21 October 2016 Aerial vehicles collision avoidance using monocular vision
Author Affiliations +
Abstract
In this paper image-based collision avoidance algorithm that provides detection of nearby aircraft and distance estimation is presented. The approach requires a vision system with a single moving camera and additional information about carrier’s speed and orientation from onboard sensors. The main idea is to create a multi-step approach based on a preliminary detection, regions of interest (ROI) selection, contour segmentation, object matching and localization. The proposed algorithm is able to detect small targets but unlike many other approaches is designed to work with large-scale objects as well. To localize aerial vehicle position the system of equations relating object coordinates in space and observed image is solved. The system solution gives the current position and speed of the detected object in space. Using this information distance and time to collision can be estimated. Experimental research on real video sequences and modeled data is performed. Video database contained different types of aerial vehicles: aircrafts, helicopters, and UAVs. The presented algorithm is able to detect aerial vehicles from several kilometers under regular daylight conditions.
Conference Presentation
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Oleg Balashov, Vadim Muraviev, Valery Strotov, "Aerial vehicles collision avoidance using monocular vision", Proc. SPIE 9988, Electro-Optical Remote Sensing X, 99880T (21 October 2016); doi: 10.1117/12.2241079; https://doi.org/10.1117/12.2241079
PROCEEDINGS
7 PAGES + PRESENTATION

SHARE
Back to Top