Aviation Image Retrieval and Simulation Based on Human Visual and Sensation

LIU Shu-qin,PENG Jin-ye
DOI: https://doi.org/10.3969/j.issn.1006-9348.2013.12.091
2013-01-01
Abstract:In this paper,accurate retrieval of aviation image was researched. The target in aviation image specif- ic areas generally show random distribution,interference objectives and recognized target are distributed in cross. The traditional image retrieval algorithms are based on the idea of goals for clues,according to the specific character- istics of the target pixel or related features,to complete the image positioning. Meanwhile,the disturbance rejection is also based on features contrast,more automated,under the circumstances of many targets,result in the extreme complex of exclusion process,efficiency and positioning accuracy is very low. To avoid these shortcomings,we pro- posed an aviation image object retrieval approach based on human visual perception. The method used local binary method to extract aviation image features,and the characteristics of the different types of human visual sense was used as essential data of aviation image retrieval,then create the human visual sensation model for aviation image re- trieval. Experimental results show that the improved algorithm for image retrieval aviation can greatly improve the re- trieval precision and recall,and for applications in different fields to provide the basis for aviation images.
What problem does this paper attempt to address?