structure-from-motion - multi-view stereo-photogrammetry

Nowadays, one of the latest methods for estimating characteristics of trees at different vegetative stages is three-dimensional models prepared by terrestrial close-range photogrammetry. In this regard, this study aims to estimate the different two- and three-dimensional characteristics seedlings using terrestrial close-range photogrammetry. Three-dimensional models were prepared using images taken by a smartphone camera and the process of “Structure from Motion with Multi View Stereo photogrammetry” (SfM-MVS). To evaluate the accuracy of the obtained results, characteristics such as diameter at the middle height of stem, height, crown height, crown diameter, stem volume, crown volume and total volume of each seedling were measured. According to the RMSE% and the calculated Bias% (less than 10%), the images and model used to estimate the two-dimensional characteristics of seedlings, had factual accuracy. Also, among the three-dimensional characteristics of the seedlings, the highest and lowest accuracy related to stem volume with R2 = 0.89, RMSE% = 23.08% and crown volume with R2= 0.67, RMSE% = 31.92% respectively. Overall, it can be concluded that the images taken by the smartphone camera and the model used in this study have remarkable accuracy to estimate characteristics of seedling such as diameter at the middle height of stem, height, crown height, crown area and stem volume, However, it is not accurate enough to estimate the crown volume characteristic.