Eye in the Sky Details Urban Tree Landscape

Prof's Imaging Technique Enables Comprehensive Study of Dallas’ Forest Areas

Dec. 8, 2009

For decades, arborists and urban planners alike have struggled with how best to evaluate and maintain Dallas’ urban forests. A vital source of greenery, environmental protection and quality of life, these forests were previously cataloged by volunteers on foot with handheld GPS devices. At best, the results were a rough, inaccurate patchwork portrayal of the variety and number of trees in a given area.

Dr. Fang Qiu

Dr. Fang Qiu, associate professor in geospatial information sciences, has produced a survey of portions of the Dallas forest.

A new technique pioneered by Dr. Fang Qiu, associate professor in geospatial information sciences in the School of Economic, Political and Policy Sciences, has produced a comprehensive survey of some of the Dallas forest — and he did it leaf by leaf. The effort marks the first time any metropolitan area has surveyed trees with the Qiu’s novel, two-step approach.

Qiu was commissioned by Steve Houser, who chairs the Dallas Urban Forest Advisory Committee.

“Trees clean our air, our water, and our soil,” Houser said. “They also greatly add to our health, our sense of well-being, our quality of life, and to our economic future. We need one thing to manage our urban forest — we need an accurate inventory.”

Houser, a longtime arborist and urban forest advocate in Dallas, learned about Qiu’s research from prior work that the scholar performed for the North Central Texas Council of Governments. The Urban Forest Advisory Committee charged Qiu with finding an accurate, speedy and reliable means to determine how many trees are in a given area, which varieties are present, the precise location of each tree — by GPS coordinates — and the size of the trees overall. Previous survey methods didn’t account for lesser, immature trees, and volunteers on foot lacked the tools or proper training to accurately measure a tree’s maturity, height or health.

“In all, we surveyed 20 square miles — a small portion of Dallas that included the Turtle Creek Corridor and Reverchon Park,” Qiu said. “A prior survey conducted for the Turtle Creek Association revealed about 2,602 trees. Our study by airplane revealed more than 3,000 trees in the same area. Earlier studies couldn’t account for the overall height or highest point of trees.”

“We’re able to survey more than just public land, by scanning from the air. The airplane allows us a more complete picture than volunteers on foot could achieve — as they can only travel through public property. We don’t intrude on private property by scanning from the sky.”

Dr. Fang Qiu

Qiu’s technique uses hyperspectral imagery to scan the distinct reflected image of a tree leaf — not unlike a tree species fingerprint. The leaf of each species of tree shows up differently in a hyperspectral scan, which means that a library of known species’ leaves can be collected and compared against the data collected when the scanning aircraft flies overhead. The advantages are legion, not only in time saved but in the accuracy of identifying the species of trees on the ground.

“We’re able to survey more than just public land, by scanning from the air,” Qiu said. “The airplane allows us a more complete picture than volunteers on foot could achieve — as they can only travel through public property. We don’t intrude on private property by scanning from the sky.”

Qiu’s remote sensing approach to surveying utilizes a second laser scanning as well, called Light Detection and Ranging, or LiDAR. The scan gathers data on all of the vertical structures in an area — including buildings, trees, and other background features — and separates the individual trees out from each other and from the other tall structures.

Qiu’s initial project surveyed only 20 square-miles of Dallas, but with additional funding he hopes to complete the picture of Dallas’ urban forest. Thus far, he’s finding predominately red oak, live oak, American elm, hackberry and pecan trees.


Media contacts: Brandon V. Webb, UT Dallas, (972) 883-2155, Brandon.webb@utdallas.edu


or the Office of Media Relations, UT Dallas, (972) 883-2155, newscenter@utdallas.edu

Text size: Increase text sizeDecrease text size

Overhead hyperspectral tree images

Above: Hyperspectral imagery identifies areas with trees, shown above in red.

Below: The approach uses more than one layer of scanning to separate individual trees from one another.

Share this page

Email this article.

Sunday,
August 31, 2014