Derin ögrenme tabanlı nesne takibi
Date
2019-09-18
Authors
Başarır, Bilen
Journal Title
Journal ISSN
Volume Title
Publisher
Bursa Uludağ Üniversitesi
Abstract
Bu tezde, derin ögrenme tabanlı SSD (Single Shot Multibox Detector) algoritmasını kullanarak, hareket eden kisileri takip eden ve bir lazer isaretçi ile hareket eden kisiye nisan alan bir sistemin tasarlanması ve gerçeklenmesi amaçlanmaktadır. SSD yöntemi nesne tespit konusunda literatürdeki en basarılı yöntemlerden biridir. Gelistirilen sistemin nisangahının yatay ve dikey hareketleri 2 adet adım motoru ile kontrol edilmektedir. Gelistirilen sistemim performansı ve isabetli vurus istatistikleri deneysel testlerle ölçülmüstür. Bulunan sonuçlar gerçek zamanlı olarak bilgisayar ortamında kaydedilerek akabinde sonuçlar istatistik olarak yorumlanmıstır.
In this thesis, using deep learning based SSD (Single Shot Multibox Detector) algorithm, it is aimed to design and implement a system that follows and aims the moving people a laser pointer. SSD method is one of the most successful methods of object detection in the literature. The horizontal and vertical movements of the developed system are controlled by 2 step motors. Improved system performance and accurate hit statistics were measured by experimental tests. The results were recorded in real time in a computer environment and interpreted as statistically.
In this thesis, using deep learning based SSD (Single Shot Multibox Detector) algorithm, it is aimed to design and implement a system that follows and aims the moving people a laser pointer. SSD method is one of the most successful methods of object detection in the literature. The horizontal and vertical movements of the developed system are controlled by 2 step motors. Improved system performance and accurate hit statistics were measured by experimental tests. The results were recorded in real time in a computer environment and interpreted as statistically.
Description
Keywords
Evrişimsel aglar, CNN, Deep learning, Derin ögrenme, SSD, Nesne tanıma, Adım motoru, Arduino, Object detection, Step motor
Citation
Başarır, B. (2019). Derin ögrenme tabanlı nesne takibi. Yayınlanmamış yüksek lisans tezi. Bursa Uludağ Üniversitesi Fen Bilimleri Enstitüsü.