Infrared and Laser Engineering, Volume. 49, Issue S2, 20200187(2020)
Scene identifiability analysis based on conditional evidential networks
According to the application requirements of target recognition, task programming and template preparation, an algorithm of scene identifiability analysis based on evidential networks was proposed to realize the quantitative analysis of the scene identifiability degree. After having acquired the support data of the research area and setting the imaging parameters, a certain number of salient ground objects were extracted as the scene nodes from the data. Then, the identifiability degree of each extracted object was assessed from three aspects, including scale significance, shape uniqueness and visualization. After that, the conditional belief function which represented the mutual support degree between scene nodes was defined by the contour point number of the extracted objects. Finally, the analysis results of the scene identifiability were obtained by the reasoning and fusion ability of evidential networks. Experimental results demonstrate that the algorithm of scene identifiability analysis is reasonable and effective, which meets the requirements of mission planning and thus exhibits great practical value.
Get Citation
Copy Citation Text
Liu Songlin, Hu Jun, Zhang Li, Gong Danchao. Scene identifiability analysis based on conditional evidential networks[J]. Infrared and Laser Engineering, 2020, 49(S2): 20200187
Category: 图像处理
Received: May. 12, 2020
Accepted: Jun. 20, 2020
Published Online: Feb. 5, 2021
The Author Email: Jun Hu (hujun25@mail.sysu.edu.cn)