Digital tobacco field production data processing method

Abstract

The invention discloses a digital tobacco field production data processing method, which comprises the following steps of: constructing a field-sky three-dimensional integrated intelligent tobacco field management system through a mode of satellite remote sensing, an unmanned aerial vehicle, a sensor, a terminal APP and a data center; collecting structured data and unstructured data in the tobacco leaf production process in real time in the tobacco field; all the collected structured data and unstructured data are effectively integrated through a data integration technology to form a complete data set; a deep learning technology is utilized to learn and analyze unstructured data, valuable information is extracted from the unstructured data, and through a field-sky-three-dimensional integrated intelligent tobacco field management system, and in combination with satellite remote sensing, an unmanned aerial vehicle, a sensor, a terminal APP and a data center, omnibearing and three-dimensional monitoring and management of a tobacco field are realized; structured data and unstructured data in a tobacco field are collected in real time, and various changes in the tobacco leaf production process can be rapidly reflected.

Description

Technical Field

[0001] The invention relates to the technical field of tobacco field management, in particular to a digital tobacco field production data processing method. Background Art

[0002] Every aspect of tobacco cultivation requires scientific planning and management to ensure an optimal growing environment, thereby improving yield, quality, and economic returns. Tobacco field management encompasses more than just agronomic operations; it encompasses ecological, technical, and economic aspects. The goal of tobacco field management is to promote healthy tobacco growth, minimize the impact of pests and diseases, and ensure that the harvested tobacco leaves meet quality standards through effective management.

[0003] In the process of tobacco field management, various data need to be analyzed and processed to ensure that the growth, quality and yield of tobacco reach the optimal level. The existing commonly used data processing methods are unable to collect multi-dimensional data from tobacco fields and cannot ensure the comprehensiveness and accuracy of the data. The accuracy and efficiency of data analysis are low, and potential problems and optimization points in tobacco field production cannot be discovered in a timely manner. Summary of the Invention

[0004] The purpose of the present invention is to provide a digital tobacco field production data processing method to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for processing digital tobacco field production data, comprising the following steps:

[0006] Through the model of satellite remote sensing + drones + sensors + terminal APP + data center, a three-dimensional intelligent tobacco field management system "Field Sky" is built;

[0007] Collect structured data and unstructured data from tobacco production in real time in tobacco fields;

[0008] Through data integration technology, all collected structured data and unstructured data are effectively integrated to form a complete data set;

[0009] Use deep learning technology to learn and analyze unstructured data to extract valuable information;

[0010] Through data models and data algorithms, the integrated data is correlated, integrated, and analyzed;

[0011] The analysis results are fed back to farm workers and decision makers through terminal APP or other management systems, and the production process is adjusted and resource allocation is optimized based on the analysis results.

[0012] As a preferred option, the technical framework of the "Field Sky" three-dimensional intelligent tobacco field management system includes satellite remote sensing, drones, sensors, terminal APP and data center.

[0013] Satellite remote sensing is used to obtain panoramic information and macro data of tobacco fields in real time;

[0014] The drone is used to provide high-resolution images of tobacco fields and more accurate field data collection;

[0015] The sensors are arranged in the tobacco field to monitor environmental parameters in real time;

[0016] The terminal APP is used to provide remote control and observation functions for field workers;

[0017] The data center is used to centrally store, process and analyze all collected data to build a comprehensive tobacco field data platform.

[0018] Preferably, the environmental parameters include soil quality, temperature and humidity, and climate conditions, and the remote control and observation functions include data viewing, remote operation, and real-time monitoring.

[0019] Preferably, the structured data includes tobacco leaf growth conditions, soil quality, and climate conditions.

[0020] Preferably, the unstructured data includes on-site images, videos, and audios, which are used to obtain more intuitive production environment information.

[0021] Preferably, the data integration technology includes a data acquisition module, a data cleaning and preprocessing module, and a data fusion and association module;

[0022] The data acquisition module is used to collect environmental parameters and unstructured data;

[0023] The data cleaning and preprocessing module is used to remove duplicates, clean and standardize the format of numbers;

[0024] The data fusion and association module is used to fuse multiple data sources and create a multi-dimensional unified data model.

[0025] Preferably, the deep learning technology includes an image processing and analysis module, an audio processing module, a time series data analysis module, a deep learning model, and a model training and prediction module;

[0026] The image processing and analysis module is used for analyzing image and video data;

[0027] The audio processing module is used for sound classification and recognition;

[0028] The time series data analysis module is used to use LSTM or GRU to model time series data and predict changes in environmental parameters or crop demand in the future;

[0029] The deep learning model is used to integrate information from different data sources into a single model, simulate the results of different strategies, and train the model to provide optimization suggestions for resource allocation and production processes;

[0030] The model training and optimization module uses pre-trained models to quickly migrate to specific agricultural environments or environmental monitoring tasks and perform data prediction.

[0031] Preferably, the data prediction includes crop prediction and pest and disease prediction and early warning. The crop prediction is based on data on soil quality, climate change, and crop growth cycle, and uses a deep learning model to predict crop growth trends, maturity time, and yield. The pest and disease prediction and early warning uses image processing and environmental data analysis, combined with historical pest and disease data, to predict the occurrence of pests and diseases and propose preventive measures.

[0032] Compared with the existing technology, the beneficial effects of the present invention are: through the three-dimensional intelligent tobacco field management system of "Tian Tiankong", which combines satellite remote sensing, drones, sensors, terminal APP and data centers, it realizes all-round and three-dimensional monitoring and management of tobacco fields; this multi-dimensional data collection method ensures the comprehensiveness and accuracy of data; real-time collection of structured data and unstructured data in tobacco fields can quickly reflect various changes in the tobacco production process; through data integration technology, all data are effectively integrated to form a complete data set, which provides a solid foundation for subsequent data analysis and decision-making; using deep learning technology to learn and analyze unstructured data, it can extract valuable information from it; this intelligent data processing method improves the accuracy and Efficiency, helps to discover potential problems and optimization points in tobacco field production; through data models and data algorithms, the integrated data is correlated, fused and calculated and analyzed; this data correlation and fusion method can reveal the inherent connections and laws between data, and provide more comprehensive and in-depth data support for farmland workers and decision makers; the analysis results are fed back to farmland workers and decision makers through terminal APP or other management systems, so that they can adjust the production process in time according to the analysis results and optimize resource allocation; this feedback mechanism ensures the flexibility and adaptability of tobacco field production, and helps to improve production efficiency and output; the comprehensive use of multiple data sources and advanced analysis technologies provides a scientific basis for tobacco field management, reduces the subjectivity and uncertainty of human judgment, and makes decision-making more scientific and reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1Schematic diagram of the method of the present invention. DETAILED DESCRIPTION

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0035] See also Figure 1 The present invention provides a technical solution: a method for processing digital tobacco field production data, comprising the following steps:

[0036] Through the model of satellite remote sensing + drones + sensors + terminal APP + data center, a three-dimensional intelligent tobacco field management system "Field Sky" is built;

[0037] Real-time collection of structured and unstructured data from the tobacco production process is performed in tobacco fields. Structured data includes tobacco growth, soil quality, and climate conditions, while unstructured data includes on-site images, videos, and audio, providing more intuitive information about the production environment. Structured data collection uses soil sensors, climate monitors, tobacco growth monitoring systems, and other equipment to collect real-time information on soil pH, soil moisture, fertility, temperature, humidity, precipitation, sunshine duration, and tobacco leaf area, growth rate, and color changes. This data is automatically collected and uploaded to the data center through sensors and weather stations. For unstructured data collection, image data is captured by drones and ground cameras to obtain high-definition images of tobacco fields, providing visual information on tobacco growth, pests, and weeds. Video data is captured by drones or ground cameras to obtain real-time dynamic videos of farmland, which are used to track changes in tobacco fields and the development of pests and diseases. Audio data is collected by audio sensors (such as microphones) to capture audio data from the field environment, such as insect chirping and wind noise.

[0038] Through data integration technology, all collected structured and unstructured data are effectively integrated to form a complete data set. The collected structured and unstructured data are cleaned to remove redundant or erroneous data. For image and video data, noise removal and quality screening are performed to ensure data accuracy. Data from different devices and sensors are synchronized according to timestamps to ensure that all types of data within the same time correspond. For spatial information, GIS (Geographic Information System) technology is used to synchronize location data in tobacco fields to ensure that all data are spatially consistent. Through data integration technologies such as ETL (Extract, Transform, and Load), structured data (such as soil moisture and temperature) and unstructured data (such as images and videos) are merged to form a unified data set.

[0039] Deep learning technology is used to learn and analyze unstructured data to extract valuable information. Deep convolutional neural networks (CNNs) are used to process images and videos of tobacco fields to extract information such as tobacco leaf growth status, pests and diseases, weeds, and environmental changes. Image recognition technology can be used to identify the health status, abnormal changes, or potential problems of tobacco leaves. Deep learning models are used to analyze audio data in field environments to identify abnormal audio patterns (such as the sounds of pests and diseases). For information extracted from images, videos, and audio, deep learning models are used to assess crop growth status and potential risks (such as pests and diseases, water shortages, etc.), classify, and score them to provide a basis for subsequent decision-making.

[0040] Through data models and data algorithms, the integrated data is correlated, integrated, and analyzed; the analysis results are fed back to farmland workers and decision makers through terminal APP or other management systems, and the production process is adjusted according to the analysis results to optimize resource allocation. The integrated structured and unstructured data are input into a variety of machine learning and deep learning models; through regression analysis, cluster analysis, time series analysis and other technologies, the temperature, humidity, climate, crop growth, etc. in the farmland are correlated and analyzed to predict future crop growth trends, climate change, pest and disease development, etc.; genetic algorithms, particle swarm optimization algorithms, etc. are used to optimize resource allocation; according to the analysis results, irrigation plans are formulated. , fertilization, pesticide spraying and other resource allocation plans to improve resource utilization efficiency, reduce waste and ensure the sustainability of farmland production; the analysis results are fed back to farmland staff and decision makers in real time through terminal APP or management system to show the production status of farmland, pest and disease risks, resource requirements, etc.; staff adjust production plans according to feedback, such as irrigation amount, fertilizer amount, application time, etc., to ensure optimal production benefits; for certain links in agricultural production, such as irrigation and fertilization, the control system can be automatically adjusted according to the analysis results to reduce manual intervention and improve efficiency; for example, the irrigation system can be automatically adjusted based on soil moisture data, or the fertilizer amount can be automatically adjusted according to crop growth data.

[0041] Among them, the technical framework of the "Tian Tiankong" three-dimensional smart tobacco field management system includes satellite remote sensing, drones, sensors, terminal APPs and data centers. Satellite remote sensing is used to obtain panoramic information and macro data of tobacco fields in real time; drones are used to provide high-resolution images of tobacco fields and more accurate on-site data collection; sensors are arranged in tobacco fields for real-time monitoring of environmental parameters; terminal APPs are used to provide remote control and observation functions for field workers; data centers are used to centrally store, process and analyze all collected data to build a comprehensive tobacco field data platform. Environmental parameters include soil quality, temperature and humidity, and climatic conditions. Remote control and observation functions include data viewing, remote operation, and real-time monitoring.

[0042] It should be noted that the satellite remote sensing system of the present invention uses satellite images to obtain macro-environmental information of a large area of ​​tobacco fields, and monitors data such as climate, soil moisture, temperature, and vegetation cover in real time. This data is obtained through remote sensing technology and integrated with other data sources. Unmanned aerial vehicles (UAVs) fly at low altitudes over tobacco fields to collect high-precision images and videos. UAVs are equipped with high-resolution cameras that can capture high-definition images and videos, and monitor local environmental factors such as climate, humidity, and light through built-in sensors. The data obtained by the UAVs can provide high-frequency, detailed local data for farmland. Sensors deployed in tobacco fields are responsible for collecting soil quality, humidity, temperature, and meteorological data in real time, such as soil sensors, climate sensors, and environmental sensors. The sensor data provides more accurate real-time information, facilitating immediate regulation of farmland. The terminal APP provides an interface for farmland workers and decision makers to view real-time data, control various equipment, receive system feedback, etc. The APP is connected to a data center and can display data from satellite remote sensing, UAVs, and sensors in real time. The data center stores, processes, calculates, and analyzes all data collected from various terminals and devices. The data center processes massive amounts of data and feeds back the analysis results to the terminal devices.

[0043] Among them, data integration technology includes data acquisition module, data cleaning and preprocessing module and data fusion and association module; the data acquisition module is used to collect environmental parameters and unstructured data; the data cleaning and preprocessing module is used to deduplicate, clean and standardize the format of numbers; the data fusion and association module is used to fuse multiple data sources to create a multi-dimensional unified data model.

[0044] It should be noted that when collecting environmental parameters in the present invention, satellite remote sensing technology is used to monitor the temperature, humidity, light, soil quality, precipitation, etc. of the tobacco fields. UAVs capture images and videos to obtain detailed data on the health status of tobacco fields, crop growth status, etc. Sensors are arranged in the fields to collect data such as soil moisture, temperature, and air pressure in real time. Meteorological stations provide local climate data to ensure that real-time weather changes are effectively monitored. Unstructured data collection is to collect high-definition images and video streams through drones or ground cameras to record the health of crops, pests and diseases in tobacco fields, etc. And audio sensors are used to collect audio information in the fields, such as the sounds of insects, weather changes, etc., to provide additional clues for subsequent analysis. All collected data are transmitted in real time to a data center or cloud platform via a communication network (such as Wi-Fi, 4G / 5G) for subsequent processing. The data cleaning and preprocessing module processes the collected raw data to ensure the consistency of data quality and format, and prepares it for subsequent analysis. The cleaning and preprocessing steps include: removing duplicate or redundant data. For unstructured data (such as images and videos), noise caused by sensor issues or environmental interference is removed to ensure data quality. Missing data from some sensors can be addressed by filling in missing values, removing missing data, or performing interpolation. Outlier detection is performed on collected data to ensure it is not affected by extreme weather or equipment failure. When fusing multiple data sources, data from different sources is converted to a unified format, such as date and time formats and numerical units (such as temperature and humidity), to ensure that subsequent analysis is not affected by format inconsistencies. Unstructured data such as images, videos, and audio is preprocessed to make it suitable for subsequent deep learning and analysis, such as image resizing, video frame extraction, and audio denoising. After cleaning and formatting, data can be optimized for storage as needed, such as compressed storage and indexed storage, to ensure efficient access during subsequent processing. The data fusion and association module fuses data from different data sources (structured and unstructured) and creates a unified multi-dimensional data model through association analysis. During operation: Data source fusion includes structured and unstructured data fusion. Structured data fusion integrates structured data (such as temperature, humidity, soil quality, and climate) collected from soil sensors, meteorological monitors, and tobacco growth monitoring systems, ensuring that data from different sources can be compared and correlated on the same platform. Unstructured data fusion effectively integrates unstructured data obtained from sources such as images, videos, and audio. Using computer vision and audio analysis technologies, information such as growth status in images, farmland dynamics in videos, and environmental changes in audio is extracted and converted into structured data for subsequent analysis. Data association involves correlation analysis of data at different time points and spatial locations to ensure temporal and spatial relationships between data.Data models are used to combine structured and unstructured data for multi-dimensional correlation. For example, climate data can be correlated with tobacco leaf growth data to identify the impact of climate change on crop growth. Data fusion technology is used to build a multi-dimensional data model, ensuring that different types of data (climate, soil, tobacco leaf growth, images, audio, etc.) can be analyzed collaboratively within a unified framework.

[0045] Deep learning technologies include image processing and analysis modules, audio processing modules, time series data analysis modules, deep learning models, and model training and prediction modules. The image processing and analysis module is used to analyze image and video data; the audio processing module is used for sound classification and recognition; and the time series data analysis module is used to model time series data using LSTM or GRU to predict changes in environmental parameters or crop demand over a period of time.

[0046] Deep learning models are used to integrate information from different data sources into a single model, simulate the results of different strategies, and train the model to provide optimization recommendations for resource allocation and production processes;

[0047] The model training and optimization module uses pre-trained models to quickly migrate to specific agricultural environments or environmental monitoring tasks, and performs data prediction. Data prediction includes crop prediction and pest and disease prediction and early warning. Crop prediction is based on data on soil quality, climate change, and crop growth cycle, using deep learning models to predict crop growth trends, maturity time, and yield. Pest and disease prediction and early warning uses image processing and environmental data analysis, combined with historical pest and disease data, to predict the occurrence of pests and diseases and propose preventive measures.

[0048] It should be noted that the present invention collects images, videos, audio, and environmental parameter data through various sensors, drones, satellites, cameras, and other devices. The collected image data needs to undergo image preprocessing, such as image cropping, scaling, denoising, etc., to ensure that the image quality is suitable for the training of deep learning models. Audio data (such as insect chirping or environmental sounds) undergoes noise filtering and feature extraction to prepare for sound recognition or classification. Time series data (such as meteorological data, crop growth cycle data, etc.) is preprocessed through time series modeling to ensure that the data has sufficient time series information and is suitable for the input of LSTM or GRU models. In the image processing and analysis module, features in the image are extracted through convolutional neural networks (CNNs) to identify crop health status, symptoms of pests and diseases, etc. For video data, key frames can also be extracted through video frame extraction technology for analysis. In the audio processing module, audio data is classified or identified through sound signal processing technology and recurrent neural networks (RNNs) to determine whether there are pests and diseases or abnormal environmental sounds. In the time series data analysis module, time series models such as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) are used to model time series data such as climate change, soil moisture, and crop requirements, predicting future changes in environmental parameters and crop requirements. Deep learning models integrate image, audio, and time series data, fusing information from multiple data sources into a unified model to better understand the interrelationships between the data and provide more accurate predictions. During the model training phase, deep learning models, including image recognition, audio classification, and time series prediction models, are trained using historical and existing data. During training, techniques such as backpropagation and gradient descent optimization are used to continuously optimize model performance. Pretrained models can be quickly transferred to new agricultural environments to more quickly adapt to changes in different regions or environments. This stage is particularly important because climate, soil, and other conditions can vary significantly across regions. Transfer learning can significantly reduce training time and computing resources. Crop forecasting relies on pretrained models and uses real-time data (such as soil quality, climate change, and crop growth cycles) to predict crop growth trends, maturity times, and estimated yields. These predictions can help farmers rationally plan planting, fertilization, irrigation, and other tasks, improving yields and quality. Pest and disease prediction and early warning uses image processing and environmental data analysis, combined with historical pest and disease occurrences, to predict potential pest and disease incidents. Based on model-based predictions, the system can issue early warnings to farmers and provide specific preventive measures, such as pesticide selection and adjusting planting density. By combining crop predictions with pest and disease early warning results, more comprehensive resource allocation and production process optimization can be achieved, ensuring efficient and sustainable agricultural production.Based on the prediction results of the deep learning model and the actual needs of agricultural production (such as water resources, soil improvement, and labor), it provides optimization suggestions for resource allocation. For example, it can adjust fertilization plans based on crop growth prediction results and adjust crop protection measures based on pest and disease warnings.

[0049] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A digital tobacco field production data processing method, characterized by: The following steps are involved: Through the model of satellite remote sensing + drones + sensors + terminal APP + data center, a three-dimensional smart tobacco field management system "Field Sky" is built; Collect structured data and unstructured data from tobacco production in real time in tobacco fields; Through data integration technology, all collected structured data and unstructured data are effectively integrated to form a complete data set; Use deep learning technology to learn and analyze unstructured data to extract valuable information; Through data models and data algorithms, the integrated data is correlated, integrated, and analyzed; The analysis results are fed back to farm workers and decision makers through terminal APP or other management systems, and the production process is adjusted and resource allocation is optimized based on the analysis results.

2. The digital tobacco field production data processing method according to claim 1, characterized in that: The technical framework of the "Field Sky" three-dimensional intelligent tobacco field management system includes satellite remote sensing, drones, sensors, terminal APP and data center. Satellite remote sensing is used to obtain panoramic information and macro data of tobacco fields in real time; The drone is used to provide high-resolution images of tobacco fields and more accurate field data collection; The sensors are arranged in the tobacco field to monitor environmental parameters in real time; The terminal APP is used to provide remote control and observation functions for field workers; The data center is used to centrally store, process and analyze all collected data to build a comprehensive tobacco field data platform.

3. The digital tobacco field production data processing method according to claim 2, characterized in that: The environmental parameters include soil quality, temperature and humidity, and climate conditions. The remote control and observation functions include data viewing, remote operation, and real-time monitoring.

4. The method for processing digital tobacco field production data according to claim 1, wherein: The structured data includes tobacco leaf growth conditions, soil quality, and climate conditions.

5. The digital tobacco field production data processing method according to claim 1, characterized in that: The unstructured data includes on-site images, videos, and audio, which are used to obtain more intuitive production environment information.

6. The method for processing digital tobacco field production data according to claim 1, characterized in that: The data integration technology includes a data acquisition module, a data cleaning and preprocessing module, and a data fusion and association module; The data acquisition module is used to collect environmental parameters and unstructured data; The data cleaning and preprocessing module is used to remove duplicates, clean and standardize the format of numbers; The data fusion and association module is used to fuse multiple data sources and create a multi-dimensional unified data model.

7. The method for processing digital tobacco field production data according to claim 1, characterized in that: The deep learning technology includes an image processing and analysis module, an audio processing module, a time series data analysis module, a deep learning model, and a model training and prediction module; The image processing and analysis module is used for analyzing image and video data; The audio processing module is used for sound classification and recognition; The time series data analysis module is used to use LSTM or GRU to model time series data and predict changes in environmental parameters or crop demand in the future; The deep learning model is used to integrate information from different data sources into a single model, simulate the results of different strategies, and train the model to provide optimization suggestions for resource allocation and production processes; The model training and optimization module uses pre-trained models to quickly migrate to specific agricultural environments or environmental monitoring tasks and perform data prediction.

8. The method for processing digital tobacco field production data according to claim 7, characterized in that: The data prediction includes crop prediction and pest and disease prediction and early warning. The crop prediction is based on data on soil quality, climate change, and crop growth cycle, and uses a deep learning model to predict crop growth trends, maturity time, and yield. The pest and disease prediction and early warning uses image processing and environmental data analysis, combined with historical pest and disease data, to predict the occurrence of pests and diseases and propose preventive measures.

Images