Fast Object Detection for Quadcopter Drone Using Deep Learning

Autor: Aurello Patrik, Andry Chowanda, Alexander Agung Santoso Gunawan, Gaudi Utama, Jarot S. Suroso, Widodo Budiharto
Rok vydání: 2018
Předmět:
Zdroj: 2018 3rd International Conference on Computer and Communication Systems (ICCCS).
Popis: The paper presents our research progress in the development of object detection using deep learning based on drone camera. The grand purpose of our research is to deliver important medical aids for patients in emergency situations. The case can be simplified into delivery of an item from start to the goal position. We will exploit the drone technology for transporting items efficiently. In sending process, our drone must detect the object target, where the items will be delivered. Therefore, we need object detection module that can detect what is in video stream and where the object is by using GPS as well. To implement the module, we use combination of MobileNet and the Single Shot Detector (SSD) framework for fast and efficient deep learning-based method to object detection. The ability of deep learning to detect and localize specific objects is studied by conducting experiments using drone camera and, as comparison, using stereo camera Minoru.
Databáze: OpenAIRE