Aplikasi Penerapan Pengolahan Citra Digital Untuk Pengenalan Objek Menggunakan Pengambangan Metode Otsu

Abstract

This research is conducted to produce an application which can imitate a bit part of human vision system's mechanism by applying digital image processing for object recognition process.   This research produces an application which consists of two main functions that referred as object training and recognition by applying digital image processing methods. The application can be implemented to recognizes simple objects that are represented in two dimensional digital images which are not influenced by the rotation. Based on the conducted testing the writer known some factors which are the major contributive factor for recognition result's accuracy, those are lightning, image acquisition's angle and distance. By considering the analysis result, the writer conclude that the application has an ability to recognizes a simple object which had previously trained, although it is not as perfect as human vision system. For further development, the writer suggest to develop a real time object processing application, minimalize the effects of lightning to image acquisition's quality, application development for recognizing a more complex object's shape and color, and implementation in robotics.