Surveying & mapping
Truth in stormwater billing
Combining 3D lidar with 2D orthophotography produces the most accurate measurement of impervious surfaces.
A fly-by of the ISM process
Woolpert's new impervious surface mapping (ISM) process requires little involvement from your staff. But a solid understanding of how the process works helps to interpret the new mapping data sets and better interpret results.
Step one: data capture
Light detection and ranging (LiDAR) is a remote-sensing technology that uses pulses from an aerial laser-scanning unit to create digital surface models (DSMs). Regardless of elevation and intensity (or density of objects), objects between the LiDAR instrument and the surface itself are captured by the instrument as multiple x, y, z coordinate points. Unlike traditional digital orthoimagery, where tree canopies, structures, and areas of shadow obscure the surface beneath, these models represent a depiction of the surface that was previously unachievable. The laser-based dataset represents only half of the data needed to perform a detailed analysis. (See the table.)
The other critical piece of data required to perform automated feature extraction is digital aerial orthoimagery of the target area. For this process, digital orthoimagery is created in 16 bit for maximum granularity, and is also divisible into a wide range of spectral channels (pan, red, green, blue, near infra-red, true color, and color infrared).
Traditional orthoimagery is created as 8-bit imagery, meaning that each color band has values between zero and 255. Imagery with 16-bit detail has color values ranging from zero to 65,536. While humans cannot see imagery in 16-bit detail, computer systems running remote-sensing software can recognize the subtle differences between the color bands, thus allowing for the additional delineation of ground features.
To create composites of the two different types of data sources, digital orthoimagery is fused with LiDAR datasets to create datasets necessary for classifying every measurable surface within the target area. The near-infrared band contained within the digital orthoimagery is especially useful when delineating characteristics for ISM purposes, such as delineating vegetation from nonvegetation areas.
Step two: analysis and segmentation
Woolpert employs an object-oriented software program (OOSP) that combines the digital orthoimagery and LiDAR datasets to perform automated feature extraction. Compared to previous methods where "heads-up" digitizing of 2D/3D drawings was subject to human error, this step relies on the software (i.e., automation) to perform its analysis. Polygonal segmentation renderings based on pixel groupings with similar values are generated to classify distinct characteristics in the surface.
Before segmentation is performed, rule sets are developed and defined based upon what type of feature is to be extracted and what type of terrain/land cover (i.e., rural or urban) exists within the project area. In this case, remote-sensing specialists used this process to apply multiple rule sets to identify the impervious surface features of the Columbus, Ohio, stormwater service area.
Step three: data delivery
The spectral analysis of the orthoimagery and the geospatial location are fused with the intensity and patterning in the LiDAR to create a polygonal segmentation. This enables multiple options for data creation and delivery. As shown in these parking lot photos, the data from the original AutoCAD files (right) differ greatly from the imagery based on LiDAR data (left).