in the seedling monitoring unit includes temperature, humidity, light intensity, soil moisture, carbon dioxide concentration and other sensors to monitor the parameters of the growth environment of grain crops in real time, helping farmers understand the soil moisture conditions and the growth environment of crops so that they can carry out irrigation and fertilization when needed. The soil moisture monitoring device is composed of soil temperature and humidity sensors, soil conductivity sensors and soil pH sensors to monitor the soil environment for the growth of grain crops. Farmers can share management information. The module can intuitively understand the information of the soil during the grain planting stage without having to make judgments based on experience. The insect monitoring unit uses the biological characteristics of pests such as phototropism, height tendency and climbing ability to set specific light sources or baits to attract and trap pests. It automatically processes the trapped insects through far-infrared technology or other non-contact heating methods to ensure the mortality rate and completeness rate. At the same time, it is equipped with high-definition cameras to collect information on the processed pest images and generate insect reports and early warning information. The disaster monitoring unit collects and analyzes various data related to the disaster through front-end data acquisition equipment, monitors meteorological elements such as temperature, humidity, wind speed and rainfall through meteorological sensors, and combines the early warning model for natural disasters to notify farmers of possible disasters in advance, so that farmers have enough time to take precautions. During the grain processing, visual sensors are used to collect information on the color, shape, size, glossiness and integrity of the grain. Moisture meters are used to detect the moisture content of the grain. Temperature and humidity meters are used to detect the temperature and humidity of the grain. Vibration screening machines are used to detect the size distribution of grain particles and transmit the information to the database to evaluate the processing and edible quality of the grain. Grain heavy metal detectors are used to detect harmful heavy metals in grain. Gas chromatographs are used to detect pesticides with high volatility and high thermal stability. Liquid chromatographs are used to detect pesticides with high polarity and low thermal stability. Mycotoxin detectors are used to detect mycotoxins in grain. Enzyme markers are used to detect non-edible chemicals, abused food additives, heavy metals, nutritional enhancers and chemical residues in grain. Grain storage monitoring components are used to monitor the temperature and humidity of the granary, the moisture content of the grain, the density of pests in the granary and the quality of the grain in the granary. The grain circulation detection component detects the quality of grain during the circulation and transportation process, mainly including the detection of food packaging and the detection of changes in grain quality during simulated transportation. The texture analyzer detects the hardness, brittleness, viscosity and fracture strength of food packaging materials. The intelligent electronic tensile testing machine is used to detect the tensile strength, peel strength and heat sealing strength of packaging materials. The impact tester is used to detect the impact resistance of packaging materials. The rubbing resistance tester is used to detect the performance of packaging when rubbing, bending, twisting and squeezing. The thickness tester is used to detect the thickness uniformity of packaging materials. The friction coefficient tester is used to evaluate the smoothness of the inside and outside of the packaging material. The packaged goods vibration tester is used to evaluate the strength of the packaging under sinusoidal vibration and resonance and the protection ability of the packaging to the grain. The vacuum sealing tester is used to detect the sealing performance of the packaging. The temperature and humidity conditioning box is used to detect the protection performance of the packaging to the grain quality under different temperature and humidity conditions. The grain detection module collects information about grain during its planting, processing, storage and circulation. The XGBoost algorithm analyzes the temperature, humidity and chemical content in the granary, builds a prediction model to predict grain quality, and outputs the prediction results. The random forest algorithm classifies the collected grain information, builds a decision tree, integrates the output results of the decision tree, automatically selects features, and obtains the final prediction through voting or averaging. The SVM algorithm uses various parameters in the grain storage process, including temperature, humidity, time, etc. to build and train the SVM model. It uses the parameters in the grain storage process to predict changes in grain quality and outputs the prediction results. The BP neural network algorithm The method predicts grain quality by building a prediction model, such as a time series quality prediction model and a temperature and humidity-quality prediction model, to predict grain storage quality indicators and output prediction results. The Bagging algorithm uses the Bootstraping method to extract training sample sets from grain information collected by the grain detection module to build a base learner. The base learner generated by the Bagging algorithm and the grain information collected by the grain detection module are input into the meta-learner for secondary learning, and a Stacking integration model is built. The Stacking integration model is trained and evaluated, and the trained Stacking integration model is used to predict and evaluate grain quality. The intelligent decision-making module is based on the information collected by the grain detection module, including multi-dimensional information such as soil moisture, light intensity, temperature, crop growth status, meteorological data and soil information, and combines the prediction results of the prediction and evaluation module on grain quality to generate the best grain planting plan, including decision support on sowing time, planting density, irrigation plan, fertilization plan, harvest time and storage conditions. The grain detection module displays indicators such as soil moisture, temperature and nutrient content collected by grain crops during the planting stage to farmers through a touch terminal, and analyzes them through the intelligent decision-making module to help farmers understand the impact of soil on grain quality. When problems occur with grain crops, farmers can use the touch terminal to understand the area, type and solution of the problem, and can solve the problems with grain crops as soon as possible. During the grain processing, storage and circulation process, managers can understand the status of grain in real time through the touch terminal, without the need for frequent manual inspection of grain. The analysis results and decision support provided by the intelligent decision-making module can ensure good grain quality.

[0025] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A multi-index prediction management platform for grain quality analysis, characterized by: The multi-index forecasting management platform includes a grain detection module and a forecasting and evaluation module, and the forecasting and evaluation module is connected to the grain detection module through signal transmission; The prediction and evaluation module receives the information collected by the grain detection module, performs analysis and prediction and evaluation, and transmits the results to the database; The food detection module is responsible for collecting data information on food production, processing, storage and circulation, and transmitting it to the database; The prediction and evaluation module includes: XGBoost algorithm, random forest algorithm, SVM algorithm, BP neural network algorithm, Stacking algorithm and Boosting algorithm; The Bagging algorithm constructs a base learner by extracting a training sample set using the Bootstraping method from the prediction results of the XGBoost algorithm, the random forest algorithm, the SVM algorithm and the BP neural network algorithm, inputs the base learner generated by the Bagging algorithm and the grain information collected by the grain detection module into the meta-learner for secondary learning, constructs a Stacking integration model, trains and evaluates the Stacking integration model, and uses the trained Stacking integration model to predict and evaluate grain quality.

2. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The grain detection module includes: a grain planting monitoring component, a grain processing detection component, a grain storage monitoring component and a grain circulation detection component; The grain planting monitoring component includes: a seedling monitoring unit, a soil moisture monitoring unit, an insect monitoring unit and a disaster monitoring unit. The grain planting monitoring component is connected to the database through signal transmission; The seedling monitoring unit monitors the growth and development of grain crops during planting, including the height, number of leaves, leaf area, number of tillers, root development, leaf color and leaf density of grain crops; The soil moisture monitoring unit monitors the moisture content, temperature, humidity, pH, and nitrogen, phosphorus, and potassium content of grain-growing soil; The pest monitoring unit monitors the pest species, pest numbers, and occurrence time and location, and uses pest monitoring lights and sex traps to automatically trap pests; The disaster monitoring unit monitors natural disasters by monitoring temperature, humidity, wind speed and rainfall through meteorological sensors.

3. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The prediction and evaluation module is connected to the intelligent decision-making module through signal transmission. The intelligent decision-making module combines the prediction results of the prediction and evaluation module with market demand, consumer preference and policy orientation to provide managers with decision support on planting structure, irrigation and fertilization plan, harvesting and storage conditions.

4. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The multi-index prediction management platform also includes a shared management module, which is a touch-controlled terminal used for managers to understand the food quality prediction results and control the food quality.

5. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The XGBoost algorithm analyzes the temperature, humidity and chemical content in the granary, builds a prediction model to predict the quality of grain, and outputs the prediction results.

6. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The random forest algorithm classifies the collected food information, constructs a decision tree, and integrates the output results of the decision tree as the prediction result.

7. The multi-index prediction management platform for food quality analysis according to claim 1, characterized in that: The SVM algorithm predicts changes in grain quality by using parameters in the grain storage process through training the SVM model and outputs the prediction results. The BP neural network algorithm predicts grain quality by building a prediction model and outputs the prediction results.

8. The multi-index prediction management platform for grain quality analysis according to claim 2, characterized in that: The food processing detection component includes: a physical detection unit, a chemical detection unit and a microbial detection unit, and the food production detection component is connected to the database through signal transmission; The physical detection unit includes a visual sensor, a moisture meter, a temperature and humidity meter, and a vibration screening machine, and the physical detection unit is connected to the database and the prediction and evaluation module through signal transmission; The visual sensor collects information on the color, shape, size, glossiness and integrity of the grain, the moisture meter detects the moisture content of the grain, the temperature and humidity meter detects the temperature and humidity of the grain, and the vibration screening machine detects the size distribution of grain particles and transmits the information to the database to evaluate the processing and edible quality of the grain. The chemical detection unit includes a food heavy metal detector, a gas chromatograph and a liquid chromatograph, and the chemical detection unit is connected to the database and the prediction and evaluation unit through signal transmission; Grain heavy metal detector detects harmful heavy metals in grain, gas chromatograph detects pesticides with high volatility and high thermal stability, and liquid chromatograph detects pesticides with high polarity and low thermal stability; The microbial detection unit includes a mycotoxin detector and an enzyme marker, and the microbial detection unit is connected with the database and the prediction and evaluation module through signal transmission; The mycotoxin detector detects the mycotoxins in the grain, and the enzyme labeler detects the non-edible chemicals, abused food additives, heavy metals, nutritional enhancers and chemical residues in the grain.

9. The multi-index prediction management platform for grain quality analysis according to claim 2, characterized in that: The grain storage monitoring component includes: a temperature and humidity monitoring unit, a moisture monitoring unit, a pest density monitoring unit and a grain quality monitoring unit; The temperature and humidity monitoring unit monitors the temperature and humidity in the granary. The temperature range of the granary is 15℃~25℃, and the humidity range is 55%~65%; The moisture monitoring unit monitors the moisture content of grain stored in the granary; The pest density monitoring unit monitors the pest density in the granary; The grain quality monitoring unit monitors the color, smell, fatty acid esters, taste scores and food safety indicators of grain in the granary.

10. The multi-index prediction management platform for grain quality analysis according to claim 2, characterized in that: The grain circulation detection component comprises: a packaging quality detection unit and a simulated transportation environment unit; The packaging quality inspection unit includes a texture analyzer, an intelligent electronic tensile tester, an impact tester, a rubbing resistance tester, a thickness tester, and a friction coefficient tester. The packaging quality inspection unit is connected to the database through signal transmission; The texture analyzer is used to test the hardness, brittleness, stickiness and breaking strength of food packaging materials. The intelligent electronic tensile testing machine is used to test the tensile strength, peel strength and heat-sealing strength of packaging materials. The impact tester is used to test the impact resistance of packaging materials. The rubbing resistance tester is used to test the performance of packaging when rubbing, bending, twisting and squeezing. The thickness tester is used to test the thickness uniformity of packaging materials. The friction coefficient tester is used to evaluate the slippery performance of the inside and outside of packaging materials. The simulated transport environment unit includes a packaged goods vibration tester, a vacuum sealing tester and a temperature and humidity conditioning box, and the simulated transport environment unit is connected to the database through signal transmission; The packaged goods vibration testing machine is used to evaluate the strength of the packaging under sinusoidal vibration and resonance and its ability to protect grain. The vacuum sealing tester is used to test the sealing performance of the packaging. The temperature and humidity conditioning box is used to test the protection performance of the packaging on grain quality under different temperature and humidity conditions.

Images