Extraction of building boundaries is a n important step towards 3D buildings reconstruction. It may be also of interest on their own, for the real estate industry, GIS and automated updating of cadastral maps. In this paper we propose a comprehensive method for an automated extraction and delineation of building outlines from raw airborne laser scanning data. The presented workflow comprises three steps. It starts with identification of the points belonging to each singular building. The second step is to trace the points that compose a building boundary. In the last step an adjustment process is applied, that aims in boundary lines regularization. The first step -building detection is a most computationally expensive process and has a fundamental importance for the whole algorithm. A proposed approach is to include building address points that give exact information about building location. This additional information highly reduces the complexity of the building points extraction.