Scripting in landscape assessment: a vegetation density case study

Abstract

Assessing landscape components in photographs is important in landscape architecture, geography, and other fields. One landscape component that is frequently assessed is density of vegetation. Assessing vegetation density, for instance, can help researchers find out how much vegetation people prefer so that they can encourage landscape designers to build more positive urban environments for people. Others assess vegetation density to find out whether sites meet urban forestry guidelines. Assessing this density from street level photographs, however, can be challenging. We need better ways to rapidly process hundreds or thousands of images to assess vegetation density within a reasonable timeframe. One promising solution is scripting, a technology that allows us to process large amounts of data in a relatively short time. This tool identifies the shades of color green detected within an eye level landscape photograph and calculates the density of those shades, similar to the NDVI method. However, a scripting tool to assess vegetation density in photographs has not been tested and implemented. To test a scripted vegetation density tool, we combined data from three different Green Infrastructure preference studies. Using 152 photographs with varying densities of GI, we compared the vegetation index retrieved from Brown Dog, an online scripting service acquired from our collaboration with the National Center of Super-computing Applications (NCSA), to the vegetation index retrieved from a manual method. Our findings suggest that Brown Dog’s scripting service accurately predicts vegetation density (Adj R^2=0.91), and that future research can use this service to speed up the process of identifying different vegetation density in photographs. This study is significant because it tells us we can use this technology to accurately assess eye level vegetation density in a photograph. It will help researchers study other aspects of human responses to greenness, such as attention restoration and stress recovery, in the future. Future research should implement and test this scripting tool with other landscape photographs, such as photographs from image-sharing websites and Google Street View images. The scripted vegetation density tool is only one of many scripts that Brown Dog offers. Other scripts include changing file formats, obtaining certain images from the webs, and obtaining landscape scenes from Google Street View. Future research can implement and test the accuracy of these scripts, thereby making landscape and geography research easier and more efficient.