基于Apriori关联规则算法的草莓叶片含水状况研究

Analysis of Association Rules of Strawberry Leaf Moisture Content Based on Spectral Reflection Characteristics

为了更便捷准确的测量草莓叶片含水率,利用叶片含水率分析指导草莓在各个生长周期的灌溉。该文通过研究数据挖掘算法中的Apriori关联规则算法及其在分析草莓叶片水分状况方面的应用,提出一种基于Apriori算法的D算子模型,采集草莓叶片不同水分状况的光谱反射值,建立相关性系数和出错率2个衡量指标,利用所提算子模型对冠层叶片含水率与不同波段的光谱反射率之间的潜在关系进行关联规则分析,找出了干旱、轻度干旱、适量和溢水4种不同的水分处理方式下草莓光谱反射特征与叶片水分状况之间的相关关系。试验结果证实,在2037.425~2560.039 nm波段下,通过光谱反射率就能判断出叶片的含水状态,即当光谱反射值在101.554~101.560时,草莓叶片含水状态为干旱;当光谱反射值在98.574~99.140时,草莓叶片含水状态为轻微干旱;当光谱反射值在96.379~97.587时,草莓叶片含水状态为适量;当光谱反射值在94.583~95.902时,草莓叶片含水状态为溢水。实证分析结果表明,所提算法简化了实际操作的步骤,节省数据空间,证明了所提算法分析草莓叶片含水状况的可行性和有效性。

In order to measure the moisture content of strawberry leaves more conveniently and accurately,the analysis of leaf moisture content was used to guide the irrigation of strawberries in each growth cycle.By studying the Apriori association rule algorithm in data mining algorithms and its application in analyzing the moisture status of strawberry leaves,a D operator model based on the Apriori algorithm is proposed.Collect the spectral reflectance values of different water conditions of strawberry leaves,two correlation indexes of correlation coefficient and error rate were established.The proposed operator model was used to correlate the potential relationship between the water content of canopy leaves and the spectral reflectance of different bands.The rule analysis found out the correlation between the spectral reflectance characteristics of strawberry and leaf water status under four different water treatment methods:drought,mild drought,moderate and overflow.The results confirmed that the water state of the leaves could be judged by the spectral reflectance in the band of 2037.25-2560.039 nm,that was,when the spectral reflectance was between 101.554 and 101.560,the water content of the strawberry leaves was drought;when the spectral reflectance was 98.574.When the temperature was between 98.574-99.140,the water content of strawberry leaves was slightly drought;when the spectral reflectance was between 96.379-97.587,the water content of strawberry leaves was moderate;when the spectral reflectance was between 94.583-95.902,the water content of strawberry leaves.The empirical analysis of the overflow showed that the proposed algorithm simplifies the actual operation steps,saves the data space,and proves the feasibility and effectiveness of the proposed algorithm for analyzing the water content of strawberry leaves.