Water classification using 3D airborne laser scanning point clouds

Airborne laser scanning (ALS), also referred to as airborne LiDAR (Light Detection And Ranging), provides highly accurate measurements of the Earth surface. In the last twenty years, ALS has been established as a standard technique for delineating objects (e.g. buildings, trees, roads) and mapping changes. Studies on hydrology or geomorphology such as monitoring of braided river structures, calculation of erosion and accumulation potential in watercourses, or floodplain mapping require all the precise location of the water surface. This paper shows a 3D point cloud based method, which allows an automatic water surface classification by using geometric and radiometric ALS information and the location of modeled lost reflections, which are called laser shot dropouts. The classification result can be used to map the watercourse, to improve DTM filtering routines or to replace water points with river bed heights for hydraulic modeling etc. The method relies on a threshold based classification using geometry and radiometric information of the 3D point cloud. The method is divided into five major steps. First, we correct the amplitude values by reducing the atmospheric and geometric influences to the laser shots. A radiometric adjustment was applied to the amplitude values of the data sets,which allows multi-temporal analysis of the amplitude values. The second step is the interpolation of the coordinates of the laser shot dropouts, which are the most important input to delineate water surfaces. In step three and four the two attributes (standard deviation of height values and the amplitude density ratio value) are calculated at a fixed distance to each reflection and dropout. These are used in step five to distinguish water and dry land points. The exploration of the attributes for the classification and the evaluation of the classification results are done by comparing the results to a terrestrial orthophoto mosaic an dGPS measurements, which were taken simultaneously to the ALS campaign. One of the major tasks is the use of modeled laser shot dropouts within a threshold based classification method to distinguish water and non-water echoes. The method is also suited to detect water under riverine vegetation, which is problematic by using data from sensors, that are not able to penetrate vegetation. The classification accuracy is about 95%. The achieved amplitude correction and the radiometric adjustment make the data sets comparable and allow to calculate changes in the channel flow paths within the different flights.