<![CDATA[Herbal plants have significant importance in traditional medicine and are often useful in various natural health products. Visual identification of these plants is usually carried out based on the shape of the leaves and often encounters difficulties in distinguishing species due to similarities in shape and color. Therefore, a system capable of automatically and efficiently recognizing and searching for herbal plant images is needed. This study aims to implement an image search engine for herbal plants based on leaf color similarity. The method used includes color feature extraction using an HSV (Hue, Saturation, Value) histogram with an 8×8×8 bin configuration, resulting in a 512-dimensional feature vector. This histogram feature is then used as input for the Random Forest classification algorithm to group images based on the type of herbal plant. The dataset used consists of 450 herbal leaf images from 9 different classes, obtained through direct image capture using a digital camera. The test results indicates that the developed system is able to classify types of herbal plants with an accuracy of 95.56%. In addition, the computation time and system response during both training and testing processes are relatively fast and efficient. The advantage of this system lies in the simplicity of feature extraction while still being able to provide high classification performance. This system has great potential to be used as an educational tool as well as an initial component in the development of mobile applications for automatic herbal plant identification.]]>
Copyrights © 2025
