The industry-focused graduation project under the name of the LIFT-UP program was carried out with the Turkish Aerospace. The project budget was provided by the Turkish Aerospace with the support of the Uludağ University Technology Transfer Office.

Academic Advisor: Assoc. Prof. Dr. Ahmet Emir DİRİK (Bursa Uludag University Computer Engineering - Head of Department), Industrial Consultant: Murat KARADERİLİ (Turkish Aerospace - Flight and Field Operations Chief), Project Student: Mehmet Sergen ERGEN (Bursa Uludag University Computer Engineering Undergraduate Student)

Abstract - Foreign object damage is one of the most important causes of aircraft damage. Foreign object; can be described as physical or economic terms that can negatively affect the safe operation of hardware or performance characteristics. The main purpose of the study is to identify the factors that may cause foreign substance damage and to inform the user. Detection is carried out by scanning the area autonomously or manually by scanning the flight area with an advanced robot equipped with electronic equipment. It obtains the position, shape and length information and sends this information to the server. The algorithm enables the detection of foreign materials by applying deep learning methods. The dataset was created in a unique way and the dataset was increased with data augmentation methods. The dataset was trained with different models. In line with the real-time operation and system requirements, the appropriate model was selected and made ready for use. The robot is autonomously or manually guided. It performs object detection and classification with a success rate of 80%.

Keywords — Foreign Object Damage(FOD), Deep Learning, Custom Object Detection, Custom Data Set, Autonomous Robot