SISTEM PENGELOMPOKAN PADA DAUN BERDASARKAN SPESIES DENGAN SEGMENTASI DAN CONVOLUTIONAL NEURAL NETWORK (CNN)

Abstract

Abstract
Identifying plant species based on leaf morphology manually requires specialized expertise, precision, and
considerable time, especially due to the variation in shape and size of leaves across different species. This
challenge often slows down the conventional process of plant identification. In this research, an automated
approach is utilized by leveraging Convolutional Neural Networks (CNN) and image segmentation techniques to
accurately classify leaf species. The image segmentation process is carried out using thresholding and morphology
methods, which aim to separate the leaf object from the background, thereby enabling better identification of key
features of leaf morphology. Once the leaf object is clearly segmented, a CNN model is employed for
classification. The CNN works by applying several convolutional layers to extract local features from the leaf and
uses pooling layers to reduce the data dimensions without losing relevant information. A diverse dataset of leaf
images is used to train the CNN model so that it can recognize the morphological patterns of leaves from various
plant species. The research findings demonstrate that this method achieves a high accuracy level in classifying leaf
species. This indicates that the combination of image segmentation techniques and CNN is an effective and
efficient approach to speeding up and enhancing the accuracy of plant species identification. This approach offers
a reliable solution in the field of botany without requiring complex manual intervention.