Leaf Morphology Classification Using Customized Convolutional Neural Networks

Abstract

Leaf morphology is a key trait in plant identification and taxonomy, crucial for understanding biodiversity and agricultural studies. Manual classification methods are often labor-intensive and subjective. Machine learning approaches, particularly convolutional neural networks (CNNs), offer automated and precise solutions. To classify plant leaves based on their morphology using CNNs, leveraging image datasets to develop an accurate and scalable model. The Plant Leaves for Image Classification dataset from Kaggle, consisting of leaf images from multiple species, was utilized. Data preprocessing included resizing, normalization, and augmentation. A CNN model was trained and optimized for classification tasks. The proposed CNN achieved an accuracy of 94.85% on the test set. Sample classifications demonstrated the model's ability to distinguish between similar leaf structures effectively. This study highlights the potential of CNNs in automating leaf morphology classification, contributing to advancements in botany and agriculture.