ANALISIS DETEKSI TEPI UNTUK MENGIDENTIFIKASI POLA DAUN

Abstract

In our life, we often see plants with various features on them. Plant variety appears from its leaf form such as oval, heart, ellips, etc. Along with progress in information and technology, it is enable to develop an application that can help human to recognize those leaf types. Digital image processing is an evolving field of digital technology with many growing applications in science and engineering. One of the area related to image processing is pattern recognition. This final project is done to build a system which can identify and recognizing a leaf pattern object. Leaf Identification started with image data acquisition, image processing, image edge detection, image thinning, and identification process using template matching method. Edge detection using 3 methods, there are Sobel edge detection, Roberts edge detection, and Prewitt edge detection. Pattern recognition will detect image as input, compare with another images in database called template. The experiment were carried out in two phases i.e leaf form identification and leaf edge identification. This experiment used 7 leaf form images which each form has 10 test images and 5 leaf edge images which each edge has 10 test image. Therefore, there are totally 120 test images. Roberts and Prewitt methods reach 75% for the recognition level of leaf form identification, Sobel method has 74%. Besides that, in leaf edge identification experiment, Sobel method reaches 90% for the recognition level, Roberts and Prewitt only reach 84%. Keywords: leaf, image processing, edge detection, pattern recognition