Derivative Data Products

Woolpert leverages the wealth of information in high-resolution imagery and lidar datasets to extract derivative products and disseminate intelligent geospatial data. We begin by collecting data via a wide variety of aerial, terrestrial and aquatic acquisition methods, including crewed, uncrewed and satellite platforms. Then, we apply a combination of proprietary processes, machine learning and manual verification to extract planimetric features such as buildings, vegetation and other landscape features, as well as how these features change over time.

Learn more about some of Woolpert's turnkey derivative data products:

By combining customized, in-house applications with the latest orthoimagery and lidar datasets, Woolpert can delineate and attribute individual building footprints. Our semi-automated approach provides a cost-effective alternative to more expensive—and time-consuming—manual techniques. Building delineation and attribution for structures approximately 100 square feet and larger are derived from applicable high-resolution orthoimagery along with classified lidar datasets. We begin by identifying sample sets of representative land cover, terrain and built areas across the project area. Then, we develop and apply tailored classification algorithms followed by rigorous statistical and visual analysis techniques to review and validate feature classification accuracy.

Woolpert is experienced in producing elevation-derived hydrography that qualifies for National Hydrography Dataset conflation. Our proprietary workflows combine automated and interactive mapping to produce complete drainage networks from lidar data and derivative breaklines. The result is a high-resolution, accurate 3D hydrographic dataset with direction of flow. Click here for more information on EDH.

Our tailored and automated workflows consistently provide clients with a reliable and economical way to develop and maintain land cover and land use data over a diverse library of landscapes. Woolpert’s process is designed to enable fully repeatable results, that will produce detailed accurate change detection over time.

As with our other feature extraction processes, Woolpert uses machine learning and the fusion of lidar data and other ancillary datasets. We train algorithms with techniques that incorporate lidar elevation data, digital imagery and other datasets to identify land cover types, and then we apply a proprietary blend of pattern recognition and scene reconstruction to deliver extraordinary land use mapping results.

Woolpert has extensive experience extracting planimetric features from orthoimagery and lidar data. Our unique processes were developed using remote sensing and photogrammetry techniques with semi-automated workflows for both commercial and proprietary software. We base the workflow on high-resolution orthoimagery and lidar datasets, using them to enhance the process and perform quality control.

Our experts can extract a wide variety of planimetric feature types, including but not limited to, impervious surfaces (buildings, driveways, parking lots, paved roads and sidewalks), swimming pools, patios and decks, unpaved roads, vegetation groups, waterbodies and waterways.

Use Case: Impervious Surface Mapping
Impervious surface mapping is a necessary component of watershed characterization, hydrologic and hydraulic modeling, community planning and stormwater analysis. We train algorithms with data incorporating elevation, intensity, patterning, spectral signature and shape. We then apply a proprietary blend of pattern recognition and scene reconstruction techniques to validate impervious areas.

Vegetation can be identified and refined into multiple classes using client-specified height ranges after (if) buildings have been classified. Common vegetation classes include low (0.00-1.0 meter high), medium (1.01-3.0 meters high) and high (3.01 and above). Woolpert applies machine learning supervised classification to further refine the classification results. One use of these classifications is the identification of wildfire risks such as invasive species or areas of deadfall trees. With results over 95% accurate (when checked against a robust ground truth network), this data helps land managers develop invasive species mitigation strategies. Data collected at approximately the same time of year over consecutive years supports change detection to monitor invasive species mitigation progress.Use Case: Controlling Invasive Species
Woolpert is mapping cheatgrass and medusahead, two invasive species in western states that contribute fuel for wildfires. The algorithms are trained using a fusion of data inputs (lidar, multispectral imagery, hydrology and historical weather information) along with any available ground truth data. The classification results are reviewed for quality assurance and used to further develop the models.