Customer Stories

SOLUTION As USAREUR had commissioned IABG, a geospatial technology company based in Ottobrunn, Germany, to update a LiDAR dataset of the HTA in 2012, it tasked IABG to produce the land-classification maps. An IABG team acquired a 1-m-resolution, LiDAR-derived DEM from 2007 and the 0.5-m-resolution LiDAR-derived DEM from the 2012 flight. They also procured 8-band, 1-m-resolution optical imagery from the Worldview-2 satellite, existing aerial photos, and ancillary datasets. That data was integrated into the eCognition object-based image analysis software to build a customized rule set. To bring the HTA down to manageable size, the team initially chose four, 0.8 x 0.8-sq-mi (2 x 2-sq-km) test sites to classify. After pre-processing and validating the raster data quality, they calculated a Normalized Vegetation Index and texture layers to separate vegetation from non-vegetation areas— detail that would be integrated into the classification process. They then wrote rules to instruct eCognition to distinguish Blackthorn from other vegetative types based on height, spectral qualities and textural features. Although the rule set took months to build, it only took 15 hours to run the workflow and produce land-use maps for 2007 and 2012, indicating the bush's growth in each test site between those two years. The IABG team presented the preliminary results to the JMRC to validate the data on the ground and to show the leadership how the data could be used for building vegetation-management strategies. For the field verification, they chose 40 different Blackthorn bushes in each site and measured them using a yardstick, comparing their real- world height and shape with their classified counterpart on the map. There was not a single mismatch between what eCognition classified as Blackthorn (including its varying heights) and what was on the ground. Based on the quality of the test sites classifications, the classification methodology was extended across the entire site. And in September 2013, IABG delivered the classification results to the Center's leadership, showing that Blackthorn was highly impacting one-third, or about 50 sq km (19 sq mi), of the training area. A subsection of the eCognition classification. IABG integrated LiDAR-derived DEMs, optical satellite imagery, aerial photos and other data into eCognition to distinguish Blackthorn from other vegetative types. There was not a single mismatch between what eCognition classified as Blackthorn (including its varying heights) and what was on the ground.

• Cover