The Development Of A Data-serving Platform For Tea Industry In Xixiang County

This paper aims to address the development of functional services for tea production and management data management,disease identification,and prevention provided by the tea industry data service platform in Xixiang County.With the rapid development of the local tea industry,various aspects such as tea planting,processing,and sales require a large amount of data support,and traditional manual data management methods can no longer meet the needs.Based on the demand analysis of the actual planting and production process of the Xixiang County tea experiment and demonstration station affiliated to Northwest A&F University,the main contributions include designing a tea industry data service platform that is in line with local characteristics,and enabling production managers to track and guide planting and management,pruning and fertilization,disease identification and prevention,and other aspects.The specific work is as follows:(1)Requirements analysis and design implementation of basic data management functions.In response to the low management efficiency and difficulty in summarizing and statistics data during the production and operation process,based on the local actual production requirements and current situation,the basic data function sub-modules of the Xixiang County tea industry data service platform were designed and implemented.The design includes database conceptual structure design and logical structure design,and the system was determined to be designed and developed based on the MINA framework We Chat end and the Express framework tea industry web management system.The implementation of various functional modules includes login and registration,unit basic information management,operating input management,output management,revenue from various channels,tea garden information,mechanical equipment management,fertilization and pruning data management,etc.(2)Self-construction of tea disease data set and model optimization.In response to the difficulty of local growers and managers in having professional knowledge of tea diseases,and the difficulty in timely prevention,this system is designed to implement the disease identification function sub-module of the data service platform.This sub-module adopts a tea disease identification process based on Mobile Net,including data set collection and construction,development environment construction,data set preprocessing and enhancement,model hyperparameter design and evaluation indicators,model adjustment optimization,training experiments,and comparative analysis.(3)Comparative analysis of tea disease classification and identification models.In response to the characteristics of the system being deployed on a mobile platform,a lightweight network model with a small parameter volume,Mobile Netv3,was selected,and Efficient Net and Mobile Netv2 network models were introduced for comparative analysis to evaluate the effectiveness of the proposed model.In order to enhance the model’s generalization ability due to the single data collection scene,offline enhancement was used to standardize the data,add motion blur,and add Gaussian noise.In addition,model hyperparameter adjustment,replacement of Mini-Batch Stochastic Gradient Descent(Mini-Batch SGD),and adaptive moment estimation(Adam)for learning rate were also performed.Finally,through comparative experiments,Precision,Recall,and F1-Score for five disease classifications(Cloudy Blotch,Tea Red Spot Disease,Tea Brown Spot,Tea Cake Disease,and Tea Anthracnose)all reached over 95%.(4)Tea industry data service platform system testing.In terms of system testing,this article conducted black-box and white-box functional testing for platform functions,including functional testing of the eight modules of basic data management and disease classification and identification,as well as performance testing of the entire platform Mini Program,evaluated the test results,and verified the stability and reliability of the platform.