representation of road environment understanding for a mobile robot by picture interpretation language PILS

Abstract

Abstract The picture interpretation language PILS·V‐1, which was already publicized, can handle essentially only the silhouette image, since the object and the background are separated in the pre‐processing. The system cannot handle a complex gray‐level image such as an outdoor scene. From such a viewpoint, PILS was improved as follows. The gray‐level image is partitioned into subregions based on the differential image in the pre‐processing. Each region is represented by global parameters such as the coordinate of the center, area, height, width and color. In the recognition stage, the object is recognized from the background, using the a priori knowledge of the object concerning the global parameters. This paper first discusses the representation of the gray‐level color image. Then using the outdoor road environment recognition by moving robot as an example, the programming in PILS for the object recognition algorithm based on the a priori knowledge is described. The result of computer processing is also shown.