A modern agricultural digital authentication large model

Abstract

The embodiment of the present application provides a kind of modern agricultural digital authentication big model, belong to agricultural big data and artificial intelligence field, agricultural data authentication method includes: obtaining the data information of target farm, the data information includes crop information, farm information and production information;According to the crop information, farm information and production information, construct block chain, the block chain includes main chain, logistics subchain, quality inspection subchain and production subchain;The block chain is stored in block chain, and the authentication certificate of each crop in the target farm is obtained, and the block chain storage includes production data storage, environmental data storage, management data storage, processing data storage and tactile data storage;According to the authentication certificate, the crop in the target farm is tracked.The method realizes the accurate estimation of agricultural yield and value and optimizes agricultural planting scheme.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural big data and artificial intelligence, in particular to a modern agricultural digital authentication large model. BACKGROUND

[0002] The current agricultural product authentication method mainly relies on manual operation, such as manual sampling, manual recording and manual auditing. This mode leads to high authentication cost and low efficiency, which directly affects the efficiency of agricultural product circulation and the loss rate of fresh products. At the same time, the lack of data credibility makes it difficult for financial institutions to provide loans to farmers based on authentication data, and the pass rate of farmers' mortgage loans is low. High-quality agricultural products are also difficult to achieve premium prices due to the lack of reliable authentication. The existing Internet of Things data cannot be combined with intelligent analysis in depth, and the traditional authentication method relies on manual operation, which is time-consuming and laborious and prone to errors, increasing the production and management costs. Agricultural projects are generally considered to be high-risk, and financial institutions have limited support for them. Consumers lack confidence in the quality of agricultural products, making it difficult to achieve high quality and high price, limiting brand building and market expansion. SUMMARY

[0003] The purpose of the embodiment of the present application is to provide a modern agricultural digital authentication large model, which realizes the accurate estimation of agricultural yield and value and optimizes the agricultural planting scheme.

[0004] In order to achieve the above-mentioned purpose, the present application provides a method for agricultural data authentication, which comprises: acquiring data information of a target farm, the data information comprising crop information, farm information and production information; constructing a blockchain according to the crop information, the farm information and the production information, the blockchain comprising a main chain, a logistics sub-chain, a quality inspection sub-chain and a production sub-chain; performing blockchain storage on the blockchain to obtain authentication certificates of each crop in the target farm, the blockchain storage comprising production data storage, environmental data storage, management data storage, processing data storage and tactile data storage; and tracking the crops in the target farm according to the authentication certificates.

[0005] Optionally, the main chain comprises key data, the key data comprising at least one of trigger data of field water holding capacity and feature hash value of pest image; the logistics sub-chain comprises transportation track and storage environment; the quality inspection sub-chain comprises pesticide detection result and quality rating; and the production sub-chain comprises farm operation record and seedling raising information.

[0006] Optionally, the blockchain notarization of the block chain obtains the authentication certificate of each crop in the target farm, comprising: determining the growth state hash value of the crop according to the main chain and the production subchain; determining the environmental parameter according to the main chain, the logistics subchain and the production subchain, and generating a warning event hash value when the environmental parameter is abnormal; determining the operation trajectory and the correlation hash value of the input according to the logistics subchain, the quality inspection subchain and the production subchain; determining the process chain hash value of each processing batch according to the logistics subchain, the quality inspection subchain and the production subchain; determining the transportation process hash value according to the logistics subchain; and storing the growth state hash value, the warning event hash value, the correlation hash value, the process chain hash value and the transportation process hash value in the chain to obtain the authentication certificate of each crop in the target farm.

[0007] Optionally, the authentication certificate includes crop variety name, grower information, planting time, growth process data, test results, authentication conclusion and risk warning label.

[0008] Optionally, the method further comprises: generating a biometric hash value according to the image information in the data information; constructing a triple according to the biometric hash value, the public key of the authentication certificate and the wallet address; and storing the triple in the blockchain underlying platform for forming a digital identity anchor point of the crop.

[0009] Optionally, the method further comprises preprocessing the data information, comprising: performing cost rule library screening, edge computing and standard formatting processing on the crop information, farm information and production information to obtain qualified data; and performing cleaning and normalization processing on the continuous data in the qualified data to obtain preprocessed data.

[0010] Optionally, the method further comprises: optimizing the crop information, farm information and production information through a deep learning model; the network neck of the deep learning model is provided with an SE attention module for recalibrating channel weights; and the channel shuffling operation of the deep learning model comprises grouping shuffling and point-by-point convolution.

[0011] Optionally, a crop category vector is set for the crop image of the target farm; the convolution kernel weight in the deep learning model is dynamically adjusted through a multi-head self-attention mechanism; the attention degree of the crop in the deep learning model is determined according to the crop category vector and the convolution kernel weight for variety identification and / or pest and disease detection; a path aggregation network is used for adaptive sharpening preprocessing of the crop image for identifying disease spot features; and deep learning model is used to determine soil and meteorological time series data for predicting growth trend and maturity, wherein the meteorological time series data is integrated into the crop growth cycle as a feature node.

[0012] In another aspect, the present application also provides an agricultural data authentication device, comprising: an acquisition module configured to acquire data information of a target farm, the data information comprising crop information, farm information and production information; a first processing module configured to construct a blockchain according to the crop information, the farm information and the production information, the blockchain comprising a main chain, a logistics sub-chain, a quality inspection sub-chain and a production sub-chain; and a second processing module configured to perform blockchain storage on the blockchain to obtain an authentication certificate of each crop in the target farm, the blockchain storage comprising production data storage, environment data storage, management data storage, processing data storage and tactile data storage, and the data of the crops in the target farm is tracked according to the authentication certificate.

[0013] Optionally, the blockchain storage on the blockchain to obtain the authentication certificate of each crop in the target farm comprises: determining a growth state hash value of the crop according to the main chain and the production sub-chain; determining an environment parameter according to the main chain, the logistics sub-chain and the production sub-chain, and generating an early warning event hash value when the environment parameter is abnormal; determining an operation trajectory and an associated hash value of an input according to the logistics sub-chain, the quality inspection sub-chain and the production sub-chain; determining a process chain hash value of each processing batch according to the logistics sub-chain, the quality inspection sub-chain and the production sub-chain; determining a transportation process hash value according to the logistics sub-chain; and performing on-chain storage of the growth state hash value, the early warning event hash value, the associated hash value, the process chain hash value and the transportation process hash value to obtain the authentication certificate of each crop in the target farm.

[0014] Optionally, the device further comprises a third processing module configured to generate a biometric hash value according to image information in the data information; construct a triple according to the biometric hash value, a public key of the authentication certificate and a wallet address; and perform on-chain storage of the triple on a blockchain underlying platform to form a digital identity anchor point of the crop.

[0015] In another aspect, the present application also provides a machine-readable storage medium having instructions stored thereon, the instructions, when executed by a processor, causing the processor to be configured to perform the above-mentioned agricultural data authentication method.

[0016] The method for agricultural data authentication comprises the following steps: obtaining data information of a target farm, wherein the data information comprises crop information, farm information and production information; constructing a block chain according to the crop information, the farm information and the production information, wherein the block chain comprises a main chain, a logistics sub-chain, a quality inspection sub-chain and a production sub-chain; performing block chain storage on the block chain to obtain an authentication certificate of each crop in the target farm, wherein the block chain storage comprises production data storage, environment data storage, management data storage, processing data storage and tactile data storage; and performing data tracking on the crops in the target farm according to the authentication certificate. The method realizes real-time collection of full-cycle agricultural data, dynamic value calculation and risk assessment based on a digital agricultural model to generate an authentication certificate, provides a quantifiable credit basis for a financial institution, converts an agricultural production process into a standard digital asset recognizable by the financial institution, provides a right confirmation basis for financial tools such as credit, futures and trust, and realizes accurate estimation of agricultural yield and value.

[0017] Other features and advantages of the embodiments of the present application will be described in detail in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings are included to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used together with the following detailed description to explain the embodiments of the present application, but do not constitute a limitation of the embodiments of the present application. In the drawings:

[0019] Figure 1 is a flowchart of a method for agricultural data authentication of the present application;

[0020] Figure 2 is a flowchart of data collection of the present application;

[0021] Figure 3 is a specific embodiment schematic diagram of the present application;

[0022] Figure 4 is another specific embodiment schematic diagram of the present application;

[0023] Figure 5 is a smart transaction review flowchart of the present application;

[0024] Figure 6 is a flowchart of embodiment two of the present application;

[0025] Figure 7 is a schematic diagram of a device for agricultural data authentication of the present application.

[0026] Explanation of reference signs

[0027] 100 - device for agricultural data authentication; 200 - acquisition module; 300 - first processing module;

[0028] 400 - second processing module. DETAILED DESCRIPTION

[0029] The specific embodiments of the present application are described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to illustrate and explain the present application, and are not intended to limit the present application.

[0030] It should be noted that the acquisition, transmission, storage, use, processing, etc. of data in the technical solution of the present application comply with the relevant provisions of laws and regulations. In the embodiments of the present application, some industry existing solutions such as software, components, models, etc. may be mentioned, which should be considered as exemplary, and the purpose is only to illustrate the feasibility of the implementation of the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0031] Embodiment one

[0032] Figure 1 is a flowchart of a method for agricultural data authentication, as Figure 1 shown, the method for agricultural data authentication comprises: step S101, acquiring data information of a target farm, the data information comprising crop information, farm information and production information.

[0033] Specifically, as Figure 2As shown, the present application can be constructed by various devices such as cameras, soil sensors, weather stations, mobile / PC terminals and data interfaces, remote sensing and unmanned aerial vehicles to form a comprehensive data acquisition network. The camera captures real-time images of crop growth in the field (including micro growth characteristics such as plant height, leaf color, fruit shape, etc.), records images of personnel and equipment activities in the field, and appearance state information of agricultural facilities such as greenhouses and irrigation systems, providing visual image basis for agricultural production; the soil sensor adopts a layered deployment method (such as one layer per 10 cm), continuously collects soil temperature and humidity, pH, electrical conductivity, and nutrient content data such as nitrogen, phosphorus, and potassium, and real-time monitors soil fertility and environmental dynamic changes. The weather station acquires meteorological parameters such as air temperature and humidity, light intensity, wind speed and direction, precipitation, and pressure at a predetermined sampling frequency, laying a foundation for agricultural production environment analysis; the mobile / PC terminal and data interface support manual input of sowing time, fertilizer amount, plant protection measures, and other agricultural operation information, as well as agricultural product sales intention data, effectively supplementing the automatically collected data that are not covered; remote sensing uses satellite platforms to carry multi-spectral, microwave, and other sensors to dynamically collect macro data such as crop growth, soil moisture, and vegetation coverage in a large range of farmland at a threshold time period; unmanned aerial vehicles carry high-resolution imaging, thermal infrared, and other sensors to obtain high-precision, real-time micro data such as crop diseases and pests, water and fertilizer conditions, plant height, and leaf area in local farmland plots.

[0034] Various devices realize data interaction through a communication network, in which the camera-acquired video stream is compressed by an edge computing node, soil sensor data is removed of high-frequency noise through sliding window polynomial fitting, weather station data is reduced in transmission volume through a time series compression algorithm, and mobile terminal data is synchronized to the cloud in real time through an encryption protocol. Key agricultural operations (such as fertilizer amount = soil EC value x 0.8 + base value) and seedling information (seed variety, batch number) are stored on the blockchain, and the smart contract automatically verifies data compliance, promoting the standardization of the agricultural production process. Enterprises can replicate high-quality production modes based on on-chain data, and farmers can work according to standardized processes, effectively solving the problem of traditional agricultural experience dominance and disorder. At the same time, the brand trust is strengthened through the tamper-proof nature of the blockchain, helping the development of agricultural product branding, and consumers can view the whole cycle of data records from sowing to picking, including daily soil temperature and humidity curves, weekly aerial images of growth taken by unmanned aerial vehicles, and manually inputted fertilizer and pesticide application records, forming a complete data chain from the field to the table, providing comprehensive and detailed data support for precise decision-making and scientific management of smart agriculture, and providing a reliable underlying data foundation for agricultural digital certification and financial services.

[0035] The method further includes: generating a biometric hash value based on the image information in the data information; constructing a triple based on the biometric hash value, the public key of the authentication certificate, and the wallet address; and storing the triple on the blockchain underlying platform to form a digital identity anchor for crops.

[0036] Specifically, during key stages of crop growth, high-resolution cameras are used to capture images of crops, ensuring clear capture of their biological characteristics, such as leaf texture and fruit shape. The captured images are first converted to grayscale, reducing data volume while preserving key feature information and simplifying subsequent calculations. Next, a Gaussian filtering algorithm is used to denoise the images. This algorithm constructs a Gaussian kernel function and applies a weighted average to each pixel in the image, effectively removing noise points caused by factors such as lighting variations and sensor noise, making the image smoother and clearer, providing a high-quality data foundation for subsequent feature extraction.

[0037] The Scale Invariant Feature Transform (SIFT) algorithm was used to extract feature points from crops. SIFT constructs a Difference of Gaussian (DoG) scale space to detect extreme points at different scales. These extreme points are highly invariant to image rotation, scaling, and illumination changes. For crop images, SIFT can accurately locate feature points such as leaf edges, vein intersections, and fruit contours. For example, when identifying apple varieties, it can accurately extract feature points such as the distribution of spots on the apple peel and the shape of the fruit stalk. The extracted feature points are then described as 128-dimensional feature vectors. These vectors contain gradient and orientation information from neighboring pixels, comprehensively and uniquely representing the biological characteristics of each feature point.

[0038] The generated feature vector is input into the Secure Hash Algorithm (SHA-256). The SHA-256 algorithm performs multiple rounds of complex operations on the input data, including logical operations and shift operations, ultimately outputting a 256-bit hash value. This hash value is a highly condensed and unique digital representation of the crop's biological characteristics. Even if there are slight differences in the crop images during the acquisition process (slightly different angles), as long as the essential biological characteristics remain unchanged, the generated hash values ​​will be highly consistent. However, if the biological characteristics undergo substantial changes (such as significant changes in leaf morphology due to pests or diseases), the hash values ​​will be drastically different, thus providing a reliable basis for subsequent identity verification and traceability.

[0039] Taking tomatoes as an example, data collection and hash value generation involve capturing images of representative tomato plants from multiple angles during the early stages of tomato growth and ripening, using professional image acquisition equipment and following standard shooting procedures. After grayscale conversion and Gaussian filtering, the acquired images are processed to extract biometric points, which are then converted into 128-dimensional feature vectors. These vectors are then used to generate biometric hash values ​​using the SHA-256 algorithm. For example, for a specific tomato plant, the generated biometric hash value is a54f321c9876b543 (256 bits).

[0040] The planting base had previously obtained agricultural product certification from an authoritative agricultural certification body. The certification certificate contained public key information, with the public key being 0x123abcdef456 (the specific public key value). Simultaneously, the base has its own wallet address on the underlying blockchain platform, with the wallet address being 0xf123456789abcdef (blockchain wallet address format).

[0041] The construction and on-chain notarization of the triple involves combining the generated biometric hash value, the authentication certificate public key, and the wallet address to construct a triple, namely a54f321c9876b543, 0x123abcdef456, and 0xf123456789abcdef. This triple data is then encapsulated into a transaction request conforming to the blockchain data format via the API interface provided by the underlying blockchain platform and sent to the blockchain network. Nodes in the blockchain network verify the transaction request, including checking the data format and signature. Upon successful verification, the triple data is packaged into a new blockchain block and reaches consensus across the entire blockchain network through consensus mechanisms (Proof-of-Stake (PoS), Practical Byzantine Fault Tolerance (PBFT), etc.), ultimately completing the on-chain notarization. Afterward, this triple data is permanently stored in the blockchain, immutable, becoming the digital identity anchor for this tomato throughout its entire lifecycle.

[0042] Biometric hash values, serving as unique digital identifiers for crops, possess high uniqueness and stability, making them difficult to forge or tamper with. The decentralized storage and consensus mechanism of blockchain ensure that the triplet data stored on the chain is distributed throughout the network, meaning that the integrity of the data cannot be affected by the failure or malicious tampering of any single node. Tampering with data on the blockchain requires control of more than half of the nodes, which is virtually impossible in practical applications, thus guaranteeing the security and reliability of digital identity information for crops.

[0043] When traceability or identity verification of agricultural products is required, it is only necessary to collect the image information of the crop again, generate a biometric hash value according to the same process, and then query the corresponding triplet data on the blockchain. If the hash value matches and the public key of the authentication certificate is consistent with the information of the relevant certification authority, the source, growth process, certification status, and other detailed information of the crop can be accurately traced, achieving precise traceability. At the same time, based on the openness and transparency of the blockchain, consumers, regulatory authorities, and other relevant parties can legally query this information, enhancing trust in the quality and safety of agricultural products. For example, in agricultural product markets, consumers can scan the traceability code on the agricultural product packaging to obtain the digital identity anchor information of the agricultural product stored on the blockchain, including changes in its biometric characteristics during its growth process (linked to image features at different stages through hash values), certification status, etc., allowing for confident purchase.

[0044] The application of this technology encourages agricultural production enterprises to pay more attention to the standardization and normalization of the production process, ensuring that the biological characteristics of crops can be accurately collected and identified, meeting certification requirements. At the same time, through blockchain-based evidence storage, regulatory authorities can more efficiently supervise agricultural production and certification processes, combat counterfeit and substandard agricultural products, maintain market order, and promote the standardization and high-quality development of the entire agricultural industry.

[0045] The method also includes preprocessing the data information, including: filtering the crop information, farm information and production information using a cost rule base, edge computing and standard formatting to obtain qualified data; cleaning and normalizing the continuous data in the qualified data to obtain preprocessed data.

[0046] Specifically, the data processing and storage layer standardizes and formats the data that has been initially processed by edge computing (such as using a unified UTC timestamp and JSON-LD data format) and stores it for a long time, in preparation for subsequent analysis and blockchain on-chaining, and retains a data archive of the entire agricultural production cycle, serving as the data hub for the precise operation of smart agriculture in smart agriculture scenarios. On the one hand, a three-level data cleaning strategy is used (such as the physical layer using an anomaly rule library to filter outliers such as vegetable data with prices > 50 yuan / kg in real time; the statistical layer using the 3σ principle combined with the isolated forest algorithm to remove outliers; and the semantic layer using agricultural economic ontology mapping tools to transform planning text into value factors) to clean and integrate multi-source data such as soil sensors and weather stations. After standardization and formatting, it provides a high-quality data source for intelligent analysis, supporting pest and disease probability analysis based on the improved YOLOv9 model and crop growth trend prediction based on the Transformer+LSTM hybrid network. It also helps to make production decisions more precise, such as irrigation (e.g., generating irrigation instructions within 100ms when soil moisture < 60% field capacity) and fertilization (nitrogen fertilizer application amount = soil EC value × 0.8 + baseline value), and promotes the transformation from experience-based planting to data-driven planting. On the other hand, by storing agricultural production data throughout the entire lifecycle through a distributed file system, a production database including soil spectra and growth images is constructed. This provides a data foundation for training regional customized irrigation strategy models based on GNNs and optimizing novel crop recognition models based on meta-learning + GANs. Combined with edge computing local processing capabilities, continuous data collection and storage are achieved in extreme environments, ensuring the stable operation of smart agriculture and improving production efficiency and quality.

[0047] This application uses an IoT gateway as the core hub for data aggregation and protocol conversion. It adopts a multi-channel design (supporting protocols such as Modbus, MQTT, and LoRa) to achieve real-time aggregation of data from various types of sensors and devices, such as soil sensors, weather stations, and 4K cameras. A dynamic load balancing algorithm ensures stable data transmission to edge computing nodes. The edge computing nodes are equipped with a lightweight processing engine based on the Intel NUC hardware platform to perform preliminary processing on the data, such as filtering outliers and calculating crop growth environment indicators in real time (e.g., accumulating photosynthetically active radiation values ​​through PAR photosynthetic radiometer data), effectively reducing the computing power pressure on the cloud.

[0048] Data transmission and edge computing utilize a hybrid edge-cloud architecture combining NB-IoT / 4G and LoRaWAN to overcome the bottleneck of high latency in traditional cloud processing. This enables real-time transmission of critical data (such as trigger data for soil moisture <60% field capacity and feature hash values ​​of pest and disease images) and off-peak transmission of non-sensitive data (such as high-definition video streams) during off-peak hours (nighttime when bandwidth utilization is low). Simultaneously, the edge computing device integrates hardware-level TEE encryption and isolation technology to encrypt critical data such as irrigation commands using the national standard SM4 encryption, preventing data tampering (e.g., intercepting and alerting to commands that illegally modify irrigation duration). Adopting an IP67 protection design, it can operate continuously in extreme environments ranging from -40℃ to 85℃, adapting to various scenarios such as greenhouses and open fields. This provides stable data support for smart agriculture, driving the transformation of production models from experience-based planting to data-driven approaches. For example, a greenhouse used this architecture to achieve early warning of tomato gray mold, reducing prevention and control costs.

[0049] Data transmission is stably delivered via the SSL / TLS 1.3 protocol, ensuring the authenticity and integrity of production entity information and equipment data. Edge computing nodes preprocess data through three levels of data cleaning (physical layer rule filtering, statistical layer outlier removal, and semantic layer ontology mapping), laying the foundation for digital identity authentication of production entities and solving the problem of identity verification in circulation. Its low latency characteristics facilitate real-time processing of authentication requests in transaction and sharing scenarios. Combined with large models and authentication centers, it shortens the time for identity and ownership authentication, improving the efficiency of agricultural digital authentication. Data processed by edge computing is transmitted and circulated within the authentication system. Financial institutions can assess credit risk based on the on-chain environmental stability index, promoting data assetization. At the same time, the privacy computing technology of edge computing helps agricultural innovation and drives high-quality industrial development.

[0050] The method further includes: optimizing the crop information, farm information, and production information using a deep learning model; the deep learning model has an SE attention module at its network neck for recalibrating channel weights; and the channel shuffling operation of the deep learning model includes group shuffling and pointwise convolution.

[0051] Specifically, MobileNetV3-Large is used as the basic skeleton to reduce the number of parameters and computational load. Meanwhile, an SE attention module (squeezing incentive mechanism) is introduced at the network neck. By recalibrating channel weights (enhancing the weights of pest and disease feature channels by 1.2 times), the accuracy of aphid identification in images is maintained. For the ShuffleNetV2 architecture, the channel shuffling operation is optimized to "group shuffling + pointwise convolution," reducing the computational loss of cross-group information interaction. While maintaining a 91% crop variety recognition rate, the model inference speed is improved, meeting the real-time requirements of edge devices.

[0052] In the raw channel shuffling of ShuffleNetV2, random channel shuffling can easily lead to redundant information exchange across groups. Grouped shuffling, based on crop feature correlation (leaf texture and fruit morphology channels are grouped together), divides the total number of channels C into G feature groups (G=4-8 is recommended for agricultural scenarios, adapting to pest and disease, and variety feature dimensions). Channel rearrangement is only performed within each group. The formula is simplified to: ;in, It is the g-th feature. To shuffle channels within a group and avoid cross-group interference from irrelevant features (soil background and crop lesions), thus reducing computational redundancy, `GroupShuffle` is a custom function used to group, shuffle, and concatenate agricultural image feature tensors. This allows the model to enhance feature diversity while preserving the semantic relevance of agriculture. `Concat` is the operation of concatenating tensors along the channel dimension, merging the sub-tensors after shuffling multiple groups. `G` is the total number of groups, adjusted according to the needs of the agricultural task. `Q` is the index of the group, from 1 to G, representing the q-th group (e.g., when G=4, q=1, 2, 3, 4 correspond to four groups).

[0053] After grouping and shuffling, we introduce a formula for linear transformation between channels using pointwise convolution:

[0054] ;in, Using a 1×1 convolution kernel, the channel compression ratio was set to (0.5-0.7) in agricultural scenarios to balance computational cost and feature preservation. This serves three purposes: first, it compensates for inter-group isolation during group shuffling with a computationally low cost (only one-ninth that of a 3×3 convolution), integrating crop features from different groups, such as leaf lesions and ambient light levels; second, it compresses high-dimensional features like 128 channels to 64-96 channels, adapting to the memory requirements of edge devices; and third, for feature channels related to aphids and downy mildew, the weights were adjusted through pre-training to enhance the response to lesion features. In experiments, the activation value of the aphid channel increased by 35%. is the feature tensor output by pointwise convolution, which is the result of shuffling the grouped data and then performing a 1×1 convolution transformation, used in subsequent model layers. B is the bias term, used to fine-tune the output of the pointwise convolution, compensate for the offset of the linear transformation, and make the feature distribution more in line with the needs of agricultural tasks, thus optimizing agricultural planting schemes.

[0055] For example, the INT8 symmetric quantization scheme is used to map 32-bit floating-point weights to the integer range of [-128, 127], reducing quantization errors, compressing model size, and lowering memory usage on edge devices. L1 regularization is used to prune convolutional layer channels, retaining the top 60% of channels (calculated through Taylor expansion to evaluate the gradient contribution of each channel to the loss function), improving the inference speed of the pruned model and reducing errors in soil EC value prediction.

[0056] A crop category vector is set for the crop images of the target farm; the convolutional kernel weights in the deep learning model are dynamically adjusted through a multi-head self-attention mechanism; the attention level of crops in the deep learning model is determined according to the crop category vector and the convolutional kernel weights, which is used for variety identification and / or pest and disease detection; an adaptive sharpening preprocessing is performed on the crop images using a path aggregation network to identify lesion features; soil and meteorological time-series data are determined through the deep learning model to predict growth trends and maturity, and the meteorological time-series data is integrated into the crop growth cycle as a feature solar term encoding.

[0057] Specifically, crop category embedding vectors are introduced (e.g., Solanaceae crop vector is [0.8, 0.2, 0.1], and Brassicaceae crop vector is [0.2, 0.7, 0.3]). The weights of the convolution kernel are dynamically adjusted through a multi-head self-attention mechanism, which improves the model's attention to the differences in leaf features of different crops. In the image, the cross-crop recognition accuracy of downy mildew spots reaches 95.3%.

[0058] The multi-head self-attention mechanism includes: setting up 4-8 parallel attention heads (adapted to crop feature dimensions), each head focusing on capturing different types of agricultural features. For example, some heads focus on leaf edge texture (distinguishing between compound leaves of Solanaceae and simple leaves of Brassicaceae), some heads focus on the color and shape of lesions (the light brown spots of downy mildew and the black clusters of aphids), and other heads specifically associate with crop growth stage features (differences in leaf size between seedling and mature stages). Each attention head independently calculates attention weights, and then the multi-dimensional features are associated by splicing and fusing.

[0059] The weight matrix output by the attention mechanism (with dimensions matching the number of convolutional kernels) is element-wise weighted with the original convolutional kernels. For high-interest feature channels (channels containing lesions), the weights of the corresponding convolutional kernels are scaled to 1.2-1.5 times (enhancing feature extraction capabilities); for low-interest background channels (soil, weeds), the weights of the convolutional kernels are compressed to 0.5-0.8 times (suppressing interference). For example, in downy mildew detection, the model identifies light brown spots on the underside of leaves as a key feature through the attention mechanism, and the weights of the corresponding convolutional kernels are dynamically increased, resulting in a more than 40% increase in the response value of the lesion area.

[0060] For example, this application employs the PANet feature pyramid, adding lateral connections and top-down / bottom-up paths between the C3, C4, and C5 feature layers output by the backbone, enhancing the feature extraction capability for small targets and improving the small target detection recall rate to 89%. Adaptive sharpening preprocessing is applied to leaf images to highlight lesion edge features, improving the accuracy of early gray mold (lesion area <5% of leaves) identification by 12 percentage points. The Transformer module has undergone technical optimization for soil and meteorological time-series data in growth trend prediction and maturity assessment. It optimizes the position encoding formula through agricultural cycle adaptive position encoding:

[0061] In this model, SEASON(t) is a solar term code (Vernal Equinox = 0.2, Grain in Ear = 0.6). Incorporating crop growth cycle characteristics reduces the tomato ripening time prediction error from 3.2 days to 1.8 days. Simultaneously, a gated recurrent unit (GRU) is added to the Transformer decoder, employing a long-short-term attention gating mechanism. This assigns high attention weight (0.7) to meteorological data (temperature, light intensity) from the past 7 days and low weight (0.3) to data from 14 days prior, effectively addressing the information dilution problem in long-term time-series dependencies and significantly improving the R-value of yield prediction. 2 The value PE(t) represents the location code, where t is the natural time step, recording the cumulative number of days from the start of crop growth or the observation start point, representing a continuous temporal process. i is the feature dimension index, which iterates through and generates location code components of different frequencies. D is the feature dimension of the model, determining the complexity of location coding and feature representation.

[0062] Step S102 involves constructing a blockchain based on the crop information, farm information, and production information. The blockchain includes a main chain, a logistics sub-chain, a quality inspection sub-chain, and a production sub-chain.

[0063] According to one specific implementation, the main chain includes key data, which includes at least one of trigger data for field water holding capacity and feature hash values ​​of pest and disease images; the logistics sub-chain includes transportation trajectories and storage environments; the quality inspection sub-chain includes pesticide test results and quality ratings; and the production sub-chain includes agricultural operation records and seedling information.

[0064] According to a specific implementation method, such as Figure 3 As shown, in the agricultural digital authentication scenario, the data processing and storage layer serves as the cornerstone of the trusted construction of agricultural digital authentication. It processes agricultural production entity information (such as farmer identity ID and enterprise qualifications) and agricultural operations (such as sowing time and fertilizer application) through standardized formatting. It uses the SHA-256 hash algorithm to ensure the authenticity and integrity of the data, providing a reliable data source for the digital identity authentication of production entities and solving the problem of identity verification in circulation.

[0065] This application can use DeepSeek large models to process the data, such as Figure 4 As shown, the process includes: a preprocessing branch that cleans and normalizes data to prepare for accurate analysis (e.g., using a three-level data cleaning strategy, the physical layer establishes an agricultural economic data anomaly rule library containing rules such as considering prices of common vegetable categories >50 yuan / kg as anomalies and logistics costs >3 times the regional average as anomalies, and uses a rule engine to filter in real time to achieve anomaly data identification rate); a statistical layer that, for continuous data, first uses the 3σ principle to remove data outside the mean ±3 times the standard deviation, and then uses the isolated forest algorithm to regard samples with a tree depth of more than 20 as outliers, improving the accuracy of outlier identification; and a semantic layer that uses agricultural economic ontology mapping tools to transform planning text into value factors such as the product premium corresponding to green certification rewards, and uses knowledge graph reasoning to achieve semantic consistency verification, reduce the dimensionality of data features, improve model training efficiency, and normalize soil nutrient data of different dimensions, converting them into comparable indicators.

[0066] Specifically, this application determines crop growth trends, environmental stresses, and the probability of pest and disease occurrence based on the aforementioned crop information, farm information, and production information. The agronomic value time-series modeling employs a three-layer bidirectional LSTM network with 256 memory units per layer. Layer Normalization layers are configured to address the gradient vanishing problem. Inputs include historical yield, accumulated temperature, precipitation, sunshine duration, and other meteorological factors, as well as crop phenological data, covering 12 months at weekly granularity. A gating mechanism is used to capture the temporal characteristics of key growth stages, such as during the rice grain-filling stage (July-August). The model uses the formula... Wherein, γ is the growth period sensitivity coefficient, and f(t) is the phenological period indicator function. The temperature factor weight is automatically increased. When the historical market price fluctuation exceeds a threshold, a cross-year pattern matching algorithm is triggered. DTW dynamic time warping is used to calculate the similarity between the current and historical sequences (if the similarity > 80%, the corresponding year's weight coefficient is reused, and the response latency is controlled within 100ms). Furthermore, image and environmental data are combined to provide early warnings of pest and disease occurrence probabilities. For example, an improved YOLOv9 model is used to identify aphids and downy mildew spots in 224×224 pixel images for early warning. A GNN graph network is used to model the relationship between crops, the environment, and pathogens. A message-passing neural network is used to simulate the pathogen spread path. Simultaneously, an adaptive position encoding fusion is performed based on a Transformer+LSTM hybrid network. By incorporating agricultural cycle characteristics such as solar terms and diurnal rhythms, the system controls prediction errors for key nodes like the corn jointing stage and rice tillering stage within predetermined timeframes. It also constructs density, ventilation, and disease correlation models based on soil moisture and light data to automatically recommend optimal planting spacing (e.g., optimizing tomato row spacing from 1.2 meters to 1.5 meters reduces powdery mildew incidence by 28%). Furthermore, it uses a genetic neural network to construct a knowledge graph of varieties, environment, and management practices, quantifying the 0.78 correlation coefficient between nitrogen fertilizer application and yield. It embeds agronomic rules such as organic fertilizer application of ≥500 kg / mu, employs a multi-task learning framework combined with soil spectral data and meteorological factors to reduce yield prediction errors, and generates multi-dimensional feature vectors such as environmental adaptability through capsule networks to dynamically analyze the potential impact of extreme weather on yield.

[0067] The execution branch outputs specific operation instructions to connect intelligent decision-making with actual production. For example, when the soil moisture is less than 60% of the field capacity, an instruction to open the solenoid valve and irrigate for 20 minutes is generated within 100ms. This instruction is sent to the smart valve via a LoRa wireless module. Before the valve is executed, the edge node confirms the receipt of the instruction through relay status feedback. If the response timeout occurs, the instruction is automatically retransmitted. After irrigation, the effectiveness of the decision is verified based on the soil moisture recovery curve (e.g., the moisture recovers to more than 70% within 2 hours). In case of anomalies, manual intervention is triggered. Another example is when the AI ​​identifies more than 50 pests per trap, it automatically triggers the linkage between the insecticidal lamp and the drone spraying, with a response time of less than 10 minutes. It can also generate fertilizer formula adjustment suggestions based on the analysis results, such as nitrogen fertilizer application rate = soil EC value × 0.8 + base value.

[0068] Intelligent decision-making systems integrate and analyze results to generate production plans adapted to local farmland, enabling intelligent decision-making in agricultural production. In smart agriculture scenarios, these systems serve as the core engine for precise and intelligent production. Through in-depth analysis of multi-source data combined with deep learning model algorithms, they accurately predict crop growth trends, the probability of pests and diseases, and the impact of environmental stresses. For example, they can predict the impact of high temperatures and drought on yields and generate corresponding strategies. Based on the analysis results, they automatically issue instructions to precisely control irrigation, fertilization, and plant protection equipment, achieving efficient resource utilization. This includes improving fertilizer utilization rates and increasing crop yield and quality. Furthermore, through continuous learning and optimization, they continuously improve production strategies, driving the transformation of agricultural production from experience-driven to data-driven approaches. For instance, they provide early warnings of tomato gray mold, reducing yield prediction errors and contributing to the modernization of agriculture.

[0069] Step S103 involves performing blockchain notarization on the blockchain to obtain certification certificates for each crop in the target farm. The blockchain notarization includes notarization of production data, environmental data, management data, processing data, and tactile data.

[0070] According to a specific implementation, the step of performing blockchain notarization to obtain certification certificates for each crop in the target farm includes: determining the growth status hash value of the crop based on the main chain and the production sub-chain; determining environmental parameters based on the main chain, the logistics sub-chain, and the production sub-chain, and generating a warning event hash value when the environmental parameters are abnormal; determining the operation trajectory and the associated hash value of the input based on the logistics sub-chain, the quality inspection sub-chain, and the production sub-chain; determining the process chain hash value for each processing batch based on the logistics sub-chain, the quality inspection sub-chain, and the production sub-chain; determining the transportation process hash value based on the logistics sub-chain; and storing the growth status hash value, the warning event hash value, the associated hash value, the process chain hash value, and the transportation process hash value on the blockchain to obtain the certification certificate for each crop in the target farm. Specifically, the certification certificate includes the crop variety name, grower information, planting time, growth process data, test results, certification conclusion, and risk warning label.

[0071] By storing processed data long-term through an object storage system and combining it with blockchain technology (using the FISCO BCOS consortium blockchain, with a single-chain throughput ≥10000 TPS), the system stores key agricultural production data such as land ownership and historical yield and quality benchmarks on the main chain; agricultural product transportation trajectories (origin, transit points, destination) and storage environment data (temperature, humidity, storage time) recorded on the logistics sub-chain; quality control data such as pesticide residue test results and quality ratings (e.g., sweetness grading) retained on the quality inspection sub-chain; and agricultural operation records (e.g., fertilizer formula, irrigation frequency) and seedling information (variety, seedling time) entered on the production sub-chain, forming an immutable agricultural data asset archive. Financial institutions can assess farmers' credit based on the on-chain data, shortening loan approval cycles. Regulatory authorities can verify the entire process data of agricultural products to achieve quality supervision and traceability; e-commerce platforms can carry out graded sales based on on-chain quality rating data, support the application of agricultural digital certification in financial, e-commerce and other scenarios, ensure the security of data circulation through encryption algorithms, promote the assetization of agricultural data and industrial collaboration, and generate tamper-proof digital ID cards for agricultural production entities through distributed storage and encryption technology. In the data sharing and transaction process, it ensures clear data ownership and traceability of data flow, supports farmers, banks, regulators and other parties to participate in maintaining the data ledger, and provides a decentralized, transparent and trustworthy operating environment for the agricultural digital certification system, promoting the digital transformation of the agricultural industry. For example, a certified tomato can achieve full-process traceability from sowing (seed RFID tags are linked to blockchain traceability codes) to processing through this system, and consumers can scan the code to obtain the certification report of the data node.

[0072] Step S104 involves tracking the crops in the target farm based on the certification certificate.

[0073] According to a specific implementation method, such as Figure 5 As shown, through multi-layered verification mechanisms such as biometrics (e.g., fingerprints, facial recognition), authentication certificates, and blockchain address binding, a unique and trustworthy digital identity is established for farmers, agricultural enterprises, and other entities. This ensures a trustworthy foundation for agricultural data on the blockchain (e.g., production data, logistics data) and financial transactions (e.g., loans, insurance). Meanwhile, the financial service interface, based on blockchain-stored production data (including historical yields, quality ratings, etc.), logistics data (transportation trajectories, storage temperature and humidity), and quality inspection data (pesticide residue test results, quality grading), uses a random forest algorithm to provide credit assessment basis for agricultural entities, supporting financial innovations such as order-backed loans and weather index insurance (e.g., automatic claims processing for cumulative high-temperature duration exceeding thresholds).

[0074] By managing device access permissions (such as tiered access control for sensor data acquisition) and verifying data authenticity (SHA-256 hash comparison on the blockchain), the credibility of production data throughout the entire process is ensured. For example, it ensures that the EC values ​​collected by soil sensors have not been tampered with, providing reliable input for intelligent decision-making systems. Financial services have innovatively launched weather index insurance products based on certified real data (such as when the temperature in the greenhouse exceeds 35°C for a cumulative period of 24 hours, the smart contract automatically triggers claims). At the same time, based on the order-pledged loan model, the approval cycle has been reduced, and the immutability of blockchain has been used to monitor credit risk in real time, promoting the intelligent and industrialized development of agricultural production and improving the loan approval rate for farmers.

[0075] In the context of agricultural digital certification, digital certification and financial services form key pillars of a trust ecosystem. Digital certification uses a Hardware Security Module (HSM) to store the subject's private key and combines it with zero-knowledge proof technology to achieve anonymous verification of identity information, establishing a unique and trustworthy digital identity for agricultural products and transactions. Consumers can scan codes to trace the entire process data from planting RFID tags (linked to variety registration information) to processing and sterilization temperatures, solving the problem of difficulty in verifying the subject's identity. Financial services rely on the certification system and, based on data assets such as soil fertility index and equipment utilization rate stored on the blockchain, construct risk control models (e.g., risk index = environmental stability × 40% + management compliance × 35% + quality rating × 25%) to achieve accurate credit assessment. Agricultural enterprises can obtain special loans with reduced profit margins based on twelve consecutive months of certified compliance records. At the same time, the future income rights of certified farmland are packaged into ABS products, promoting the deep integration of agricultural data assetization and financial services, and driving the digital transformation of the agricultural industry. For example, a certified strawberry can achieve a premium in the end market through this system and obtain automatic supply chain finance loans.

[0076] By aggregating data from various blockchain sub-chains, the platform enables consumers and regulatory authorities to query information on the entire process of agricultural products "from farm to table" (such as production environment, production process, and quality inspection results), thus supporting quality supervision and brand building. In smart agriculture scenarios, the traceability platform is a crucial tool for optimizing production management and enhancing industrial competitiveness. It integrates data from the entire process, including soil environment, meteorological conditions, and agricultural operations, helping farmers and businesses trace the production process and analyze historical data to optimize planting and breeding strategies, such as improving plant protection plans by reviewing pest and disease control records. Simultaneously, the platform's data supports the iteration of intelligent decision-making systems, providing historical references for precision irrigation and scientific fertilization, thereby improving resource utilization efficiency. Furthermore, the platform can display information such as the growing environment of agricultural products and pesticide use to consumers, enhancing market trust in products, helping to build high-quality agricultural brands, and increasing the added value and market competitiveness of agricultural products.

[0077] In the context of agricultural digital certification, traceability platforms serve as a core hub for building a trustworthy agricultural ecosystem. They visualize certification data stored on the blockchain, including the identity of the production entity, agricultural product quality inspection reports, and logistics trajectories. This allows regulatory authorities to quickly verify product compliance, promptly identify risks such as data fraud, and ensure the authority of the certification system. For consumers, scanning a code verifies the digital identity of agricultural products, providing access to information across the entire supply chain from farm to table, enhancing trust in the certification results. In the financial sector, financial institutions can use the platform to trace enterprise production and operational data, assess credit risk, and provide more precise financial services to agricultural entities, promoting the digitalization and trustworthiness of the agricultural industry.

[0078] Example 2

[0079] like Figure 6 As shown, this invention constructs a complete ecosystem from physical perception to digital applications through multi-dimensional technology integration and full-chain data value mining. At the technical implementation level, the perception layer devices form a three-dimensional monitoring network: the ground monitoring system deploys a high-definition camera array, using OpenCV and AI models to achieve machine vision analysis of growth characteristics such as plant height, leaf color, and fruit morphology, while also serving as a security monitoring function; soil sensors are buried in layers at threshold intervals, integrating sensing units for temperature, humidity (pH value, EC value, and NPK nutrient content, etc.), converting soil physicochemical properties into electrical signals through physical and chemical principles and conditioning them through a local controller; the meteorological monitoring system collects microclimate data such as air temperature and humidity, light intensity, wind speed and direction, precipitation, and atmospheric pressure at fixed frequencies; the low-altitude remote sensing system uses drones equipped with multispectral / thermal infrared cameras and lidar to achieve pest and disease identification, crop lodging monitoring, and plant height measurement; the macro-monitoring system relies on satellite or large drone multispectral / hyperspectral / microwave sensors to acquire macro-data such as NDVI vegetation index and soil moisture in large areas of farmland at preset intervals; and the human-computer interaction system supports farmers to input structured management data such as sowing time and fertilizer application amount through mobile apps and PC platforms, supplementing the information gaps of automated equipment.

[0080] The edge intelligent computing layer deploys a lightweight engine to implement three levels of data cleaning: the physical layer filters out-of-range data, the statistical layer removes outliers using the IQR method, and the semantic layer verifies the logical rationality based on an agricultural knowledge graph. Simultaneously, it calculates key parameters such as DLI (Photosynthetically Active Radiation Integral) and soil water potential thresholds in real time, and extracts pest and disease image features using a lightweight model. A security mechanism integrates a trusted execution environment and a federated learning framework to ensure the security of data processing and cross-domain collaboration, keeping response latency within a set timeframe, and employing a staggered transmission strategy for non-sensitive data.

[0081] The blockchain evidence storage process utilizes the FISCOBCOS consortium blockchain to store key business data such as fertilizer calculation formulas and planting batches on the chain. Smart contracts automatically verify the compliance of data like fertilizer application rates, standardizing agricultural processes and ensuring operational trustworthiness. In terms of data value application, precision agricultural decision-making integrates data such as soil moisture and ET (evaporization-transpiration) for precise irrigation. Fertilization plans are dynamically adjusted based on soil NPK and crop spectral data (e.g., the formula on the chain: fertilizer application rate = soil EC value × 0.8 + baseline value). Drone imagery and AI models provide early warnings of pests and diseases such as tomato gray mold, and data-driven optimization of greenhouse environmental control is employed. Agricultural product traceability integrates data from the entire planting, growth, and harvesting cycle. Consumers can scan a code to access tamper-proof information across the entire chain. A strawberry brand leverages this to achieve premium pricing and provides electronic certification for organic products. In the agricultural finance sector, data such as environmental stability indices and agricultural compliance rates improve farmers' loan approval rates, reduce default rates, and provide objective data for insurance pricing and loss assessment. Furthermore, the production data from successful farms on the blockchain promotes the replication of high-quality models, and the cleaned data assets are circulated for farm rating, realizing full-chain empowerment from technological perception to industrial value.

[0082] As a key hub connecting physical and digital agriculture, the data transmission and edge computing module integrates multi-layered technologies such as heterogeneous data aggregation, edge intelligent processing, secure transmission, and digital authentication, forming a complete closed loop from data acquisition to value release. At the heterogeneous data aggregation level, the IoT gateway adopts a modular protocol stack design, integrating multiple communication protocols and achieving unified data modeling through a protocol abstraction layer: 224-band spectral data from soil sensors is converted to JSON format after transmission via LoRaWAN; temperature and humidity data sampled at 1Hz from the weather station is parsed and timestamped via Modbus-RTU; and 25fps 4KH / 265 video streams are accessed via the RTSP protocol to extract metadata. The gateway has a built-in protocol parsing engine based on the EclipseKura framework. It manages high-concurrency data streams with the help of a circular buffer, and a single gateway can support more than 500 nodes. At the same time, it adopts a QoS-based priority scheduling algorithm to set key data such as soil moisture alarms as the highest priority and connect them directly to edge nodes in real time. 4K video streams are transmitted at night through a bandwidth prediction model to avoid peak hours. Combined with forward error correction coding, it ensures that the packet loss rate is less than 0.1%. When the network is congested, it automatically switches to the NB-IoT / 4G backup link.

[0083] The data processing stage implements a three-level cleaning process: For soil temperature and other data streams, the IQR interquartile range method is applied according to a set time window to calculate Q1 and Q3, removing outliers (such as ±10℃ jumps) outside the range of [Q1-1.5×IQR, Q3+1.5×IQR]. PAR (Photosynthetically Active Radiation) is collected using a spectrometer calibrated with NIST standard light sources, and the daily cumulative light integration error of DLI (Digital Light Integrator) is controlled to <3% using a linear compensation algorithm. For transmission optimization, second-level data from weather stations is compressed using a rotating door algorithm, retaining only inflection points and deviation points to achieve compressed data. Critical data is directly connected to edge nodes via a dedicated network, while non-sensitive data is transmitted using LoRaWAN during off-peak hours at night. Dual-stack communication proxies are deployed at edge nodes to achieve automatic data routing. For security protection, critical data such as irrigation commands are encrypted using SM4-CTR mode, with the key stored in a secure area, and the command hash value is verified in real time to prevent tampering. The data-driven agricultural digital authentication system constructs a trusted data chain through a three-level cleaning process: the physical layer filters hardware fault data outside the pH range of 0-14; the statistical layer combines IQR and sliding windows to smooth out random noise; and the semantic layer verifies the logical rationality based on a crop growth rule base. This application, through deep coupling of protocol fusion, edge computing, secure encryption, and blockchain notarization, addresses the requirements for low power consumption, low latency, and anti-interference in agricultural data transmission. Furthermore, through data cleaning and trustworthiness processing, it provides core support for precision agriculture, financial risk control, and brand value enhancement, forming a technological closed loop of perception, transmission, computing, and application, thus driving the paradigm shift in agricultural production from experience-driven to data-driven.

[0084] Large files can be stored using the Ethereum public blockchain combined with IPFS, replacing consortium blockchains. This solution features low deployment costs and high flexibility, making it suitable for small and medium-sized agricultural cooperatives or individual farmers. On the public blockchain, smart contracts are used for data authentication and management, while IPFS's distributed storage features are utilized to store large files, reducing data storage costs.

[0085] For edge devices with limited computing resources (such as field gateway devices), the large agricultural digital authentication model can be lightweighted by adopting a lightweight neural network architecture (such as MobileNet, ShuffleNet, etc.) to reduce the number of model parameters and computational load. Through model compression and quantization techniques, the model parameters can be reduced by more than 60%, enabling it to run on edge devices, achieving localized real-time analysis, reducing dependence on cloud servers, and improving the real-time performance and reliability of authentication.

[0086] This application can also employ ensemble learning methods to replace single large models, such as combining multiple different types of machine learning models (e.g., random forests, support vector machines) for authentication analysis. Each model is responsible for processing a specific type of data or task, and the results of each model are combined through voting mechanisms or weighted averaging to improve the accuracy and robustness of authentication. In field crop planting scenarios, drone inspections can replace fixed cameras for data collection. Drones equipped with multispectral or high-resolution cameras regularly conduct aerial photography of field crops to acquire canopy images and multispectral data (e.g., NDVI values, red-edge index) for analyzing crop growth status, pest and disease occurrence, and yield prediction. Drone inspections have the advantages of wide coverage and high collection efficiency, making them suitable for monitoring and authentication of large areas of farmland. For livestock and poultry farming scenarios, smart wearable devices (e.g., smart collars, ear tags) can replace some sensors for data collection. Smart wearable devices can monitor physiological indicators such as body temperature, heart rate, and activity level of livestock and poultry in real time, and transmit the data to a cloud server via Bluetooth or wireless communication technology. This method can obtain more accurate real-time data on individual livestock and poultry, providing a more detailed basis for the health management and certification of livestock and poultry.

[0087] The data processing and storage layer technology architecture of smart agriculture revolves around the management of the entire data lifecycle. Through the integration of multiple technologies, including standardized processing, intelligent analysis, trusted storage, and security assurance, it constructs an efficient, accurate, and secure data value transformation system. In terms of data standardization and formatting, the NTP protocol achieves UTC timestamp synchronization with ±1ms precision, eliminating time zone differences across different scenarios. The W3C standard JSON-LD format is used to semantically encapsulate multi-source data, giving it structured semantic relationships. For example, soil sensor data, after encapsulation, includes information such as device ID, geographical location, measurement indicators, and timestamps. A lightweight format conversion middleware pre-parses the raw binary data at edge nodes, and the cloud uses the Apache NiFi pipeline to complete the final formatting, ensuring data format uniformity. Simultaneously, a three-level data cleaning process is implemented: the physical layer filters outliers in real time based on a rule base, such as identifying abnormal market data where vegetable prices exceed 50 yuan / kg; the statistical layer combines the 3σ principle and the isolated forest algorithm to remove outliers in soil nutrient data that deviate from the mean by three times the standard deviation; and the semantic layer quantifies planning text into cost coefficients through agricultural economic ontology mapping.

[0088] In the field of pest and disease identification, an improved attention mechanism is used to accurately identify small targets such as aphids and gray mold. Once edge computing identifies pest characteristics, agricultural drones can plan their operational paths within a predetermined time. Crop growth prediction employs a fusion model of Transformer and LSTM, with the former encoding historical meteorological sequences and the latter decoding the growth stage status.

[0089] The agricultural digital authentication and trustworthy system achieves data storage and value application through a multi-chain collaborative blockchain architecture. The main chain stores land ownership and yield baselines; the logistics sub-chain records temperature and humidity trajectories, ensuring data security through channel isolation; the quality inspection sub-chain uses zero-knowledge proofs to verify pesticide residue compliance; and the production sub-chain uses smart contracts to automatically verify the compliance of agricultural operation formulas. The storage process encompasses data hashing, joint signatures (between farmers and regulators), and cross-chain anchoring (main chain aggregating sub-chains). Significant data assetization applications are achieved.

[0090] This application constructs a three-in-one digital identity verification system integrating biometrics, national cryptographic certificates, and on-chain addresses; it innovates the on-chain insurance actuarial model by writing continuous variables such as the cumulative duration of high temperatures into smart contracts; and it transforms farmland revenue rights into tradable ABS through environmental stability indices. This module not only solves the financial pain point of insufficient credibility of agricultural data, but also expands agricultural financing channels through data assetization, promoting the deep integration of agricultural production and financial services.

[0091] Heterogeneous data normalization processing adopts a classification and transformation strategy: sensor data is converted into JSON-LD format, such as soil sensor data being encapsulated into structured data with timestamps and geographic locations; video data is extracted with one key frame per second and compressed to 1080P, then timestamped and stored; text data is mapped to feature vectors through ontology, such as "500 kg / mu of organic fertilizer" being mapped to a crop fertilizer requirement node in a knowledge graph.

[0092] This application constructs a spatiotemporal data cube, integrating geographic information, time series data, and blockchain business data to achieve millisecond-level penetrating queries of soil EC values ​​at specific times; it realizes the dual value release of data assets, internally optimizing plant protection models through historical pest and disease data. Through technological integration, the platform not only solves the trust problem of agricultural product traceability but also provides support for production optimization and financial services through data mining, promoting the digital and value-added transformation of the agricultural industry chain.

[0093] Example 3

[0094] This invention also proposes an agricultural data authentication device, namely a large-scale model for modern agricultural digital authentication, such as... Figure 7As shown, the agricultural data authentication device 100 includes: an acquisition module 200 for acquiring data information of a target farm, the data information including crop information, farm information, and production information; a first processing module 300 for constructing a blockchain based on the crop information, farm information, and production information, the blockchain including a main chain, a logistics sub-chain, a quality inspection sub-chain, and a production sub-chain; and a second processing module 400 for performing blockchain notarization on the blockchain to obtain authentication certificates for each crop in the target farm, the blockchain notarization including production data notarization, environmental data notarization, management data notarization, processing data notarization, and tactile data notarization, and tracking the crops in the target farm based on the authentication certificates. The process of obtaining certification certificates for each crop in the target farm by storing blockchain evidence includes: determining the growth status hash value of the crop based on the main chain and the production sub-chain; determining environmental parameters based on the main chain, the logistics sub-chain, and the production sub-chain, and generating a warning event hash value when the environmental parameters are abnormal; determining the operation trajectory and the associated hash value of the input based on the logistics sub-chain, the quality inspection sub-chain, and the production sub-chain; determining the process chain hash value of each processing batch based on the logistics sub-chain, the quality inspection sub-chain, and the production sub-chain; determining the transportation process hash value based on the logistics sub-chain; and storing the growth status hash value, the warning event hash value, the associated hash value, the process chain hash value, and the transportation process hash value on the blockchain to obtain the certification certificate for each crop in the target farm.

[0095] The device also includes: a third processing module, used to generate a biometric hash value based on the image information in the data information; construct a triple based on the biometric hash value, the public key of the authentication certificate, and the wallet address; and store the triple on the blockchain underlying platform to form a digital identity anchor for crops.

[0096] This application, through data integration and technological fusion, forms a closed loop in areas such as improving efficiency in smart agricultural production, building trustworthy blockchain, empowering digital authentication security, and supporting innovative financial services. It provides a full-chain solution for the transformation of agricultural modernization, promotes the coordinated development of agriculture towards precision, intelligence, trustworthiness, and financialization, and helps achieve high-quality upgrading and value enhancement of the agricultural industry.

[0097] An agricultural data authentication method of the present invention includes: acquiring data information of a target farm, the data information including crop information, farm information, and production information; constructing a blockchain based on the crop information, farm information, and production information, the blockchain including a main chain, a logistics sub-chain, a quality inspection sub-chain, and a production sub-chain; performing blockchain notarization on the blockchain to obtain authentication certificates for each crop in the target farm, the blockchain notarization including production data notarization, environmental data notarization, management data notarization, processing data notarization, and tactile data notarization; and tracking the crops in the target farm based on the authentication certificates. This method generates authentication certificates by collecting full-cycle agricultural data in real time and performing dynamic value calculation and risk assessment based on a digital agriculture big data model, providing financial institutions with quantifiable credit granting basis. The authentication certificate transforms the agricultural production process into standardized digital assets that financial institutions can recognize, providing a basis for confirmation of rights for financial instruments such as credit, futures, and trusts, and achieving accurate prediction of agricultural output and value.

[0098] This invention provides a storage medium storing a program that, when executed by a processor, implements the method for agricultural data authentication. This invention also provides a processor for running the program, wherein the program, when running, executes the method for agricultural data authentication. The memory may include non-persistent memory in computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0099] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0100] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0101] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for agricultural data authentication, characterized in that, The method includes: Obtain data information from the target farm, including crop information, farm information, and production information; Based on the crop information, farm information and production information, a blockchain is constructed, which includes a main chain, a logistics sub-chain, a quality inspection sub-chain and a production sub-chain. The blockchain is stored in a blockchain database to obtain certification certificates for each crop in the target farm. The blockchain storage includes production data storage, environmental data storage, management data storage, processing data storage, and tactile data storage. Based on the certification, data tracking is performed on the crops in the target farm; The step of performing blockchain notarization to obtain certification certificates for each crop in the target farm includes: Based on the main chain and the production sub-chain, determine the growth status hash value of the crop; Based on the main chain, logistics sub-chain, and production sub-chain, environmental parameters are determined, and when the environmental parameters are abnormal, a warning event hash value is generated. Based on the logistics sub-chain, quality inspection sub-chain, and production sub-chain, determine the operation trajectory and the associated hash value of the inputs; Based on the logistics sub-chain, quality inspection sub-chain, and production sub-chain, determine the process chain hash value for each processing batch; Based on the aforementioned logistics sub-chain, determine the hash value of the transportation process; The growth status hash value, early warning event hash value, association hash value, process chain hash value, and transportation process hash value are stored on the blockchain to obtain the certification certificates for each crop in the target farm.

2. The method according to claim 1, characterized in that, The main chain includes key data, which includes at least one of the following: trigger data of field water holding capacity and feature hash values ​​of pest and disease images. The logistics sub-chain includes transportation routes and warehousing environment; The quality inspection sub-chain includes pesticide testing results and quality ratings; The production sub-chain includes agricultural operation records and seedling information.

3. The method according to claim 1, characterized in that, The certification certificate includes the crop variety name, grower information, planting time, growth process data, test results, certification conclusion, and risk warning label.

4. The method according to claim 1, characterized in that, The method also includes: Generate a biometric hash value based on the image information in the data information; Based on the biometric hash value, the public key of the authentication certificate, and the wallet address, a triple is constructed and stored on the blockchain underlying platform to form a digital identity anchor for crops.

5. The method according to claim 1, characterized in that, The method also includes preprocessing the data information, including: The crop information, farm information, and production information are filtered using a cost rule base, edge computing, and standard formatting to obtain qualified data. The continuous data in the qualified data is cleaned and normalized to obtain preprocessed data.

6. The method according to claim 1, characterized in that, The method also includes: The crop information, farm information, and production information are optimized using a deep learning model; The deep learning model has an SE attention module in the network neck, which is used to recalibrate the channel weights; The channel shuffling operation of the deep learning model includes group shuffling and pointwise convolution.

7. The method according to claim 6, characterized in that, Set crop category vectors for the crop images of the target farm; The convolutional kernel weights in the deep learning model are dynamically adjusted using a multi-head self-attention mechanism. The crop attention level in the deep learning model is determined based on the crop category vector and the convolution kernel weights, and is used for variety identification and / or pest and disease detection. An adaptive sharpening preprocessing method using a path aggregation network is employed to identify lesion features in the crop images. Soil and meteorological time-series data are determined by a deep learning model to predict growth trends and maturity. The meteorological time-series data is incorporated into the crop growth cycle as a feature solar term encoding.

8. A device for agricultural data authentication, characterized in that, The device includes: The acquisition module is used to acquire data information of the target farm, including crop information, farm information and production information; The first processing module is used to construct a blockchain based on the crop information, farm information and production information. The blockchain includes a main chain, a logistics sub-chain, a quality inspection sub-chain and a production sub-chain. The second processing module is used to perform blockchain notarization on the blockchain to obtain certification certificates for each crop in the target farm. The blockchain notarization includes production data notarization, environmental data notarization, management data notarization, processing data notarization, and tactile data notarization. The module also performs data tracking on the crops in the target farm based on the certification certificates. The step of performing blockchain notarization to obtain certification certificates for each crop in the target farm includes: Based on the main chain and the production sub-chain, determine the growth status hash value of the crop; Based on the main chain, logistics sub-chain, and production sub-chain, environmental parameters are determined, and when the environmental parameters are abnormal, a warning event hash value is generated. Based on the logistics sub-chain, quality inspection sub-chain, and production sub-chain, determine the operation trajectory and the associated hash value of the inputs; Based on the logistics sub-chain, quality inspection sub-chain, and production sub-chain, determine the process chain hash value for each processing batch; Based on the aforementioned logistics sub-chain, determine the hash value of the transportation process; The growth status hash value, early warning event hash value, association hash value, process chain hash value, and transportation process hash value are stored on the blockchain to obtain the certification certificates for each crop in the target farm.

9. The apparatus according to claim 8, characterized in that, The device also includes: The third processing module is used to generate biometric hash values ​​based on the image information in the data information; A triple is constructed based on the biometric hash value, the public key of the authentication certificate, and the wallet address. The triple is then stored on the blockchain underlying platform to form a digital identity anchor for crops.

10. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the method of agricultural data authentication as described in any one of claims 1-7.

Images