Classifying Chilli Plants Using Digital Images And Multiple Linear Regression

Abstract

The present study focuses on the application of classification methods using digital images and multiple linear regression to identify types of chili plants based on texture and shape features extracted from leaf images. In the process, digital images of chili plants undergo a pre-processing stage to enhance image quality, followed by feature extraction using methods such as the Gray-Level Co-occurrence Matrix (GLCM). The present study utilised 100 datasets of chili plant images obtained from the BRIN website, which were then divided into training data and test data to train a multiple linear regression model. However, the findings of the study indicated that the multiple linear regression model was not adept at encapsulating the intricacies of the data, as evidenced by the negative R-squared value and substantial prediction errors. Consequently, it is recommended that dimensionality reduction and cross-validation techniques be applied to enhance model performance and increase accuracy in classifying chili plant types in future.