Determination and Accuracy Analysis of Individual Tree Crown Parameters Using UAV Based Imagery and OBIA Techniques

Abstract In the process of producing information from images with very or ultra-high spatial resolution (VHR and UHR), the most accurate results are achieved by using object-based image analysis (OBIA) techniques. The most economical method to obtain UHR images is to use sensor systems that are integrated into unmanned aerial vehicles (UAV). In this study, which combines UHR-UAV-based images and OBIA-based analyzes, individual tree crown parameters were obtained, and the results were examined using various accuracy analysis techniques. For this purpose, the UAV data acquisition was performed at the altitude of 40 m above ground level, and a ground sample distance (GSD) of 1.28 cm was obtained. Photogrammetric processes were performed using the structure-from-motion (SfM) techniques, and orthomosaic and photogrammetric point cloud data were generated with 2.46 cm RMSE. OBIA-based techniques were applied to these data, and the individual tree heights, crown borders and related parameters were derived. For the accuracy analysis, actual tree heights were collected with terrestrial measurements. The reference tree crown borders were stereoscopically interpreted from UAV-based images. In this study, the accuracy of the tree crown borders and tree heights were tested over 31 parameters. Recommendations were presented by interpreting the ultimate accuracy values to determine the accuracy of the data obtained using OBIA techniques. As a result, OBIA techniques will increase the effectiveness and preciseness forest inventory applications, such as determination of the stand structural characteristics (canopy cover, canopy gaps, stand height etc.).