DETECTION OF CITRUS TREES FROM UAV DSMS
| Autor: | Asli Ozdarici-Ok, Ali Ozgun Ok |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
lcsh:Applied optics. Photonics
010504 meteorology & atmospheric sciences lcsh:T business.industry Computer science Orientation (computer vision) 0211 other engineering and technologies Symmetry in biology lcsh:TA1501-1820 Pattern recognition 02 engineering and technology Integrated approach lcsh:Technology 01 natural sciences Maxima and minima lcsh:TA1-2040 Detection performance Artificial intelligence Symmetry (geometry) lcsh:Engineering (General). Civil engineering (General) Precision and recall business Simulation 021101 geological & geomatics engineering 0105 earth and related environmental sciences |
| Zdroj: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol IV-1-W1, Pp 27-34 (2017) |
| ISSN: | 2194-9050 |
| Popis: | This paper presents an automated approach to detect citrus trees from digitals surface models (DSMs) as a single source. The DSMs in this study are generated from Unmanned Aerial Vehicles (UAVs), and the proposed approach first considers the symmetric nature of the citrus trees, and it computes the orientation-based radial symmetry in an efficient way. The approach also takes into account the local maxima (LM) information to verify the output of the radial symmetry. Our contributions in this study are twofold: (i) Such an integrated approach (symmetry + LM) has not been tested to detect (citrus) trees (in orchards), and (ii) the validity of such an integrated approach has not been experienced for an input, e.g. a single DSM. Experiments are performed on five test patches. The results reveal that our approach is capable of counting most of the citrus trees without manual intervention. Comparison to the state-of-the-art reveals that the proposed approach provides notable detection performance by providing the best balance between precision and recall measures. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|