Deep Learning-Based Real-Time Object Detection in Inland Navigation
| Autor: | Metzli Ramirez-Martinez, Philippe Brunet, Wided Hammedi, Mohamed Ayoub Messous, Sidi-Mohamed Senouci |
|---|---|
| Přispěvatelé: | Département de Recherche en Ingénierie des Véhicules pour l'Environnement [Nevers] (DRIVE), Université Bourgogne Franche-Comté [COMUE] (UBFC)-Université de Bourgogne (UB) |
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
Computer science
Object detection 02 engineering and technology Machine learning computer.software_genre Convolutional neural network Domain (software engineering) [SPI]Engineering Sciences [physics] 0502 economics and business 0202 electrical engineering electronic engineering information engineering Training Inland navigation Adaptation (computer science) 050210 logistics & transportation Artificial neural network business.industry Deep learning 05 social sciences Data models Navigation Roads Data set 020201 artificial intelligence & image processing Artificial intelligence business computer Neural networks |
| Zdroj: | 2019 IEEE Global Communications Conference (GLOBECOM) 2019 IEEE Global Communications Conference (GLOBECOM), Dec 2019, Waikoloa, United States GLOBECOM |
| Popis: | International audience; Semi-autonomous and fully-autonomous systems must have knowledge about the objects in their environment to ensure a safe navigation. Modern approaches implement deep learning techniques to train a neural network for object detection. This project will study the effectiveness of using several promising algorithms such as Faster R-CNN, SSD, and different versions of YOLO, to detect, classify, and track objects in near real-time fluvial domain. Since no dataset is available for this purpose in literature, we first started by annotating a dataset of 2488 images with almost 35 400 annotations for training the convolutional neural network architectures. We made this data set openly accessible for the community working on this area. The other contribution of this research is the adaptation and the configuration of deep learning techniques used in other domains such as maritime and road domain to fluvial domain for autonomous vessels in which high accuracy and fast processing are vital. Experiments demonstrated that detecting objects in such environment is plausible in near real time with the selected algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|