Tree detection from data obtained by unmanned aerial vehicle

With the advancement of technology today, it is easier to reach information in cartography as in every subject. The data collection phase, which used to take months, started to be realized in a short time with unmanned aerial vehicles. Unlike terrestrial data collection methods, when the data was collected by unmanned aerial vehicle, we started to have information about the land, to the smallest details. However, there was a new problem. Although this problem related to the processing of this collected data was tried to be solved by the operators spending a long time processing the data at the computer, it was not efficient enough. In this thesis study, it is aimed to save time and transfer these processed data to various geographic information systems by paying attention to the accuracy of data processing as well as data collection. The thesis was done in two different areas with RGB (Red-Green-Blue) tapes. In the study, digital images collected with unmanned aerial vehicles were evaluated in pix4D software and orthophoto, digital height model and digital terrain model were produced from images. With the help of these data, object-oriented classification is aimed in the eCognition program. When making object-based classification, segments with similar characteristics are classified according to the index values determined by dividing the image into groups during segmentation. Again, control data has been generated in eCognition software and accuracy analysis has been made with the help of error matrix. The agreement between accuracy analysis and classification was calculated.