Research On Application Of Crop Environmental Information Fusion Evaluation In Guizhou Mountain Blueberry Par

Blueberries are a highly nutritious and valuable fruit that are highly sought after in the consumer market,and have gradually become an important economic crop in China’s agricultural development.The tremendous economic benefits of the blueberry industry have led to widespread cultivation of blueberries throughout the country,with increasing planting areas and yields,and higher incomes for farmers.In order to further improve planting efficiency and reduce environmental pollution,it is necessary to conduct a detailed analysis of the growth environment of blueberries,scientifically evaluate their suitability,and select planting areas rationally.This not only helps to improve the quality and yield of blueberries,but also makes an important contribution to the sustainable development of blueberry cultivation.This thesis takes the crop environment of the mountain blueberry orchard in Majiang County,Guizhou Province as the research object,analyzes and integrates the growth environment factors of blueberries,establishes a blueberry growth suitability dataset,and uses multiple machine learning algorithms to construct a blueberry growth environment suitability evaluation model and divide the suitable growing areas for blueberries.This provides important reference value for the scientific selection of blueberry planting areas.The main research contents of this thesis are as follows:1.Aiming at the problem that blueberry growth is affected by many factors,a data set of blueberry growth environment characteristics was constructed.First,combined with the planting environment information of the mountainous blueberry park in Majiang County,the influence of different environmental factors on the growth of blueberries was analyzed,and the meteorological,topographic and soil environmental factors that affected the growth of blueberries were excavated;then,data information from different sources was collected and analyzed,and each Finally,918 cases of blueberry growth suitability samples in Majiang County were determined to establish a data set.2.To solve the problem that redundant features affect model performance,carry out feature engineering on the original data.First,construct new environmental features through the original basic data,and extract 13 features such as altitude,rainfall,and soil p H;second,preprocess various environmental data,and use two statistical methods of variance expansion coefficient and information gain Filtering feature selection was performed on the environmental features;finally,according to the feature selection results,3 redundant features were eliminated,and 10 most effective and direct blueberry growth environment suitability features were screened out.3.Aiming at the problems of slow convergence speed,low model precision and poor generalization performance of the model constructed by traditional methods,a blueberry growth suitability evaluation model based on environmental information fusion was proposed.First,the Borderline-SMOTE algorithm was used to solve the imbalance problem of the sample data set;then,various environmental information data were integrated,and four machine learning algorithms,including logistic regression,support vector machine,random forest and CatBoost,were used to build the suitability evaluation of blueberry growth environment Models were evaluated and compared to obtain the best model;finally,the influence of different environmental factors on blueberry growth was analyzed by calculating the importance scores of different features.4.In view of the defects of low applicability of the model,the best model is optimized by using the swarm intelligence optimization algorithm to improve its applicability.Firstly,use the particle swarm optimization algorithm,whale optimization algorithm,and sparrow search algorithm to find the optimal parameters of the model respectively;secondly,based on different optimal parameter combinations,respectively build blueberry growth environment suitability evaluation models and conduct evaluation and comparison;finally,according to different models Based on the classification results,ArcGIS software was used to draw the blueberry growth suitability map in Majiang County,and compared with the actual blueberry planting conditions in Majiang County,to analyze the applicability of the best model.