Automated People Detection in UAV Imagery
Flash required for video playback in older browsers.
UAV platforms are increasingly being investigated for
search, rescue and surveillance operations both on land and at sea.
Their use allows the autonomous search coverage of large search areas using multiple platforms with minimal operator intervention as a set of remote eyes capable of searching a given area from above.
However, a key remaining problem is the manual analysis of the resulting video footage transmitted back from the UAV platform for “objects of interest” in the search operation.
Automatic detection approaches have to cope with significant motion blur, changes in platform orientation and possible transmission/compression artefacts.
Here we present a novel approach for the real-time automatic detection of people based on using multiple trained classifiers. Our approach facilitates the real-time detection of both static and moving people within thermal imagery.
The results presented show the successful detection of thermal people signature varying conditions in both isolated rural and cluttered urban environments with minimal false positive detection. Notably we target the detection of every person (target object) within the environment over the duration of the flight rather than the traditional object detection paradigm of every person (target object) in every image.
The technique is realised on aerial imagery obtained at 1Hz from an optical camera on the Blue Bear Systems Research medium UAV B-MAV platform The video below shows real-time people detection from an imagery stream transmitted back live from the UAV platform (including frame drop-outs).
Overall we see an ability to identify thermal people signatures over a range of different environments and automatically alert a human operator to their presence (and position) within the search environment. Future work is investigating the secondary tracking of people in UAV imagery.
This work formed part of the Stellar Team SATURN system that was awarded the R.J. Mitchell Trophy in winning the 2008 challenge. Stellar Team comprises of Cranfield University, Blue Bear Systems Research, Selex Galileo, Marshall System Design Group, TRW Conekt and Stellar Research Services.
|||Real-time People and Vehicle Detection from UAV Imagery , In Proc. SPIE Conference Intelligent Robots and Computer Vision XXVIII: Algorithms and Techniques, Volume 7878, No. 78780B, 2011. [demo]Keywords: UAV image analysis, people detection, aerial image analysis. [bib] [pdf] [doi]|