Integration of LiDAR Data and Orthophoto for Automatic Extraction of Parking Lot Structure

摘要

To overcome the challenges of parking lot structure extraction using optical remote sensing images, this study proposes an automatic method for the extraction of parking lot structure by integrating LiDAR data and orthophoto, which consists of three steps. The first step is to extract vehicles from LiDAR data and then to identify the corresponding central axes for each vehicle. In the second step, orientations of the identified vehicle central axes are used as principle orientation constraints for parking lines extraction from orthophoto. The third step is the determination of parking lot structure with vehicle central axes and parking lines, in which parking lot parameters are calculated and an adaptive growth method is used for parking lot structure determination. In this method, vehicle central axes identified from LiDAR data and parking lines extracted from orthophoto are integrated for the extraction of parking lot structures. The main novelty of this study lies in two new algorithms: an algorithm on parking lines extraction with principal orientation constraints and an algorithm on parking lot structure determination based on parameter solution and adaptive growth. The experiment shows that the proposed method can effectively extract parking lot structure with high correctness, high completeness, and good geometric accuracy.