|
Time and spatial variations in atmospheric aerosol plumes can be factors in the effectiveness of electro-optical (EO) systems. Of particular interest are obscuration statistics including the frequency and duration of optically thick and optically thin cloud regions that intermittently affect sensor operation. Related to this problem is an increasing use of modern visualization techniques in simulations and computer graphics to portray realistic looking clouds. It is obviously important that such visualizations accurately model obscuration statistics if they are to be used in EO design and evaluation. Most plume models in use are either based on well-established, physical descriptions of mean aerosol transport and diffusion, buoyancy and radiative transfer, or they are based on modern artistic computer image texturing to provide variations in cloud appearance. The former, traditional methods visualize aerosol plumes as smooth distributions resulting from time-averaged contributions of turbulence to the mean flow. The latter computer graphics techniques visualize instantaneous spatial and time variations. However, the graphics techniques are usually devoid of much physics, using only a subjective criteria of 'looking realistic' as a validation basis. This paper begins a process of reconciling the two approaches. A method for texturing the mean plume concentrations is examined that has much in common with fractal- based texturing in computer graphics. The method also includes time-dependent evolution of eddy phases within the plume. Results of actual transmittance measurements of smoke and dust plumes are presented that show an apparent measured change in fractal dimension or Hurst parameter with increasing distance or time separations. This behavior is discussed in terms of contributions from different eddy sizes to the path integrated concentration. It is then shown that the fixed fractal dimension of the concentration texturing scheme can be reconciled with the changing dimension in the transmittance data if one is careful to modify the texture model to account for the fact that we see real plumes by the transmittance of the background radiance through the cloud and from the path radiance produced by scattering or emission from the cloud itself. Additional efforts to analyze images to include the scattering and emission contributions to cloud statistics are briefly discussed.
|
