Enhancing virtual natural scenes using quick and dirty image based recipes

Computer generated scenes are getting popular in the scientific communities concerned by plant, crop and forest structural and functional modeling. We show that realistic images can be produced from scientific visualizations at low costs, with post-processing functions applied to output images with or without their associated depth maps. We illustrate the use of empirical quick and dirty image based functions to mimic weather conditions and photographic effects. 2D particle systems, depth gradient processes, image with depth maps composing, allow to mimic rain, snow deposit and depth of field effects. Giving moreover a viewing matrix, on restrictive hypothesis related to the view and the illumination, images can be added sky-domes, fake reflections among other process. Compared to existing tools (often complex), the application of these experimental recipes shows fast dedicated parametrization, easy to handle and apply, covering a wide range of applications from single plant to landscape levels, including small forest stands, at various seasons and day times.