Estimating crown defoliation of Scots pine (Pinus sylvestris L.) trees using small format digital aerial images

Abstract

This study focuses on the possibilities of using small format digital aerial images for the estimation of tree crown condition. The test area was located in the eastern part of Lithuania where Scots pine (Pinus sylvestris L.) trees prevail and was photographed using a Canon EOS-1DsMark II digital camera installed on-board a SkyArrow ultra-light aircraft. The camera lenses were adopted to capture images corresponding to conventional color-infrared photography. In addition, the test area was photographed using a large format digital frame aerial camera (Vexcel UltraCam D) installed on board a Rockwell Turbo Commander 690A high performance commuter aircraft. The ground sampling density of the images taken was around 9-10 cm. Crown defoliation was assessed in the field for more than 500 Scots pine trees located in 46 sample plots representing stands of trees that were either 65 years old or 170 years old. Spearman’s correlations coefficients were used to check for relationships between tree crown defoliation and image characteristics. The defoliation was also predicted using the non-parametric k-Nearest Neighbor method applied on data available from aerial images alone. The results were validated using the “Leave One Out” technique by comparing the obtained data with data from the field assessed defoliation rates. The prediction root mean square errors were calculated using data from the small format aerial images as being 11.5% for the younger trees, whereas those calculated using conventional aerial images were between 9.5 and 9.9%. The differences in predicted root mean square errors disappeared in the older stands and both methods produced errors of between 8.1 and 8.5%. Defoliation class was correctly predicted for approximately 84-88% of the older tree crowns and correctly for 75-85% of the younger tree crowns. These results showed that small format aerial images had the potential to predict defoliation in tree crowns and were comparable with results obtained using conventional aerial images. Their main advantage is that small format images are much cheaper to obtain than conventional images when the areas targeted are thousands of hectares in size.