基于机器学习的土壤机械压实应用研究

Research and Application of Soil Compaction Based on Machine Learning

农业机械作业造成的土壤压实已成为制约农业可持续发展的重要因素,过度机械压实使土壤理化性质恶化,甚至成为降低作物产量的主要原因。已有的土壤机械压实研究多是基于简单的数学统计分析,且研究重点为试验方案的设计,无法挖掘数据内部关系,也无法进行土壤机械压实程度的预测。近几年,随着机器学习的不断发展,越来越多的学者开始将其引入农业领域及土壤机械压实的研究。为此,分析了机械压实对土壤理化性质及作物生长的影响,总结了土壤机械压实的表征属性和常用机器学习算法及评价标准,并归纳了近几年基于机器学习的土壤压实的研究成果,给出了相关应用研究的建议。

Soil compaction caused by agricultural machinery operations has become an important factor restricting the sustainable development of agriculture.Excessive mechanical compaction worsened the soil physical and chemical properties of and even became the main reason for reducing crop yield.Most of the existing soil compaction studies are based on simple mathematical statistical analysis,and the research priority is the design of the test scheme,so the internal relationship of data cannot be mined,and the degree of compaction cannot be predicted.Recently,with the continuous development of machine learning,more and more scholars have begun introducing it into the agricultural field and soil mechanical compaction research.In this paper,the effects of mechanical compaction on soil physical and chemical properties and crop growth were analyzed,and the characterization properties,common machine learning algorithms and evaluation criteria of soil mechanical compaction were summarized.Finally,the research results of soil compaction based on machine learning recently are summarized,and some suggestions for related application research are given.