Multiple Object detection using Autonomous UAVs based on YOLOv3 algorithm

• Published in: IOP conference series. Materials Science and Engineering 2020-09, Vol.925 (1), p.12025
• Main Authors: Murugendrappa, N, Gayathri, KE, Harshitha, GM, Gupta, Sony
• Format: Article
• Language: English
• Subjects: Algorithms; Convolution Neural Networks (CNNs); Graphics Processing Units (GPUs); Image quality; Inspection; Object recognition; Regions with Convolutional Neural Network (RCNN); Unmanned Aerial Vehicles (UAVs); You Only Look Once (YOLO)
• ISSN: 1757-8981; 1757-899X
• DOI: 10.1088/1757-899X/925/1/012025

Currently, there has been an enormous increase in interest towards autonomous UAVs and applications with respect to autonomous UAVs such as scientific data collection and security monitoring, infrastructure inspection, search and rescue and much more. Recognizing objects visually is a key step in such applications and is very pivotal to build a complete autonomous system. Nevertheless, the chore of detecting an object is very much challenging and is even more difficult when we have low quality images aboard low cost consumer UAVs. More frequently due to the motion of UAVs, images get blurred, noisy and even onboard cameras have low resolution hence object detection becomes more challenging since identifying objects are quite small and these tasks get even more difficult as there is a necessity of real time detection of objects. This paper focuses on all these aspects and brings an outcome for the object detection and identification.