Galyna Kriukova
National University of Kyiv-Mohyla Academy Kyiv, Ukraine
Adaptive sensor multi-parameter selection
Abstract:
Motion detectors have found wide use in commercial applications, e.g. surveillance video
cameras with infrared night vision, passive infrared sensors (PIR), thermal cameras, and
radar sensors. Recently, the radar-based sensors have been employed in various active safety
applications and robotics, as low power and cost effective devices for motion detection,
able to track and classify a number of objects simultaneously and estimate corresponding
distances and shapes. Sensor fusion of radar-based sensor and video-camera enables rich
data stream for object detection, tracking and classification.
In the talk we discuss motion zone estimation for static multi-sensor camera and following
tuning of sensor parameters. The approach implies inverse problem requiring multi-parameter
regularization. We discuss interconnection between sensor’s parameters and corresponding
parameters of inverse problem. Methods for de-noising, tracking and motion zone estimation
are considered. We illustrate the approach with a few real-world applications and experiments.
<< back