Leaf and Flower Recognition System (e-Botanist)

Abstract—The objective of this research is to build a computer system that can recognize a plant by using either its leaf or flower image. The system consists of 4 main modules, 1) image acquisition, 2) image preprocessing, 3) image recognition, and 4) display result. In the image acquisition module, the system captures a leaf or flower image, which is taken with a digital camera. The image is taken by using white paper as a background, with a ping-pong ball at a side of the leaf or flower. In the image preprocessing module, the system applies several image processing techniques to prepare a leaf or flower image for the features extraction process. In the image recognition module, the system extracts eight main features from the leaf or flower image and recognizes it by the Euclidean distance algorithm. In the display result module, the system displays the recognition results. The experiment was conducted on 30 kinds of leaves with a total of 1,075 images, and 30 kinds of flowers with a total of 1,075 images. The precision of the system is 76.8 percent and 74 percent for recognizing leaf and flower images, respectively. The average access time for the system is 6.21 seconds per leaf image and 5.69 seconds per flower image.