Development of a Framework for Stereo Image Retrieval With Both Height and Planar Features

摘要: The wide availability and increasing number of applications for high-resolution optical satellite stereo images (HrosSIs) have created a surging demand the development effective content-based image retrieval methods. However, this is challenge existing methods since they were designed collected from close-range imaging sensors. Thus, successful not assured given mismatch between characteristics HrosSIs. Moreover, none remote sensing takes account specific HrosSIs such as viewing multiview angles. This paper proposes generic framework to exploit unique data so allow efficient accurate HrosSI retrieval. executed by similarity matching features obtained digital surface models (DSMs) orthoimages, both extracted In addition, significance height information was investigated. A prototype system implemented method validation using ISPRS benchmark test dataset. Experimental results show that proposed techniques are suitable spaceborne but might also be airborne well. alone inefficient unstable retrieval; however, combination planar stable.