Summary of Paper: “Individual Tree Crown Delineation from Airborne Laser Scanning for Diseased Larch Forest Stands”

Chloe Barnes, Postgraduate Researcher at the University of Leicester’s Department of Geography, shares with Lidar News a summary of the paper “Individual Tree Crown Delineation from Airborne Laser Scanning for Diseased Larch Forest Stands”.

Summary of Paper: “Individual Tree Crown Delineation from Airborne Laser Scanning for Diseased Larch Forest Stands”

The published research successfully isolated individual larch trees within a forest in South Wales affected by larch disease using data collected from airborne surveys. It provides the first example for the use of remote sensing to segment diseased larch trees within forest environments. The isolation of individual tree crowns is particularly important in the study of tree disease, as it allows the condition of each tree to be assessed and symptomatic individuals to be targeted.

The methods employed by the research converted point clouds collected via airborne laser scanning to canopy height models, representing the canopy surface in a raster format. Canopy height models were generated at three different pixel sizes (0.15 m, 0.25 m and 0.5 m), providing reduced quantities of height variability within crowns as pixel size increased. Subsequently, canopy height models were then filtered to reduce irregularities across the canopy surface. Treetops were pinpointed by identifying points of local maxima across the canopy height models. These were used as the seed input to perform two different commonly used segmentation algorithms: watershed and region growing, to delineate tree crown boundaries. The study explored the best combination of canopy height model generation method, canopy height model pixel size and segmentation algorithm for the isolation of diseased larch trees.

Larch disease caused by the pathogen Phytothpthora ramorum, has affected large areas of larch forest in Southwest England, South Wales and Southwest Scotland. Current approaches to assess and monitor Larch disease across the UK have been conducted manually by tree-health surveyors during helicopter surveys conducted by the Forestry Commission and Natural Resources Wales. Whilst the condition of larch forests can be assessed in this way, larch trees within the landscape cannot be individually assessed and located.

Identifying individual trees in forests and woodlands automatically from airborne laser scanning has successful been achieved for a range of forest types across the globe. However, in cases where trees are diseased and have lost leaves or needles, the use traditional techniques can be difficult as trees no longer exhibit full crowns. This research has identified a successful method to identify larch trees which are completely or partially defoliated as a result of larch disease.  The findings of the research highlight the potential use of airborne laser scanning to automatically isolate individual trees within diseased forests. It provides the first step in developing a more automated approach to monitoring larch disease using remote sensing. The researchers aim to apply the results from the study to consider the use of airborne laser scanning for disease detection and assessment in larch trees.

The full article is available for download at http://www.mdpi.com/2072-4292/9/3/231.

Leave a Reply

Your email address will not be published. Required fields are marked *