Object Detection Methods for Improving UAV Autonomy and Remote Sensing Applications
Panagiotis Aposporis
Abstract
The last decades the Unmanned Aerial Systems (UASs) are being used in a variety of applications, such as civil protection, security, agriculture, armed forces, that need real time object detection of observed information by their sensors. Moreover, the development of fully autonomous UAS is heavily dependent on their capability to detect and track steady or moving objects in a robust, powerful and reliable manner. In this review, we present a comprehensive literature survey and discussion on object detection methodologies for improving UAV autonomy and remote sensing applications. Emphasis is placed on Convolutional Neural Networks (CNN) implementing different object detectors and exploiting cloud processing. Based on these works, we provide a brief discussion and summary of related proposals for UAV-based object detection using different methodologies and approaches, share views for future research directions and draw conclusive remarks.