Generation a adaptation of transferable roadway model in naturally complex scene

Abstract

To expand the scope of humanspsila perception on satellite-roadway-vehicle network, roadway images in encountered scene are matched with birdpsilas eye view. Based on the geometric consistency of the satellite images with as-is local terrain, in this paper, a computational mechanism is presented for generation and adaptation of roadway model transferable through the network. Laplacian-Gaussian filter is applied to the distribution of random texture in the scene image for generating a fractal model of the perspective of the roadway area. The structure of chromatic complexity is extracted via fractal sampling of the scene image and adapted to the associated area of the birdpsilas eye view. By transferring the chromatic complexity the roadway pattern is segmented arid extended in the birdpsilas eye view. Through experimental studies, roadway patterns were demonstrated to be extended beyond the perspective in the scene images prior to physical arrival.