Elaeis Guineensis Nutritional Lacking Identification based on Statistical Analysis and Artificial Neural Network

In this study, nutritional disease classification of Elaeis Guineensis or widely known as oil palm is discussed. At present, nitrogen, potassium, magnesium are the main category nutrition deficient prevalent in oil palm plantation and these deficiencies can be identified based on the affected leaves surface appearance. Hence in this work, an alternative method based on image processing technique is proposed for identification of nutritional lacking in Elaeis Guineensis. Firstly, twenty seven features are extracted from three main groups that represent the Elaeis Guineensis leaf surface images namely RGB color features, RGB histogram based texture features as well as gray level co-occurrence matrix attributes. Next, feature selection via ANOVA and Multiple Comparison Procedure is conducted. Further, to verify the effectiveness of feature extraction and feature selection done, ANN is chosen as classifier. Initial findings based on classification accuracy attained confirm that the proposed method is capable to categorize nutritional lacking in Elaeis Guineensis with above 83% success rate prior to statistical analysis and over 86% with ANOVA as subset selection.