Intelligent platform system for agricultural science and technology communication learning

Abstract

The invention discloses an intelligent platform system for agricultural science and technology communication learning. The intelligent platform system comprises a technical support layer, a three-end application layer and a data center layer. The system has the advantages that the data center layer integrates a multi-source agricultural knowledge base, the problem of information fragmentation is thoroughly solved, and a user can obtain planting technology, policy and regulation and market quotation data in a one-stop mode; the intelligent matching module introduces a regional weight factor optimization collaborative filtering algorithm to realize accurate matching of technical contents and user portraits, so that the learning efficiency is remarkably improved; the learning path planning module is used for decomposing a crop growth cycle technology into customizable learning nodes based on a structured agricultural knowledge graph and helping a user to systematically master skills, and the data intelligent analysis module is used for automatically identifying regional technology demand hotspots and generating a visual report through a TF-IDF model and a clustering algorithm and providing a decision basis for agricultural technology popularization; a'consulting-learning-practice-feedback 'closed loop is formed, and the agricultural technology conversion efficiency and the farmland economic benefits are improved.

Description

TECHNICAL FIELD

[0001] The application relates to a wisdom platform system, in particular to a wisdom platform system for agricultural technology exchange and learning, and belongs to the technical field of agricultural technology informatization. BACKGROUND

[0002] With the development of society and the improvement of technology, more and more researches on agricultural production are carried out in China. In order to improve the productivity of agricultural production, more and more agricultural technology learning platforms have been produced. For example, the agricultural technology training method and system based on a cloud platform disclosed in the authorized announcement No. CN117350695B can effectively adapt to the differences between the trainees and the trainers and select the training mode, so that the maximum performance of the training resources can be fully utilized. In the mobile system for online exchange, learning and management of agricultural technology disclosed in the authorized announcement No. CN106777100B, the system comprises an expert online module, an agricultural technology guidance module, an agricultural technology knowledge module, a market information module, an agricultural technology exchange module and a notification module. Through the expert online module, the problems and answers can be viewed in time, and the problems can be switched between 'new', 'latest', 'essence' and 'to be solved'. Through the agricultural technology guidance module, the guidance records can be viewed. Through the agricultural technology knowledge module, the agricultural technology knowledge can be extracted from the agricultural technology network library.

[0003] Although the existing technology has many innovations and improvements related to agricultural technology learning platforms, the current agricultural technology learning platforms generally have the problem of information fragmentation. The technical data is scattered in different databases and government websites, and there is a lack of unified integration mechanism, which leads to the need for repeated search of multiple platforms by the user, and the low efficiency. The existing platforms mostly adopt a one-way information publishing mode, and an effective real-time interaction channel has not been established. Farmers cannot obtain expert guidance in time, the response cycle of technical consultation is too long, and the problem solving efficiency of agricultural production is seriously affected. In terms of personalized service, the platform usually provides general course content, and cannot dynamically adjust the recommendation strategy according to the planting type, regional characteristics and skill level of the user, which leads to the disconnection between the learning content and the actual demand. In addition, the user behavior data is not deeply mined, the platform only records the basic learning time, and does not construct a data analysis model to identify regional technology hotspots, which is difficult to support precise agricultural technology popularization and cannot meet the core needs of efficient dissemination and landing of agricultural technology. SUMMARY

[0004] The purpose of the present application is to provide a wisdom platform system for agricultural technology exchange and learning, which can integrate agricultural technology resources, realize precise knowledge pushing, promote real-time interaction between experts and farmers, and support agricultural technology popularization decision-making.

[0005] The application achieves the above-mentioned purpose by the following technical solutions: a smart platform system for agricultural science and technology exchange and learning, which comprises a technical support layer arranged at the front end, a three-end application layer and a data center layer arranged at the rear end, and the technical support layer is arranged in a cloud server;

[0006] The technical support layer comprises:

[0007] The micro-service architecture realizes service decoupling of each application end of the three-end application layer and supports horizontal expansion of the platform;

[0008] Real-time communication, real-time audio and video communication is adopted to support expert field guidance video connection;

[0009] Distributed search engine, realizing million-level agricultural technology document second-level retrieval with retrieval response time ≤ 500ms;

[0010] Blockchain storage alliance chain, ensuring that the technical achievements and consultation records are not tamperable and being used for intellectual property protection;

[0011] The three-end application layer comprises:

[0012] The user end is used for farmers, agricultural technicians and experts to browse knowledge, learn courses and interactively ask questions and answer questions;

[0013] The management end is used for agricultural technology resource auditing, user management and data visualization analysis;

[0014] The expert end is used for online question answering, course publishing and technical scheme customization of experts;

[0015] The data center layer is provided with a data center, which integrates an agricultural knowledge base, a user behavior database and an expert resource library, and the data center adopts MongoDB to store unstructured data and MySQL to store structured data;

[0016] The data center is integrated with:

[0017] An intelligent matching module, based on user portraits and technical labels, realizes technical content recommendation through an improved collaborative filtering algorithm;

[0018] A real-time interaction module, realizing double-mode message push of "agricultural technology question and answer square" and "one-on-one consultation" through WebSocket;

[0019] A learning path planning module, based on an agricultural technology knowledge graph, generates a personalized learning plan;

[0020] A data intelligent analysis module, through clustering analysis of user behavior data, generates a regional technical demand report.

[0021] As a further scheme of the present application: the agricultural knowledge base integrated by the data center includes planting / cultivation technology, policy and regulation and market data, and multi-source data is regularly synchronized through an ETL tool.

[0022] As a further scheme of the present application: the improved collaborative filtering algorithm of the intelligent matching module introduces a regional weight factor, and the similarity calculation formula is:

[0023]

[0024] wherein N(item) represents the number of users who have scored the item, and the regional weight is dynamically adjusted according to the matching degree of the user region and the technical application area.

[0025] As a further scheme of the present application: the consultation process of the real-time interaction module includes:

[0026] user question→system automatically matches relevant field experts→expert takes an order or system assigns an order→real-time audio and video dialogue→evaluation and archiving, and the consultation response time is within 15 minutes.

[0027] As a further scheme of the present application: the knowledge graph of the learning path planning module contains technical nodes in the whole growth cycle of crops, and the node relationship structure is:

[0028] crop type→technical subfield→specific technical point→solution→associated resource.

[0029] As a further scheme of the present application: the data intelligent analysis module performs the following steps:

[0030] S1, collect user consultation keywords and regional IP;

[0031] S2, quantize the technical demand heat by a TF-IDF model, and the formula of the TF-IDF model is:

[0032] TF(t, d) = the number of occurrences of word t in document d / the total number of words in document d;

[0033] IDF(t, D) = log (total number of documents / number of documents containing word t + 1);

[0034] TF-IDF(t, d) = TF(t, d) x IDF(t, d)

[0035] wherein t is the heat keyword, d is a single consultation record, and D is a set of all consultation documents in the region, and the technical demand heat is quantized by the model;

[0036] S3, when the keyword TF-IDF value exceeds the threshold value 0.8, it is determined as a high-priority demand and a visual report is generated.

[0037] As a further scheme of the present application: the assignment rule of the regional weight factor is:

[0038] When the user region and the technical application area are completely matched, the weight is increased by 30%;

[0039] When partially matched, the weight is increased by 15%;

[0040] When not matched, the weight is 1.

[0041] As a further scheme of the present application: the real-time interaction module supports the field guidance scene, and realizes mobile terminal video connection through WebRTC.

[0042] As a further scheme of the present application: the user data integration in the user behavior database is provided with a user registration and portrait generation process program, specifically including:

[0043] The user fills in the basic information -> the system generates an initial portrait;

[0044] The user completes the first learning / advisory -> the behavior data is written into the Redis cache;

[0045] A timing task is set every 24 hours to synchronize the cache data to the Hadoop cluster, and the user portrait is updated through MapReduce.

[0046] As a further scheme of the present application: the user data integration in the user behavior database is also provided with a technical recommendation process program, specifically including:

[0047] The user logs in -> the front end sends a request to the recommendation service;

[0048] The service obtains the user portrait ID, and reads the recent browsing records from Redis;

[0049] A similarity calculation function is called to obtain Top10 recommended technologies;

[0050] The results are rendered by the front end and displayed, and the user click behavior is fed back to the intelligent matching module in real time to realize technical content recommendation.

[0051] The present application has the advantages that: the intelligent platform system integrates multiple agricultural knowledge bases through the data center layer, completely solves the problem of information fragmentation, and users can obtain planting technology, policies and regulations and market data in one-stop;

[0052] The intelligent matching module introduces a regional weight factor to optimize the collaborative filtering algorithm, realizes accurate matching of technical content and user portrait, and significantly improves learning efficiency. The real-time interaction module combines WebSocket and WebRTC technologies to support "question and answer square" and "one-on-one video guidance" dual modes, ensuring instant communication between experts and farmers and significantly shortening technical consultation response time;

[0053] The learning path planning module divides the crop growth cycle technology into customizable learning nodes based on the structured agricultural knowledge graph, helping users systematically master skills. The data intelligent analysis module automatically identifies regional technology demand hotspots and generates visual reports through the TF-IDF model and clustering algorithm, providing decision-making basis for agricultural technology popularization;

[0054] The blockchain storage technology ensures the non-tamperability of technical achievements and consultation records, strengthens intellectual property protection, the micro-service architecture ensures the high scalability of the platform, the distributed search engine realizes second-level retrieval of massive data, and finally forms a "consultation-learning-practice-feedback" closed loop, significantly improving the efficiency of agricultural technology transformation and the economic benefits of farmland. BRIEF DESCRIPTION OF DRAWINGS

[0055] Fig. 1 The figure is a schematic diagram of the system architecture of the present application;

[0056] Fig. 2 The figure is a timing diagram of the real-time interaction module of the present application. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0058] Embodiment one

[0059] As shown in Figs. 1-2 A smart platform system for agricultural technology exchange and learning, the smart platform system includes a technical support layer arranged at the front end, a three-end application layer and a data center layer arranged at the back end, and the technical support layer is deployed on a cloud server.

[0060] The technical support layer includes:

[0061] Micro-service architecture, realizing service decoupling of each application end of the three-end application layer, supporting platform horizontal expansion;

[0062] Real-time communication, using real-time audio and video communication, supporting expert field guidance video connection;

[0063] A distributed search engine is implemented to realize the second-level retrieval of ten million agricultural technology documents, and the retrieval response time is less than or equal to 500 ms.

[0064] A blockchain storage alliance chain is used to ensure that the technical achievements and consultation records are not tampered with, and is used for intellectual property protection.

[0065] The three-end application layer comprises:

[0066] The user end is used for farmers, agricultural technicians and experts to browse knowledge, learn courses and interactively ask and answer questions.

[0067] The management end is used for agricultural technology resource auditing, user management and data visualization analysis.

[0068] The expert end is used for online question answering, course publishing and technical scheme customization.

[0069] The data center layer is provided with a data center, which integrates an agricultural knowledge base, a user behavior database and an expert resource library. The data center uses MongoDB to store unstructured data and MySQL to store structured data.

[0070] The data center is integrated with:

[0071] The intelligent matching module is based on user portraits and technical tags, and realizes technical content recommendation through an improved collaborative filtering algorithm.

[0072] The real-time interaction module realizes the double-mode message push of the "agricultural technology question and answer square" and "one-on-one consultation" through WebSocket.

[0073] The learning path planning module generates a personalized learning plan based on the agricultural technology knowledge graph.

[0074] The data intelligent analysis module generates a regional technical demand report by clustering user behavior data.

[0075] Embodiment two

[0076] In addition to including all the technical features in embodiment one, this embodiment also includes:

[0077] The agricultural knowledge base integrated by the data center includes planting / cultivation technology, policy regulations and market trend data, and periodically synchronizes multi-source data through an ETL tool. The data integration rate is more than 95%, solving the problem of information fragmentation.

[0078] Further, the improved collaborative filtering algorithm of the intelligent matching module introduces a regional weight factor, and its similarity calculation formula is:

[0079]

[0080] Wherein, N(item) represents the number of users who have scored for item, and the regional weight is dynamically adjusted according to the matching degree of user region and technology applicable area.

[0081] Further, the consultation process of the real-time interaction module includes:

[0082] User asks questions -> system automatically matches relevant field experts -> expert takes orders or system assigns orders -> real-time audio and video dialogue -> evaluation and archiving, and the consultation response time is within 15 minutes.

[0083] Further, the knowledge graph of the learning path planning module contains technical nodes in the whole growth cycle of crops, with more than 200,000 nodes and more than 500,000 node relationship edges, and the node relationship structure is:

[0084] Crop type -> technical sub-field -> specific technical point -> solution -> associated resource.

[0085] Further, the data intelligent analysis module performs the following steps:

[0086] S1, collect user consultation keywords and regional IP;

[0087] S2, quantify technical demand heat by TF-IDF model, and the formula of TF-IDF model is:

[0088] TF(t,d)=the number of occurrences of word t in document d / total number of words in document d;

[0089] IDF(t,D)=log(number of documents / number of documents containing word t+1);

[0090] TF-IDF(t,d)=TF(t,d)×IDF(t,d)

[0091] Wherein, t is the heat keyword, d is a single consultation record, and D is the set of all consultation documents in the region. By quantifying the technical demand heat through the model, for example, when the TF-IDF value of "wheat rust" exceeds the threshold value of 0.8, it is determined as a high-priority demand;

[0092] S3, when the keyword TF-IDF value exceeds the threshold value of 0.8, it is determined as a high-priority demand and a visual report is generated.

[0093] Further, the assignment rules of the regional weight factor are:

[0094] When the user region and the technology applicable area are completely matched, the weight is increased by 30%;

[0095] When partially matched, the weight is increased by 15%;

[0096] When not matched, the weight is 1.

[0097] Furthermore, the real-time interaction module supports field guidance scenarios, enabling mobile video connections via WebRTC with a latency of ≤300ms.

[0098] Example 3

[0099] A smart platform system for agricultural technology exchange and learning adopts a front-end and back-end separation architecture. The front-end is developed based on Vue.js, and the back-end uses the Java Spring Boot framework, deployed on a cloud server (Alibaba Cloud ECS). The data center uses Kafka for real-time data synchronization and Redis to cache frequently accessed data, improving response speed. Specifically, it includes:

[0100] 1. Design of core modules of the intelligent platform system:

[0101] User profile building module: Collects user registration information (planting type, region, years of experience), learning behavior (course completion rate, test scores), and consultation records, and maps them to database tables through DjangoORM. Example table structure:

[0102] classUserProfile(models.Model):

[0103] user=models.OneToOneField(User,on_delete=models.CASCADE)

[0104] planting_type=models.CharField(max_length=50) # Planting type (rice, corn, etc.)

[0105] region = models.CharField(max_length = 50) # The region to which the model belongs

[0106] experience_years = models.IntegerField() # Planting years

[0107] #Behavioral data fields (learning duration, test scores, etc.)

[0108] Recommendation Algorithm Module: Implements an improved collaborative filtering algorithm based on TensorFlow. Key code snippets:

[0109] defcalculate_similarity(item_i,item_j,region_weight):

[0110] # Calculate the number of users who gave the same rating to items i and j

[0111] common_users=set(item_i.users)&set(item_j.users)

[0112] #Introduce regional weights (e.g., weighting when the user's location matches the applicable technology region)

[0113] ifitem_i.region==user_region:

[0114] weight=region_weight

[0115] else:

[0116] weight=1

[0117] # Calculate cosine similarity

[0118] similarity=len(common_users) / math.sqrt(len(item_i.users)*len(item_j.users))*weight

[0119] returnsimilarity

[0120] 2. Intelligent Platform System Program Flow:

[0121] User registration and profile generation process:

[0122] User fills in basic information → System generates initial profile;

[0123] User completes initial learning / consultation → Behavioral data is written to Redis cache;

[0124] A scheduled task (every 24 hours) synchronizes cached data to the Hadoop cluster and updates user profiles via MapReduce.

[0125] Technology recommendation process:

[0126] User login → Frontend sends a request to the recommendation service;

[0127] The service retrieves the user profile ID and reads the recent browsing history from Redis.

[0128] Call the similarity calculation function to obtain the Top 10 recommendation technology;

[0129] The results are displayed after being rendered by the front end, and user click behavior is fed back to the algorithm module in real time.

[0130] 3. Example of physical steps (regional technology requirements analysis):

[0131] Step 1: Collect consultation data from users in a certain region within one month (physical data: consultation text, timestamp, geographic IP).

[0132] Step 2: Extract keywords (such as "wheat rust" and "fertilizer application rate") using NLP technology;

[0133] Step 3: Use the TF-IDF algorithm to calculate keyword weights and filter high-frequency technology requirements;

[0134] Step 4: Generate a visualization report (such as a word cloud or trend curve) and push it to the local agricultural technology station.

[0135] Mathematical model explanation (TF-IDF formula):

[0136] TF(t,d) = Number of occurrences of word t in document d / Total number of words in document d

[0137] IDF(t,D) = log(total number of documents / number of documents containing word t + 1)

[0138] TF-IDF(t,d) = TF(t,d) × IDF(t,d)

[0139] Where t represents a keyword (e.g., "rust disease"), d represents a single consultation record, and D represents the set of all consultation documents for the region. This model can quantify the intensity of technology demand; for example, when the TF-IDF value of "wheat rust disease" exceeds the threshold of 0.8, it is determined to be a high-priority demand.

[0140] It should be noted that this platform is deployed in a cloud service environment and requires no special hardware modifications.

[0141] Server configuration: 3 servers with 4 cores and 8GB of memory each (application server), 2 servers with 8 cores and 16GB of memory each (database server).

[0142] Network requirements: Bandwidth of 100Mbps or higher, with support for WebSocket persistent connections;

[0143] If offline training equipment is required, only a regular projector and a computer with internet access are needed; the system is adapted via a web interface.

[0144] Working principle: Users register through the user terminal and fill in basic information such as planting type, region and years of experience. The system generates an initial user profile based on this. When users start browsing technical information or initiating inquiries, behavioral data is written to the Redis cache in real time. A scheduled task synchronizes the cached data to the Hadoop cluster every 24 hours and updates the user profile attributes through MapReduce calculation.

[0145] After the user logs in, the recommendation process of the intelligent matching module is triggered, the front end sends a request to the recommendation service, the server obtains the user portrait ID and extracts the recent browsing records from Redis, and calls the improved collaborative filtering algorithm to calculate the similarity: the algorithm first identifies the technical projects of the user's historical interaction, retrieves other technical projects with common scoring users, introduces a regional weight factor to weight the similarity, and finally returns the Top10 recommended technology list to the front end for display;

[0146] When the user initiates a technical consultation, the real-time interaction module starts a double-mode process, the question request is pushed to the expert end through WebSocket, the system automatically matches the field experts according to the technical tags and starts the order grabbing mechanism, the expert establishes a low-delay audio and video connection through WebRTC after accepting the order to realize real-time guidance in the field, and the conversation record is archived to the blockchain storage alliance chain after evaluation to ensure that it cannot be tampered with; The learning path planning module synchronously responds to user behavior, locates the user's current knowledge weak points based on the pre-built agricultural knowledge graph, and dynamically generates a learning plan including sub-fields such as seedling raising and pest control;

[0147] The test scores generated by the user learning and the consultation keywords are captured by the data intelligent analysis module, combined with the regional IP information, and quantified through the TF-IDF model to quantify the technical demand heat (calculate the product of keyword frequency and inverse document frequency in regional consultation documents), when the TF-IDF value of a specific technical demand exceeds the preset threshold, the early warning mechanism is automatically triggered, and the clustering analysis module aggregates the hot spot demand to generate a visual regional report and push it to the management end. The agricultural technician of the management end deploys offline promotion resources according to the report, and the farmer's practice effect is fed back through the user end to enter the behavior data flow again, forming a "consultation-learning-practice-feedback" closed loop to drive the recommendation algorithm to continuously optimize.

[0148] The entire system operates through the micro-service architecture coordination module, the distributed search engine handles massive data queries, and the cloud server ensures elastic expansion of computing resources, ultimately realizing efficient dissemination and precise landing of agricultural technology.

[0149] It is apparent to those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and that the present application can be implemented in other concrete forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, and the scope of the present application is defined by the appended claims rather than the foregoing description, and it is intended to encompass all changes falling within the meaning and scope of the equivalent elements of the claims. Any reference signs in the claims should not be considered as limiting the claims involved.

[0150] Furthermore, it should be understood that although the specification is described in terms of embodiments, not every embodiment includes every feature described. The specification can include implicit combinations of explicitly mentioned features and / or explicit combinations of implicitely mentioned features. Each embodiment depends on the explicit combinations of features and / or the implicit combinations of features made specifically within that embodiment, and each such embodiment can be combined with every other such embodiment to create further embodiments.

Claims

1. A smart platform system for agricultural technology exchange learning, characterized in that: The wisdom platform system comprises a technical support layer arranged at a front end, a three-terminal application layer and a data center layer arranged at a rear end, and the technical support layer is arranged in a cloud server; The technical support layer comprises: a micro-service architecture, which realizes service decoupling of each application terminal of the three-terminal application layer and supports horizontal expansion of the platform; real-time communication, which adopts real-time audio and video communication and supports video connection for field guidance by experts; a distributed search engine, which realizes million-level agricultural technology document second-level search with a search response time of ≤500 ms; a blockchain storage alliance chain, which ensures that technical achievements and consultation records are not tamperable and is used for intellectual property protection; The three-terminal application layer comprises: a user terminal, which is used for knowledge browsing, course learning and interactive question and answer by farmers, agricultural technicians and experts; a management terminal, which is used for agricultural technology resource auditing, user management and data visualization analysis; an expert terminal, which is used for online question answering, course publishing and technical scheme customization by experts; The data center layer is provided with a data center, which integrates an agricultural knowledge base, a user behavior database and an expert resource database, and the data center uses MongoDB to store unstructured data and MySQL to store structured data; The data center is integrated with: an intelligent matching module, which realizes technical content recommendation based on user portraits and technical tags through an improved collaborative filtering algorithm; a real-time interaction module, which realizes double-mode message push of "agricultural technology question and answer square" and "one-on-one consultation" through WebSocket; a learning path planning module, which generates personalized learning plans based on an agricultural technology knowledge graph; a data intelligent analysis module, which generates regional technical demand reports by clustering user behavior data. 2.The smart platform system for agricultural technology exchange learning according to claim 1, wherein: The agricultural knowledge base integrated by the data center comprises planting / cultivation technology, policy regulations and market trend data, and is regularly synchronized with multi-source data through an ETL tool. 3.The smart platform system for agricultural technology exchange learning according to claim 1, wherein: The improved collaborative filtering algorithm of the intelligent matching module introduces a regional weight factor, and the similarity calculation formula is: ； Wherein, N(item) represents the number of users who have scored item, and the regional weight is dynamically adjusted according to the matching degree of user region and technical application area.

4. The smart platform system for agricultural technology exchange learning according to claim 1, wherein: The consultation process of the real-time interaction module comprises: user asking questions→system automatically matching relevant field experts→expert taking orders or system assigning orders→real-time audio and video dialogue→evaluation and archiving, and the consultation response time is within 15 minutes. 5.The smart platform system for agricultural technology exchange learning of claim 1, wherein: The knowledge graph of the learning path planning module contains technical nodes in the whole growth cycle of crops, and the node relationship structure is: crop type→technical sub-field→specific technical point→solution→associated resource. 6.The smart platform system for agricultural technology exchange learning according to claim 1, wherein: The data intelligent analysis module performs the following steps: S1, collect user consultation keywords and regional IP; S2, quantify technical demand heat by a TF-IDF model, and the formula of the TF-IDF model is: TF(t,d)=the number of occurrences of word t in document d / total number of words in document d; IDF(t,D)=log(number of all documents / number of documents containing word t+1); TF-IDF(t,d)=TF(t,d)×IDF(t,d) Wherein, t is a heat keyword, d is a single consultation record, and D is a set of all consultation documents in the region, and the technical demand heat is quantified by the model; S3, when the keyword TF-IDF value exceeds the threshold value 0.8, it is determined that the high priority demand and generates a visual report. 7.The smart platform system for agricultural technology exchange learning according to claim 3, wherein: The assignment rule of the regional weight factor is: When the user region and the technology application area are completely matched, the weight is increased by 30%; When partially matched, the weight is increased by 15%; When not matched, the weight is 1. 8.The smart platform system for agricultural technology exchange learning of claim 1, wherein: The real-time interaction module supports field guidance scenarios and realizes mobile video connection through WebRTC. 9.The smart platform system for agricultural technology exchange learning of claim 1, wherein: The user data integration in the user behavior database is provided with a user registration and portrait generation process program, specifically including: User fills in basic information → system generates initial portrait; User completes the first learning / advisory → behavior data is written into Redis cache; Set a timing task every 24 hours to synchronize the cache data to the Hadoop cluster, and update the user portrait through MapReduce. 10.The smart platform system for agricultural technology exchange learning of claim 1, wherein: The user data integration in the user behavior database is also provided with a technology recommendation process program, specifically including: User login → front-end sends request to recommendation service; Service acquires user portrait ID, reads recent browsing records from Redis; Call similarity calculation function to get Top10 recommended technologies; The results are rendered by the front-end and displayed, and the user's click behavior is fed back to the intelligent matching module in real time to realize technology content recommendation.

Images