Description:
OBJECTIVE: To develop an integrated software application that processes small-footprint discrete airborne LiDAR (Light Detection and Ranging) data of terrain for detection and classification of micro-terrain features defined by surface discontinuities and various conditions of surface roughness. DESCRIPTION: Small-footprint, discrete-return LiDAR (Light Detection and Ranging) sensors have become increasingly useful for natural resource management and for enhanced military knowledge of the Battlespace for surveillance and mission planning. These airborne laser scanning systems have been successfully applied to terrain modeling and analysis in such diverse applications as determining forest structure, extraction of urban features, and tracking changes in dynamic coastal environments. High point densities produced by modern commercial scanning systems and high precision Global Positioning System (GPS) data allow for high accuracy in mapping ground features and in representing vertical structure. Frequent scanning missions can reveal newly-formed features or recently changes in their appearance. This capability has not yet been fully exploited for accurately capturing the position, extent, and surface structure of micro-terrain features that often escape mapping efforts. Higher-fidelity knowledge of small-scale terrain irregularities would better inform efforts to map obstacles to cross-country movement and improve mobility models in difficult terrain for military and emergency planning. The high-frequency return pulse information from scanning systems can be processed into point clouds that model the reflective surfaces of ground features as well as the internal structure of vegetative canopies. In addition, LiDAR pulse return positions can be interpolated to create a digital elevation model (DEM). Non-ground pulse returns can be filtered to create a digital terrain model (DTM) of the ground surface. However, in spite of high sampling rates, LiDAR pulse footprints are non-contiguous, resulting in under-sampling of surface features. Yet small-scale terrain features with discontinuous elevation characteristics have been detected and mapped from airborne LiDAR scanning data in the absence of dense forest overstory. This topic seeks to develop a capability to process and interpret small-footprint LiDAR data for the detection of small-scale terrain surface discontinuities as micro-features in open areas and under canopy with elevation differences on the order of a few meters or less, and to categorize terrain areas defined by these features and by the degree and type of surface roughness. The software solution should be able to distinguish vegetated areas from essentially non-vegetated areas. An attempt should also be made to extract the types of micro-features and areas of surface roughness described above in under-canopy settings. Terrain roughness may be described by degree (defined by mean relative elevation change) as well as texture (defined by spatial frequency or azimuthal trending of features). The topic objective includes the ability to represent individual breaklines as linear features that help to define micro-terrain features and regions of surface roughness. The solution should also allow for the extraction and categorization of breaklines associated with man-made structures as distinct from discontinuities on the bare-earth surface. These would be converted to layers for display and analysis in a Geographic Information System (GIS) application. Return intensity values may be treated as a value-added discriminator in the extraction of surface discontinuity features or canopy height models. The LiDAR scanning data may be analyzed as point clouds or as interpolated surface matrices in order to fully exploit the data for the identification and analysis of micro-feature breaklines and regions of surface roughness. The intent is to take full advantage of the LiDAR elevation information in the point cloud or derived gridded models to develop algorithms to detect and extract potentially subtle breaks-in-slope and their azimuthal trends, to understand their inter-relationships or connectedness in the formation of localized micro-terrain features and/or areas of surface roughness as separate features. In the case of voluminous point cloud data, compression strategies may be pursued to reduce the computational load and better facilitate data transfer and storage during processing. The contractor may take advantage of available commercial software for visualizing point cloud data and elevation models such as TerraScan or QT Modeler/Reader. PHASE I: The contractor needs to accomplish two research goals using Government-provided small-footprint LiDAR test data. First, develop a methodology and preliminary software design that would perform detection, classification, and display of surface discontinuities and various conditions of surface roughness. Discontinuities expressed as linear micro-terrain features might include incised stream channels, gullies, small escarpments due to the surface expression of resistant beds or movement along faults, small ridges, or other features defined by breaks-in-slope including those caused by human earth-moving activities. Unpaved roads might be characterized by edge channels or wheel-caused depressions. Regions of varying surface roughness are characterized by extended areas that contain surface discontinuities such as boulder fields, talus slopes, areas of downed trees, heavily plowed fields, or other kinds of disturbed land. The contractor would have to state specifically in this design how he intends to perform these functions and with what (if any) COTS software. While the Army prefers that the proposed solution be compatible with ESRI (ArcGIS) or ERDAS IMAGINE, the Army will consider others. For the second research goal, using test data provided by the Government, the contractor must evaluate his technical approach to perform the detection, classification, and display of surface discontinuities associated with micro-terrain features and various classes or conditions of surface roughness in open areas not under canopy cover. Individual extracted micro-terrain features will have a relative elevation range of 3 meters or less. Areas of surface roughness will also show a mean relative elevation change of 3 meters or less. PHASE II: In Phase II, the contractor will complete the system design and development and integrate the processing capabilities that are defined in Phase I. The contractor will further develop and enhance capabilities developed in Phase I for 3D scene visualization and analysis of small-footprint LiDAR data for the extraction of micro-terrain and surface roughness features, and will integrate the system into an industrystandard GIS or image processing environment compatible with Army collection/dissemination programs such as Buckeye. In addition to processing open areas, the Phase II system will have the ability to perform this extraction under conditions of canopy cover in which less-than-optimal pulse energies are reaching the forest floor, and will be able to distinguish vegetated and essentially non-vegetated areas. Individual extracted micro-terrain features in open areas and in areas under canopy cover will have minimum relative elevation ranges of 1.5 meters and 3 meters or less, respectively. Areas of surface roughness will also show a minimum mean relative elevation change of 1.5 meters and 3 meters or less, in open areas and in areas under canopy cover, respectively. The system will include the ability to display mean azimuthal trending of micro-terrain features and a measure of texture for areas of surface roughness. Testing will occur with as much data as time and budgetary constraints allow. Testing will progress with data provided by the Government. The software prototype must be able to ingest and export standard data formats for imagery and vector data as functions in a 3D visualization prototype for rapid scene assessment. PHASE III: The contractor will create a software product as a standalone application suitable for use on a 32-bit desktop machine or other service-oriented architecture and also integrate it with other commercial GIS software. The final product will be relevant to use by the Army for off-road mobility, mission planning, target detection, and possibly line-of-sight analysis by providing potentially time-sensitive and mission-critical 3D information for rapid decision-making in both open area and under-canopy environments. As timely, high-resolution LiDAR data comes into greater use and becomes more available, this technology will be directly applicable to emergency management functions important to Homeland Security such as post-natural disaster site assessment and disaster relief operations. These functions may include post-flood, post-hurricane, or tornado debris location and volume mapping. Forestry applications might include biomass assessment of forest floor debris or forest crown fuel. REFERENCES: (1) Blakely, R.J., B.L. Sherrod, J.F. Hughes, M.L. Anderson, R.E. Wells, and C.S. Weaver, 2009. Saddle mountain fault deformation zone, Olympic Peninsula, Washington: western boundary of the Seattle uplift. Geosphere 5(2):105-125. (2) Engelkemeir, R.M., and S.D. Khan, 2008. Lidar mapping of faults in Houston, Texas, USA. Geosphere 4(1):170-182. (3) Slatton, K.C., W.E. Carter, R.L. Shrestha, and W. Dietrich. 2007. Airborne Laser Swath Mapping: Achieving the resolution and accuracy required for geophysical research. Geophysical Research Letters vol. 34, L23S10, doi:10:1029/2007GL031939, 2007. (4) Carter, W.E., R.L. Shrestha, and K.C. Slatton, 2007. Geodetic laser scanning. Physics Today, December 2007. (5) Miller, S.N., S.R. Shrestha, and D. Semmens, 2004. Semi-automated extraction and validation of channel morphology from LIDAR and IFSAR terrain data. Proceedings of the American Society for Photogrammetry and Remote Sensing (ASPRS) Annual Conference, Denver, Colorado, May 2004.
