Abstract:

Several studies have been conducted on object recognition by using high spatial resolution remote sensing images, which have more than three spectral bands and using object‐oriented classification. However, using three bands (Red, Green, Blue) images derived from UAV is difficult to recognize an object based only on spectral information. This paper focuses on improving accuracy of object recognition from UAV images with digital height model (DHM). The study results showed that the overall accuracy of object recognition of three band UAV images has been increased significantly from 85,48% to 94,72% when the combination of spectral bands and digital height model was made.