详情

原文传递 Advancing Cots UAV Capability to Provide Vision-Based SA/ISR Data.

题名：	Advancing Cots UAV Capability to Provide Vision-Based SA/ISR Data.
作者：	Teo, W. S.
关键词：	Unmanned aerial vehicles, Computer vision, Off the shelf equipment, Electro-optical sensors, Target detection, Image processing, Search and rescue, Reconnaissance, Systems engineering, Aerial photography, Moving targets, Surveillance, Optical detectors, Unmanned aerial system, Object detection, Optical-flow, Payload sensor, Manifold, Gpu
摘要：	Unmanned systems are gaining popularity in many modern-day applications, and their growth potential in unmanned technologies is infinite. These systems have created research and development opportunities for enabling autonomous behavior to reduce human workload and involvement in tedious operations. This thesis assesses autonomy-enabling technologies for conducting search-and-rescue (SA) operations and intelligence, surveillance, and reconnaissance (ISR) missions using a small unmanned system (sUAS). These technologies include an electro-optical sensor, onboard processor, and computer-vision (CV) algorithms. In a previous masters thesis by Wee Kiong Ang, a commercial off-the-shelf (COTS) quadcopter sUAS was integrated with a suite of hardware and multiple-moving-target-detection software. Building upon that work, this thesis aims to advance the systems capabilities by exploring the applicability of the aforementioned three technologies on an sUAS. Using the systems engineering approach, the baseline system deficiencies are identified first. Next, a technology enabler review is conducted to explore the relevant COTS products and paradigms. Then, through the implementation of a set of changes, the baseline system architecture is reassessed and consequently redesigned, followed by an assessment of state-of-the-art CV algorithms. After being tested in a field experiment based on SA/ISR-type mission scenarios, the developed prototype was found to be successful.
报告类型：	科技报告