Ground truth label attachment to tree crown map

After segmentation, we classified segmented images into the following seven classes: deciduous broad-leaved tree, deciduous coniferous tree, evergreen broad-leaved tree, Chamaecyparis obtuse, Pinus elliottii or Pinus taeda, Pinus strobus, and non-forest. The ‘non-forest’ class included understory vegetation and bare land, as well as artificial structures. For deciding these classes, we conducted field research. We set three rectangular plots sized 30 m × 30 m and checked the tree species, regarding the classes we decided could be identified from the November 20 drone images. The Pinus elliottii or Pinus taeda class consisted of two Pinus species, because these two species are difficult to identify from drone images. At the ground truth map-making phase, we visually attached the class label to each tree crown, using nearest neighbour classification in the eCognition software to improve operational efficiency, which was then used for forest mapping38 (Fig. 3). More specifically, we chose some image objects as training samples and applied that algorithm to the overall tree crowns. In subsequent steps, by adding wrongly classified objects to correct classes of the training samples, we improved the accuracy of the ground truth map.

Segmentation and ground truth map-making result. The tree classes found in the image on the left are represented by the colours explained in the legend in the figure on the right. This figure was created using ArcGIS Desktop v10.6 software (https://www.esri.com, Environmental Systems Research Institute, Inc., Redlands, United States).

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.