Using Near-Infrared Aerial Imagery to Map Oak Trees

Caitlin Dempsey


The first ordinance passed by the California city of Santa Clarita after it was incorporated was the Oak Tree Ordinance.  Thousands of oak trees cover the Santa Clarita Valley and the presence of the trees is an important part of the city’s landscape (the city logo prominently displays an oak tree).  The ordinance covers oak trees of the Quercus species native to the area which includes Valley Oak, California Live Oak, Canyon Oak, Interior Live Oak, and Scrub Oak.   As well as regulating the pruning, encroachment and removal of oak trees, the city sought to protect “heritage oaks” under the ordinance.  Heritage oaks are the largest and oldest oaks in the city and are defined as oaks measuring at least 108 inches in circumference for a single trunk, or 72 inches in diameter for multiple trunks, at 4 1/2 feet above the ground.

Having been incorporated in 1987, the city of Santa Clarita is still experiencing significant growth and development.  Development plans submitted to the city have to be carefully scrutinized to make sure they are in compliance with the Oak Tree Ordinance.  Because of this, it had long been a priority of the city’s GIS group to develop a geographic layer identifying the location of oak trees within the city.  When the GIS group began participating in Los Angeles County’s inaugural aerial imagery acquisition program (called LAR-IAC), the opportunity arose to experiment with remotely sensing tree locations.

After reviewing different academic and commercial options, the project manager, Edgardo David, along with the rest of the GIS team (Kristina Jacob, Anthony Calderon, and Kelly Minniti) opted to contract with the Los Angeles firm Engineering Systems.  The goal of the project was to utilize 4-inch resolution near-infrared aerial imagery to extract the locations of specific oak tree species within the city.  The challenge was in using aerial imagery to develop a spectral signature for the oak trees that could be automated in order to extract all tree locations.

For the pilot study, oak tree locations were identified by the city’s arborist.

A pilot project was developed that involved the identification of oak trees by the City’s arborist for a single aerial tile.  Staff at Engineering Systems then used those marked locations to create a polygon layer in AutoCAD of all oak trees present.  Engineering Systems then developed an application using Microsoft .NET that scanned the TIFF image (a 000 x 8000 4-inch pixel grid comprising 64 million pixels).  Those pixel within the polygons were extract and were analyzed to prepare histograms to represent the frequency of individual Red, Green and Blue (RGB) values to determine the peak values representing the spectral signature of the oak trees.

Free weekly newsletter

Fill out your e-mail address to receive our newsletter!

Three histograms representing frequencies of Red, Green, and Blue values.

The spectral signature was then used as input parameters for a second application that scanned the TIFF image tile and extracted pixels that matched the signature.  The process went through several iterations matched against field surveys to verify that the correct species of trees were being selected.  The overall analysis found that the spectral signature was accurate in identifying more mature oak trees but younger trees with smaller canopies were not being identified.  Engineering Systems is working on refining the process to be able to identify those younger trees.  Over 166,000 trees were located using this automated process.

The resulting geographic layer also identifies the diameter of the oak tree canopy.  Since the oak tree locations are now georeferenced, the oak trees were spatially identified with the parcel number.  This now allows the staff within the various city departments to know when a property has an oak tree and is subject to the constraints of the Oak Tree Ordinance when plans are submitted by developers and property owners.

Identified oak tree locations.

See Also

Photo of author
About the author
Caitlin Dempsey
Caitlin Dempsey is the editor of Geography Realm and holds a master's degree in Geography from UCLA as well as a Master of Library and Information Science (MLIS) from SJSU.