ADVANCED LIDAR DATA PROCESSING WITH LASTOOLS

1 ADVANCED LIDAR data PROCESSING with LASTOOLSC. Hug a,*, P. Krzystek b, W. Fuchsca GeoLas Consulting, Sultenstra e 3, D-85586 Poing, Germany Department of Geoinformatics, Munich University of Applied Sciences, Karlstra e 6, D-80333 Munich, Germany fpi fuchs Ingenieure GmbH, Aachener Stra e 583, D-50226 Frechen K nigsdorf, Germany @fpi WORDS: LIDAR , PROCESSING , Software, DEM/DTM, Classification, Segmentation, Digital, OrthoimageABSTRACT:This paper introduces LASTOOLS , a new software suite for the operational PROCESSING of data from ADVANCED airborne lidarsensor provides the tools required to generate DSMs and DTMs from raw or basically preprocessed LIDAR data in a stand-alone application. It features intelligent management of project data , import and geocoding of raw LIDAR and image data ,system calibration, filtering and classification of LIDAR data , generation of elevation models, and the export of the results invarious common formats.

2 Special emphasis is laid on an intuitive graphical user interface and streamlined workflow to enableefficient and rapid production. Additionally, LASTOOLS provides features for the handling and PROCESSING of ADVANCED lidardata like return signal waveforms and true surface color, as well as the rapid integration of LIDAR and digital image data intodigital paper describes the approach taken by LASTOOLS to efficiently organize and handle the large amounts of data generated byemerging ADVANCED LIDAR systems with waveform sampling and digital image registration. It discusses the segmentation-basedobject-oriented filtering technique at the heart of the LASTOOLS classification engine that achieves a high degree of automationfor LIDAR point classification and feature extraction. The user interface philosophy for rapid data visualization and highlyefficient interactive editing is highlighted.

3 Finally, the instruments for quality control incorporated into LASTOOLS at allstages of the PROCESSING workflow are presented. * Corresponding author. Paper presented at the XXth ISPRS Congress, 12-23. July 2004, Istanbul, Turkey1. MotivationAirborne LIDAR mapping has been gaining widespreadacceptance during the last several years. While there are anumber of established manufacturers offering standard lidarinstrumentation, LIDAR data PROCESSING is mostly still doneusing proprietary in-house developments of the serviceproviders. Obviously, service providers have little interestin making their developments available to the generalmarket, and thus potentially to competitors in their ownfield. Therefore, the market of commercially available lidarprocessing software is to date quite small, with TerrasolidOy of Finnland offering the only LIDAR -specific products thatcover LIDAR PROCESSING in some breadth.

4 Other remote-sensing software packages are beginning to include limitedlidar data support. However, most of these only providetools for the regularization of LIDAR points ( gridding)and for bare-earth filtering/classification but do not addresstypical tasks specific to LIDAR data PROCESSING likegeocoding and strip adjustment. A framework that handlesthe entire production workflow is so far not , we have had to observe shortcomings in userinterfaces of most of the currently available LIDAR processingsystems and LIDAR add-on components to other remote-sensing software packages with regards to intuitivness andefficiency, which put high demands for training andeducation on the potential users and make the productionprocess unduly , new requirements for LIDAR PROCESSING software areon the horizon due to new hardware capabilities (integratedhigh-resolution digital cameras, or waveform digitization).

5 They demand flexible data and PROCESSING structures tohandle the explosively expanding data volumes efficientlyand to make best use of the additional information thatbecomes available through these goal with LASTOOLS is to provide a next-generation integrated LIDAR PROCESSING environment that addresses theseissues and puts an emphasis on ease-of-use and OverviewIn its first incarnation, LASTOOLS handles LIDAR dataprocessing from the egg (raw data ) to the chicken (elevationmodels, orthoimages), integrated with a project managementand quality control PROCESSING features are available: raw data geocoding calibration and strip adjustment model adjustment and ground control classification model generation orthoimage generation datum transformation and projectionOn the project management side LASTOOLS provides workflow management storage resource management quality control documentation and report generation simultaneous multi-user and network capabilities flexible process scheduler for running time-consumingtasks automatically during suitable periodsThe user meets a consistent, easy-to-use interface, including graphical project organizer, process scheduler, andworkflow monitor fast and flexible data viewer giving visual access to thedata in a variety of display modes in multiplesynchronized windows and with efficient displaycontrol mechanisms.

6 Intuitive and efficient data editor with instant visualfeedback for interactive operations comprehensive online help features on several is able to manage, process, and visualize 3D-point,vector, 2D-raster, and volume data , it is ready to handle the output of ADVANCED lidarsystems recording echo waveforms. An open data interfacearchitecture allows LASTOOLS to be easily expanded withadditional PROCESSING capabilities in the future, as well asgiving third-party developers and users with specialrequirements access for customized the scope of this paper, we will focus on two aspectsof LASTOOLS in more detail: project management PROJECT AND data Project Organization and data ManagementA project in LASTOOLS is organized as one or several regions,each of which contains one or more blocks. Blocks areconsidered elementary coherent survey areas.

7 data isacquired in flights (between one takeoff and landing)collecting one or more tracks of data that may cover a singleor multiple blocks. While flights and tracks result fromsequential data acquisition and thus represent the temporalstructure of a project, regions and blocks reflect its spatialorganization. A flight will usually be done with a singlelidar sensor and one set of calibration values. A block may,however, contain data from several flights with differentsensors and calibration sets. LASTOOLS maintains a databasethat relates flights, tracks, blocks, and regions with therelevant sensor descriptions and calibration data 1. Project layoutRegions and blocks are defined by their perimeters (outlinepolygons) that are imported when a project is import of track data (or its creation in the process ofgeocoding), outline polygons of tracks are generated toallow easy access to track-related information at laterprocessing stages when data of multiple tracks has 2.

8 Subblock file layoutThe geocoded LIDAR measurement points are held andorganized in blocks. Blocks are subdivided into squaresubblocks for manageable file sizes (subblocks are th estorage units) which themselves are divided into tiles forrapid location-based access. Subblock files hold in a meta- data structure basic LIDAR data (point coordinates) andattribute information (time stamps, track identifier, returnsignal intensity, surface point color, point class, etc.) inmultiple segments or layers. The sequence of data in alllayers is identical allowing the relevant information of anypoint to be accessed directly and rapidly by indexing fromthe layer start while maintaining a fully bi-directionalrelationship between coordinates and attribute values on thelevel of single measurements. This organization also makesit easy to include additional attributes without having tochange the basic file structure and access example, the inclusion of digitized waveform data fromthe ADVANCED LIDAR systems is easily accomplished as anadditional layer of data .

9 For each laser measurement, avariable-length array of waveform samples plus azimuth andincidence angle values of the beam direction and thecoordinate of the last sample is appended to the subblockfiles. While file sizes increase substantially by includingwaveform data , the layered organization of the files does notrequire the entire file to be read for performing geometricoperations, so PROCESSING speed is not further speed up data access and display at coarser levels,each subblock contains aggregated surface height, heightdifference, intensity and color information at severalresolutions in a data manages subblocks invisibly to the user, who mayselect any area (rectangular, polygonial or linear) within ablock for display and PROCESSING and may pan and zoom thedisplay freely over the entire block without having to workwith subblocks or even encountering their traces in thedisplayed WorkflowLasTools handles all steps of the LIDAR data processingworkflow from project setup to output of products andprovides guidance and support to the user on required steps,production progress, quality control, and report first step in the workflow is project preparation.

10 Itinvolves setting up a project by defining its location in acomputer environment, and creating the requireddirectory structures (which happens automatically); anew project may be created from scratch or as a copy ofan existing project if some of the parameters used in theold project also apply to the new project, importing project-defining data (perimeters, targetcoordinate system(s) and projections, flight plans,products, output formats and tiling requirements,accuracy requirements, tie-points and control surfaceinformation, geoid, etc.), definition of the PROCESSING steps and timeline forprocessing the second step the data to be processed is imported: import of pre-processed trajectory data import of raw or pre-processed LIDAR along with sensordescription and calibration files the import processincludes geocoding of raw LIDAR data , generation oftrack outlines, conversion and storage of LIDAR data inthe meta- data subblock files, import of digital camera images and associatedexposure location and calibration next step covers LIDAR and image PROCESSING .