Intelligent research and production control method for high value-added products of deep processing of corn

Abstract

The present application relates to a corn deep processing high value-added product intelligent research and development and production management and control method, belonging to the field of intelligent agriculture technology. The method comprises: through AI intelligent agent, corn product ranking data on an e-commerce platform is grabbed and analyzed, and high value-added deep processing categories are identified; soil, climate and market data are fused, and accurate planting decisions are generated by an AI large model; in response to product categories automatically calling process parameters, through an industrial control system, a full-automatic production line is driven to complete processing, and closed-loop regulation and control is realized by using online quality detection; through Internet of Things sensors, field growth data is collected in real time and pushed to consumer terminals, supporting corn field adoption and periodic subscription distribution. The present application realizes intelligent closed loop from product selection, planting, processing to sales, solves the problem of small farmers' production and marketing disconnection and low added value, and improves the industry chain collaboration efficiency and product standardization level.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, specifically to a method for intelligent research and development and production control of high value-added corn deep-processing products. Background Technology

[0002] Corn cultivation in my country is mainly carried out by small-scale family farms, whose production and management generally face the following technical difficulties: First, planting decisions lack data support. Key agricultural operations such as sowing time, fertilization plans, and irrigation strategies rely heavily on traditional experience, making it difficult to accurately align with the actual needs of the target market. This leads to fierce competition due to homogeneity and a sharp drop in prices when the harvest season is concentrated. Second, post-harvest processing capabilities are weak. More than 60% of small farmers can only sell raw grains, lacking deep processing technology and equipment. They are forced to rely on traditional intermediary channels, and information isolation results in low bargaining power, making them highly susceptible to the passive situation of "low grain prices hurting farmers." Third, the production and sales chain from field to table is broken. Farmers cannot directly obtain product preference data from consumers and find it difficult to establish branded sales channels. Grain stockpiling and low returns coexist, seriously restricting the sustainability of agricultural production.

[0003] To address the aforementioned issues, some existing agricultural informatization solutions exist. For example, some solutions monitor soil moisture by deploying sensors in the fields or provide weather warnings through mobile applications; others focus on building e-commerce platforms for agricultural products, attempting to broaden sales channels. However, these solutions are mostly limited to the digital transformation of a single link, failing to establish a closed loop across the entire chain of "product selection—planting—processing—sales." Specifically, existing technologies lack mechanisms to link e-commerce consumption data with front-end planting decisions, making it impossible to guide planting varieties and processing directions based on high-value-added products in the market. Simultaneously, existing automated processing equipment is mostly geared towards large grain enterprises, lacking lightweight production lines adapted to small farmers' raw material batches, capable of rapid category switching, and possessing online quality closed-loop control capabilities. Furthermore, existing adoption or subscription-based e-commerce systems are not deeply integrated with field IoT sensor data, preventing consumers from obtaining real-time information on crop growth status and resulting in weak trust mechanisms. Therefore, systematically solving the problems of production-sales disconnect, low added value, and information asymmetry for small farmers from a technological perspective is a crucial technological direction that urgently needs breakthroughs in this field. Summary of the Invention

[0004] To address the problems of existing technologies, this invention provides a method for intelligent R&D and production control of high value-added corn deep-processed products.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: a method for intelligent R&D and production control of high value-added corn deep-processed products, comprising the following steps performed in sequence: Step S1: AI-driven product selection based on e-commerce data; Collect corn product ranking data from at least one e-commerce platform. The data includes product name, sales ranking, price range, and user review tags. Use an AI agent to call a large language model to analyze the ranking data, identify the corn deep-processed product category with the highest added value, and output the target product type identifier. Step S2: AI-driven planting decision-making based on multi-source data fusion; Collect soil parameters, climate parameters, and market data for the target planting area. Input the parameters and data, along with the target product type identifier output in step S1, into a pre-trained AI model. After calculation, the AI ​​model outputs precise planting decision information, which includes at least sowing time, planting density, fertilization plan, and irrigation plan. Step S3: Automated production control based on product category; In response to the target product type identifier output in step S1, the corresponding production process parameter set is retrieved from the process database; control commands are sent to the fully automated processing production line through the industrial control system to process the corn raw material into the target deep-processed product, and the quality inspection data of the production line is acquired in real time. Step S4: IoT adoption and subscription sales management; The system collects real-time sensor data on the corn growth process by deploying IoT sensor nodes in the field and pushes the sensor data to the terminal user interface; it establishes a virtual unit of "one acre of cornfield" in response to the user's adoption instruction, and records the corn variety and fertilizer type selected by the user; and it generates a periodic product delivery plan in response to the user's subscription instruction.

[0006] In one specific implementation, the construction and operation of the AI ​​agent in step S1 is as follows: an analysis instruction is generated through DeepSeek, an AI agent instance is created in the Tencent Yuanqi platform, a large language model is loaded to parse the ranking data of the top 100 corn products on the e-commerce platform, and "corn pancake" is output as the target product type identifier.

[0007] In one specific implementation, the soil parameters in step S2 include: pH value, organic matter content, soil moisture and electrical conductivity; the climate parameters include: daily average temperature, precipitation, light intensity and wind speed; the market data includes: corn spot price and local purchase price fluctuation curve.

[0008] In one specific implementation, the soil moisture is collected in the range of relative water content from 10% to 90%, and the collection frequency is once every 30 minutes; the daily average temperature is collected in the range of -10℃ to 40℃, and the collection frequency is once every 10 minutes.

[0009] In one specific implementation, the fertilization plan generated in step S2 includes the ratio and application amount of nitrogen, phosphorus and potassium fertilizers, wherein the nitrogen-phosphorus-potassium ratio is in the form of N-P2O5-K2O, and the application amount ranges from 30 kg to 80 kg per mu; in the irrigation plan, an irrigation command is triggered when the soil moisture content is lower than 55% relative water content.

[0010] In one specific embodiment, the fully automated processing production line in step S3 is a fully automated corn pancake production line, and the set of production process parameters includes: grinding fineness of 60 to 120 mesh, fermentation temperature of 25 to 35°C, fermentation time of 2 to 6 hours, frying temperature of 150 to 200°C, frying time of 30 to 90 seconds, and finished product moisture control threshold of ≤10%.

[0011] In one specific implementation, the quality inspection data in step S3 includes: the thickness, diameter, color uniformity, and moisture content of the finished pancake. When any indicator exceeds a preset threshold, the industrial control system automatically adjusts the frying temperature or frying time.

[0012] In one specific implementation, the IoT sensor node in step S4 includes: a soil moisture sensor, a light intensity sensor, and a crop image acquisition camera, wherein the image acquisition camera acquires field images with a resolution of 1080p or higher once a day.

[0013] In one specific implementation, the adoption instruction in step S4 includes: the consumer selecting the corn variety type (sweet corn, waxy corn, or regular corn) and fertilizer type (organic fertilizer or compound fertilizer) through an interactive interface; the subscription instruction includes selecting the subscription period (monthly, quarterly, or annual) and flavor preference tags.

[0014] An integrated management system for corn planting and sales based on artificial intelligence and the Internet of Things is used to implement intelligent R&D and production control methods for high value-added corn deep-processing products, including: The data acquisition module is configured to acquire e-commerce platform ranking data, soil parameters, climate parameters, and market data. The AI ​​intelligent agent product selection module is connected to the data acquisition module and is used to identify high value-added corn deep-processed product categories; The AI ​​large-scale model planting decision module is connected to the data acquisition module and the AI ​​intelligent agent product selection module respectively, and is used to generate precise planting decision information; An automated production control module, connected to the AI ​​intelligent agent product selection module, has a built-in process parameter database and is used to drive a fully automated corn pancake production line. The IoT sales management module is connected to the field sensor nodes in the data acquisition module and is configured to display crop growth sensing data and process cornfield adoption orders and subscription delivery requests.

[0015] The beneficial effects of this invention are as follows: 1. This invention significantly improves the intelligence level of the entire corn industry chain through a collaborative mechanism of AI-driven product selection based on e-commerce data and AI-powered large-scale model planting decisions based on multi-source data fusion. Specifically, in step S1, the AI ​​agent automatically captures and analyzes e-commerce platform ranking data to identify high-value-added deep-processed product categories, solving the technical problems of blind product selection and lack of market foresight among traditional small farmers, enabling the production end to dynamically respond to market demands. In step S2, soil parameters, climate parameters, and market data are integrated, and the AI-powered large-scale model generates precise planting decisions (including sowing time, planting density, fertilization plan, irrigation threshold, and harvest window), overcoming the defects of resource waste and unstable yields caused by experience-based planting, and achieving a technological leap from "experience-driven" to "data-driven." The linkage of the above two steps creates a data closed loop between planting varieties and processed product categories, ensuring the market adaptability and production efficiency of agricultural products from the source.

[0016] 2. This invention constructs a closed-loop system from the field to the consumer end through the bidirectional coupling of automated production control and IoT sales management. In the processing stage, step S3 automatically retrieves process parameter databases based on product categories, drives a fully automated production line through an industrial control system, and utilizes an online quality inspection module to monitor key indicators such as moisture, thickness, and color in real time, achieving closed-loop feedback adjustment. This overcomes the technical defects of traditional manual processing, such as large quality fluctuations and low pass rates, ensuring the standardization and stability of deep-processed products. In the sales stage, step S4 collects field growth data in real time through IoT sensors and pushes it to consumer terminals, establishing a virtual adoption unit and a periodic subscription delivery model. This directly opens up the data channel between producers and consumers, eliminating the information asymmetry and profit squeeze problems caused by multiple intermediaries in traditional sales. Meanwhile, the interactive data display of the adoption model enhances consumers' trust in the production process, and the subscription service enables the demand side to guide the supply side in reverse, ultimately forming a virtuous cycle of "production based on sales, processing based on production, and planting based on processing". From a technical perspective, it systematically solves the fundamental problems faced by small farmers, such as the disconnect between production and sales, low added value, and weak risk resistance. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall process of the method of the present invention.

[0018] Figure 2 This is a schematic diagram of the system architecture of the present invention.

[0019] Figure 3 This is a schematic diagram of the control process of the fully automatic corn pancake production line of the present invention. Detailed Implementation

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

[0021] like Figures 1 to 3 The method for intelligent R&D and production control of high value-added corn deep-processed products is shown.

[0022] System overall architecture.

[0023] The method described in this invention relies on a distributed system and includes: Data acquisition layer: IoT sensor node clusters deployed in the target planting area, e-commerce platform data crawler module, and market data interface.

[0024] AI Decision Layer: Includes AI intelligent agent product selection module (deployed on cloud server) and AI large model planting decision module (based on pre-trained model of Transformer or LSTM architecture).

[0025] Execution layer: Industrial control system (Programmable Logic Controller (PLC) or edge computing gateway) and fully automated corn pancake production line.

[0026] Interaction Layer: A consumer-facing mobile / web IoT platform that supports adoption, subscription, and crop growth visualization.

[0027] The data flow between the layers of the system is as follows: data acquisition layer → AI decision layer → execution layer; at the same time, the interaction layer communicates bidirectionally with the data acquisition layer (acquiring sensor data and receiving user instructions).

[0028] Step S1: E-commerce data-driven AI product selection Input: Corn category product data from at least one e-commerce platform (such as JD.com or Taobao), including: product name, sales volume in the past 30 days, price range, high-frequency keyword tags in user reviews, and processing method field in the product description.

[0029] Processing procedure: Use web crawling tools (such as the Scrapy framework) to crawl the aforementioned data daily and store it in a NoSQL database.

[0030] Building an AI Agent: Create an agent instance in the Tencent Elements platform and set the system commands as follows: "You are an agricultural market analyst. Please analyze the following list of corn products and identify the type of deep-processed products with the highest added value. Value-added evaluation indicators include: unit price (yuan / 500g), sales growth rate, and frequency of keywords such as 'convenient,' 'nutritious,' and 'taste' in user reviews. Output format: product name, reasons for recommendation, and suggested processing technology." Call the DeepSeek large model to generate analysis instruction template, input the data of the top 100 products into the agent, and the large model outputs the results.

[0031] If the probability of "corn pancake" appearing in the output exceeds the threshold (e.g., 80%), then it is identified as the target product category.

[0032] Output: Target product type identifier (e.g., product_id = "corn_pancake"), and corresponding suggested processing attributes (flavor, packaging specifications).

[0033] Parameter range: capture frequency once a day; sales statistics period is the past 30 days; price normalized to yuan / 500g.

[0034] Step S2: AI-driven planting decisions based on multi-source data fusion.

[0035] Input: Soil parameters: pH value (4.5~9.0, accuracy ±0.1), organic matter content (0.5%~5%, accuracy ±0.1%), soil moisture (relative water content 10%~90%, accuracy ±2%), electrical conductivity (0~10 mS / cm).

[0036] Climate parameters: daily average temperature (-10℃~40℃, accuracy ±0.5℃), precipitation (0~200 mm / day), light intensity (0~150 kLux), wind speed (0~20 m / s).

[0037] Market data: Corn spot price (yuan / kg) and local purchase price fluctuation curve over the past 30 days.

[0038] The target product type identifier is output in step S1.

[0039] Processing procedure: The above data is normalized and then concatenated into a multidimensional feature vector.

[0040] Call pre-trained AI large models (such as time series prediction models based on LSTM or GPT architecture, which have been trained on historical data from the main corn producing areas in Northeast and North China).

[0041] The large model performs inference and outputs the following decision variables: Sowing time (in ten-day periods, such as "early April"); Planting density (plants / acre); Fertilization plan: N-P2O5-K2O ratio and dosage (kg / mu); Irrigation trigger threshold (soil moisture percentage); Expected harvest time window (start of harvest date ~ end of harvest date); Output: Structured planting decision instructions, which can be sent to farmers' mobile devices or agricultural machinery operation terminals.

[0042] Explanation of the reasonableness of the parameter range: The pH value ranges from 4.5 to 9.0, covering soil types in major corn-growing areas of my country.

[0043] Soil moisture below 55% triggers irrigation as a general threshold for water-saving irrigation of maize (derived from the "Technical Specification for Maize Irrigation").

[0044] The fertilizer application rate of 30-80 kg / mu complies with the "Technical Specifications for Formula Fertilization of Maize" (NY / T 2911-2016).

[0045] Step S3: Automated production control based on product category.

[0046] Input: The target product type identifier (e.g., "corn pancake") output from step S1, and the batch information of the corn raw material to be processed (moisture content, bulk density).

[0047] Processing procedure: Based on the product type identifier, retrieve the corresponding preset process parameter set from the process parameter database. For corn pancakes, the parameter set includes:

[0048] The industrial control system (programmable logic controller, PLC) receives the above parameters and sends control signals to each station of the fully automated corn pancake production line: Grinding motor speed adjustment (corresponding to grinding fineness); Fermenter heating rod power and stirring speed (corresponding to fermentation temperature and time); PID control of grilling drum temperature (corresponding to grilling temperature); Cutting and packaging servo motor stroke (corresponding to pancake thickness).

[0049] The online testing module acquires quality data in real time. Infrared moisture meter measures the moisture content of finished product; Laser displacement sensor for thickness measurement; Industrial cameras capture images, and edge computing is used to analyze color uniformity (color difference ΔE).

[0050] Closed-loop adjustment logic: If the thickness deviation exceeds ±0.2 mm, the PLC will automatically adjust the gap between the pressure rollers; If the moisture content is greater than 10%, the frying time will be automatically extended (in 5-second increments, up to a maximum of 90 seconds). If the color is abnormal (color difference value ΔE > 3), an alarm will be triggered and the batch will be rejected.

[0051] Outputs: Production line control instruction sequence, quality inspection log, and qualified product count.

[0052] Step S4: IoT adoption and subscription sales management.

[0053] Input: Field sensor data: soil moisture, light intensity, crop images (once daily, resolution 1080p or higher).

[0054] Consumer terminal interaction instructions: adoption request (including plot number, variety selection, fertilizer selection), subscription request (cycle, flavor preference).

[0055] Processing procedure: A virtual unit of "one acre of cornfield" is established. Each virtual unit corresponds to the GPS coordinate range of an actual plot of land. After the consumer pays the adoption fee, the system assigns the unit.

[0056] Real-time sensor data is uploaded to the cloud platform via the MQTT protocol, processed, and then pushed to the consumer terminal via WebSocket. It is displayed in the form of charts and images overlay: soil moisture curve, growth stage identification (based on image classification model), and expected harvest countdown.

[0057] Subscription module: Based on consumers' selected flavor preference tags (such as "locust flower," "pumpkin," and "red date"), and combined with the seasonal crop phenological calendar, a delivery plan is generated. For example, locust flower corn pancakes are launched in spring (March to May), and pumpkin corn pancakes are launched in autumn (September to November).

[0058] Delivery instructions are automatically sent to the partner logistics provider's API, generating a tracking number and providing feedback to the consumer.

[0059] Output: Adoption order records (including variety, fertilizer, and plot), subscription cycle schedule, and daily delivery task list.

[0060] Example

[0061] Example 1: Applied to the corn planting area of ​​Gengzhuang Village, Shenze County, Shijiazhuang City, Hebei Province.

[0062] 1. Overview of the pilot areas Gengzhuang Village currently has 126 smallholder corn farmers, cultivating approximately 800 mu (about 53 hectares). In the three years prior to the project's implementation, the corn purchase price dropped from 1.2 yuan / kg to 0.9 yuan / kg, resulting in an average net profit of only 150-200 yuan per mu for farmers, with some years even showing losses. The project team deployed the system described in this invention in the village.

[0063] 2. Step execution process.

[0064] S1 execution status: On March 1, 2025, the system crawled the top 100 corn products on Taobao and JD.com. Data shows that corn pancakes occupied 5 of the top 10 spots, with an average price of 28 yuan per 500g. Frequently used user reviews included "convenient breakfast," "additive-free," and "healthy whole grains."

[0065] The AI ​​agent (Tencent Yuanqi + DeepSeek) output the conclusion: "It is recommended to develop corn pancakes, and it is suggested to add red dates and purple rice flavors, and use individual small packages." The system sets "Corn Pancakes (Red Date / Purple Rice Flavor)" as the target product.

[0066] S2 execution status: The average soil parameters collected from five representative plots in Gengzhuang Village were: pH 7.2, organic matter 1.8%, soil moisture (March average) 68%, and electrical conductivity 0.45 mS / cm.

[0067] Climate data: The average temperature from April to September over the past three years is 19.5℃, the average precipitation is 450mm, and the average daily sunshine intensity is 98 kLux.

[0068] Market conditions: The local corn purchase price is 0.9-1.0 yuan / kg, showing a downward trend.

[0069] The AI-powered model (trained based on 1.5 million data points of maize from North China) outputs planting decisions: Sowing time: April 5th to April 15th; Planting density: 4500 plants / acre; Fertilization plan: Compound fertilizer (N-P2O5-K2O=25-10-16), application rate 50kg / mu, applied twice, as basal fertilizer and topdressing at the jointing stage; Irrigation threshold: Start drip irrigation when soil moisture is below 55%, replenishing to 75% each time; Expected harvest: September 20th to September 30th; Farmers who followed this decision reduced fertilizer use by 18%, reduced irrigation by 2 times, and increased corn kernel bulk density by 5% compared to traditional planting methods.

[0070] S3 execution status: After harvesting in September 2025, select plump corn kernels with a moisture content of ≤14% as raw materials.

[0071] Start the fully automatic corn pancake production line with the default process parameters: grinding fineness 80 mesh, fermentation temperature 30℃, fermentation time 4h, frying temperature 180℃, and frying time 60 seconds.

[0072] Online inspection: Finished product moisture content 8.5~9.2%, thickness 1.7~1.9mm, uniform color. Pass rate 98.3%.

[0073] A total of 12.6 tons of corn pancakes with red dates and purple rice flavors were produced, packaged in 180g bags.

[0074] S4 execution status: The "One Mu Pancake" adoption program was launched on Douyin and WeChat mini programs. Consumers pay 199 yuan to adopt a plot of corn (about 0.1 mu) and can view the soil moisture and crop images in real time. A total of 342 consumers participated in the adoption, covering an area of ​​34.2 mu.

[0075] "Four Seasons Flavors" subscription service: launched a monthly subscription (49 yuan / month, including 6 bags of pancakes), with 520 subscribers in the first month and a repurchase rate of 32%.

[0076] The system automatically generates delivery orders daily, which are then shipped nationwide through partner logistics companies.

[0077] The yield of corn per mu (a Chinese unit of area, approximately 0.067 hectares) increased from 430 kg to 485 kg (thanks to precision irrigation and fertilization), and the corn raw materials were sold to processing plants at a price 5% higher than the market price. Calculations show that the average cost per mu in the planting stage decreased by approximately 18 yuan (due to water and fertilizer savings), and the average income per mu increased by approximately 15 yuan (due to increased yield and premium), resulting in a total increase in profit per mu of 33.16 yuan (consistent with the original data). If the profit sharing from the deep processing stage is included, the overall increase in farmers' average income per mu is even more significant.

[0078] Resource conservation: Saves 36 cubic meters of water and about 9 kilograms of fertilizer (pure nutrients) per acre.

[0079] Quality Improvement: The pass rate of finished corn pancakes has increased from 85% for manually produced ones to over 98%.

[0080] User feedback: 96% of adopting users are satisfied with the crop growth visualization function, and the repurchase rate of subscribing users is 32%.

[0081] The above embodiments demonstrate that the method of the present invention can effectively solve the technical problems of small farmers' blind selection of products, extensive planting, lack of deep processing, and single sales channels, and realize the digital closed loop of the corn industry chain.

[0082] The system upon which the above method relies includes the following functional modules: a data acquisition module configured to acquire e-commerce platform ranking data, soil parameters, climate parameters, and market data. This corresponds to the e-commerce data crawler in step S1 and the soil / climate / market data acquisition interface in step S2.

[0083] AI-powered product selection module: Connected to the data acquisition module, it is used to identify high-value-added corn deep-processed product categories. This corresponds to the AI ​​agent built on the Tencent Elements platform in step S1.

[0084] The AI ​​large-scale model planting decision module is connected to both the data acquisition module and the AI ​​intelligent agent product selection module, and is used to generate precise planting decision information. This corresponds to the pre-trained large-scale model invoked in step S2.

[0085] Automated production control module: Connected to the AI ​​intelligent agent product selection module, it has a built-in process parameter database and is used to drive the fully automated corn pancake production line. This corresponds to the industrial control system (PLC) and process parameter database in step S3.

[0086] The IoT sales management module communicates with the field sensor nodes in the data acquisition module, is configured to display crop growth sensor data, and process cornfield adoption orders and subscription delivery requests. This corresponds to the field sensor nodes, cloud platform, and terminal interface in step S4.

[0087] The above modules can be integrated on the same server or deployed in a distributed manner, and communicate via wired or wireless networks.

[0088] The IoT sensors, AI large-scale models, and fully automated production lines used in this invention are all mature technologies with controllable system costs, making them suitable for widespread adoption in rural areas. By applying transfer learning to the geographical and climatic data of the planting area, this method can be extended to other staple crops such as wheat and rice, demonstrating promising industrial application prospects.

[0089] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent R&D and production control of high value-added corn deep-processed products, characterized in that, The following steps are performed sequentially: Step S1: AI-driven product selection based on e-commerce data; Collect corn product ranking data from at least one e-commerce platform. The data includes product name, sales ranking, price range, and user review tags. Use an AI agent to call a large language model to analyze the ranking data, identify the corn deep-processed product category with the highest added value, and output the target product type identifier. Step S2: AI-driven planting decision-making based on multi-source data fusion; Collect soil parameters, climate parameters, and market data for the target planting area. Input the parameters and data, along with the target product type identifier output in step S1, into a pre-trained AI model. After calculation, the AI ​​model outputs precise planting decision information, which includes at least sowing time, planting density, fertilization plan, and irrigation plan. Step S3: Automated production control based on product category; In response to the target product type identifier output in step S1, the corresponding production process parameter set is retrieved from the process database; control commands are sent to the fully automated processing production line through the industrial control system to process the corn raw material into the target deep-processed product, and the quality inspection data of the production line is acquired in real time. Step S4: IoT adoption and subscription sales management; The system collects real-time sensor data on the corn growth process by deploying IoT sensor nodes in the field and pushes the sensor data to the terminal user interface; it establishes a virtual unit of "one acre of cornfield" in response to the user's adoption instruction, and records the corn variety and fertilizer type selected by the user; and it generates a periodic product delivery plan in response to the user's subscription instruction.

2. The method according to claim 1, characterized in that, In step S1, the construction and operation of the AI ​​agent are specifically as follows: an analysis command is generated through DeepSeek, an AI agent instance is created in the Tencent Yuanqi platform, a large language model is loaded to parse the ranking data of the top 100 corn products on the e-commerce platform, and "corn pancake" is output as the target product type identifier.

3. The method according to claim 1, characterized in that, The soil parameters in step S2 include: pH value, organic matter content, soil moisture and electrical conductivity; the climate parameters include: daily average temperature, precipitation, light intensity and wind speed; the market data includes: corn spot price and local purchase price fluctuation curve.

4. The method according to claim 3, characterized in that, The soil moisture was collected from a relative moisture content of 10% to 90% at a frequency of once every 30 minutes; the average daily temperature was collected from -10℃ to 40℃ at a frequency of once every 10 minutes.

5. The method according to claim 1, characterized in that, The fertilization plan generated in step S2 includes the ratio and application amount of nitrogen, phosphorus and potassium fertilizers, wherein the nitrogen-phosphorus-potassium ratio is in the form of N-P2O5-K2O, and the application amount ranges from 30kg to 80kg per mu; in the irrigation plan, an irrigation command is triggered when the soil moisture content is lower than 55% relative water content.

6. The method according to claim 1, characterized in that, The fully automated processing production line in step S3 is a fully automated corn pancake production line. The set of production process parameters includes: grinding fineness of 60 to 120 mesh, fermentation temperature of 25 to 35°C, fermentation time of 2 to 6 hours, frying temperature of 150 to 200°C, frying time of 30 to 90 seconds, and finished product moisture control threshold of ≤10%.

7. The method according to claim 1, characterized in that, The quality inspection data in step S3 includes: the thickness, diameter, color uniformity, and moisture content of the finished pancake. When any indicator exceeds the preset threshold, the industrial control system automatically adjusts the frying temperature or frying time.

8. The method according to claim 1, characterized in that, The IoT sensor nodes in step S4 include: a soil moisture sensor, a light intensity sensor, and a crop image acquisition camera. The image acquisition camera acquires field images with a resolution of 1080p or higher once a day.

9. The method according to claim 1, characterized in that, The adoption instructions in step S4 include: consumers selecting corn variety and fertilizer type through the interactive interface; the subscription instructions include selecting the subscription period and flavor preference tags.

10. A corn planting and marketing integrated management system based on artificial intelligence and the Internet of Things, used to execute the method according to any one of claims 1 to 9, characterized in that, include: The data acquisition module is configured to acquire e-commerce platform ranking data, soil parameters, climate parameters, and market data. The AI ​​intelligent agent product selection module is connected to the data acquisition module and is used to identify high value-added corn deep-processed product categories; The AI ​​large-scale model planting decision module is connected to the data acquisition module and the AI ​​intelligent agent product selection module respectively, and is used to generate precise planting decision information; An automated production control module, connected to the AI ​​intelligent agent product selection module, has a built-in process parameter database and is used to drive a fully automated corn pancake production line. The IoT sales management module is connected to the field sensor nodes in the data acquisition module and is configured to display crop growth sensing data and process cornfield adoption orders and subscription delivery requests.

Images