Indonesia is one of the state that has many
plants in the world. Indonesia has tropical climate, has many vitals than
another subtropical and pole. Various types of plants are found in Indonesia,
including types of medicinal plants. Around 9,600 types of medicinal plants are
found in Indonesia, but only about 300 species are used as ingredients for
traditional medicines. . As time goes by, the research about herbal is still
developed up to now. It's done because many kinds of herbal but people still
don't know about it. The process of identifying plant species is very dependent
on the knowledge of botanists with manual methods that rely on the sense of
sight based on morphological characteristics. With advances in technology, leaf
image recognition can be done using computer vision. By the technology, someone
can easily recognize kinds of herbal. The process can be applied to many parts
of plants, like fruit, flower, seed, leaf, or others. In this study, it is
proposed to identify medicinal plant species based on leaf image patterns. This
identification process uses the KNN (K-Nearest Neighbors) method. In this
experiment, 15 types of medicinal plant leaves were used with 20 different
images each to be used as training data. Before the testing project, the first
process is preprocessing and feature extraction. The identification process
uses the KNN method using a K value of 10 and the Euclidean distance. After
testing on 75 leaf images, 3 unsuitable leaves were obtained and 73 leaves were
identified correctly. From the experiments conducted in this study, the
percentage of success of 96% was identified.
Keywords— Identification, image
processing, K-Nearest Neighbors, Leaves Image, medical plant