Moving towards a Network of Autonomous UAS Atmospheric Profiling Stations for Observations in the Earth's Lower Atmosphere: The 3D Mesonet Concept.

The deployment of small unmanned aircraft systems (UAS) to collect routine in situ vertical profiles of the thermodynamic and kinematic state of the atmosphere in conjunction with other weather observations could significantly improve weather forecasting skill and resolution. High-resolution vertical measurements of pressure, temperature, humidity, wind speed and wind direction are critical to the understanding of atmospheric boundary layer processes integral to air-surface (land, ocean and sea ice) exchanges of energy, momentum, and moisture; how these are affected by climate variability; and how they impact weather forecasts and air quality simulations. We explore the potential value of collecting coordinated atmospheric profiles at fixed surface observing sites at designated times using instrumented UAS. We refer to such a network of autonomous weather UAS designed for atmospheric profiling and capable of operating in most weather conditions as a 3D Mesonet. We outline some of the fundamental and high-impact science questions and sampling needs driving the development of the 3D Mesonet and offer an overview of the general concept of operations. Preliminary measurements from profiling UAS are presented and we discuss how measurements from an operational network could be realized to better characterize the atmospheric boundary layer, improve weather forecasts, and help to identify threats of severe weather.

Adaptive Kalman filtering for histogram-based appearance learning in infrared imagery.

Targets of interest in video acquired from imaging infrared sensors often exhibit profound appearance variations due to a variety of factors, including complex target maneuvers, ego-motion of the sensor platform, background clutter, etc., making it difficult to maintain a reliable detection process and track lock over extended time periods. Two key issues in overcoming this problem are how to represent the target and how to learn its appearance online. In this paper, we adopt a recent appearance model that estimates the pixel intensity histograms as well as the distribution of local standard deviations in both the foreground and background regions for robust target representation. Appearance learning is then cast as an adaptive Kalman filtering problem where the process and measurement noise variances are both unknown. We formulate this problem using both covariance matching and, for the first time in a visual tracking application, the recent autocovariance least-squares (ALS) method. Although convergence of the ALS algorithm is guaranteed only for the case of globally wide sense stationary process and measurement noises, we demonstrate for the first time that the technique can often be applied with great effectiveness under the much weaker assumption of piecewise stationarity. The performance advantages of the ALS method relative to the classical covariance matching are illustrated by means of simulated stationary and nonstationary systems. Against real data, our results show that the ALS-based algorithm outperforms the covariance matching as well as the traditional histogram similarity-based methods, achieving sub-pixel tracking accuracy against the well-known AMCOM closure sequences and the recent SENSIAC automatic target recognition dataset.