Selection of optimal filtration and prediction block in the single target tracking system with one imaging sensor under changeable daylight | ||||
| The International Conference on Electrical Engineering | ||||
| Article 45, Volume 8, 8th International Conference on Electrical Engineering ICEENG 2012, May 2012, Page 1-13 PDF (405.12 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/iceeng.2012.30791 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
| Mikluc Davorin; Antonijević Goran; Andrić Milenko; Bondžulić Boban | ||||
| University of Defence, Military Academy, Belgrade, Serbia. | ||||
| Abstract | ||||
| The paper presents a system with one imaging sensor. Serial communication interface enables azimuth and elevation control, zoom setting, and changing the mode into infrared. Single target tracking in plane in multiple scenarios has been analysed in detail. By comparative analysis of CA models, Singer model and IMM algorithm with two models, and on the basis of root mean square error of position estimation, optimal filtration and prediction block has been selected. The first scenario shows linear target tracking at variable velocity. In the second scenario, the sensor tracks a maneuvering target. The third and fourth scenario present target tracking with infrared sensor with the cinematic characteristics identical to the first two scenarios. These are followed by scenarios in which the imaging and infrared sensors are controlled in target tracking, and the influence of different filtration and prediction blocks on the tracking is analysed. The last scenario shows target tracking under changeable conditions with automatic sensor control. All the scenario analyses are based on root mean square error of position. | ||||
| Keywords | ||||
| Image restoration; remote sensing and image blur models | ||||
|
Statistics Article View: 105 PDF Download: 197 |
||||
