Estimating the velocity and direction of moving targets using a single optical VHR satellite sensor image

摘要

The present contribution demonstrates the feasibility and explores the limits of a new method for estimating the velocity and direction of moving targets using a single, optical, very-high-resolution (VHR) satellite sensor dataset. The method is based on the fact that there is a time lag between the data collection of the panchromatic (P) and multi-spectral (MS) sensors in the same VHR platform. Consequently, it is developed around three main steps: (1) accurate image-to-image registration between MS and P images with a sub-pixel displacement error, (2) precise location of barycentre of targets by mathematical morphology-based image transforms, and (3) estimation of the ground velocity and direction of the target using the MS-P spatial displacement, the known time lag, and an image-to-ground transformation taking into account the interior and exterior orientation of the sensors and a terrain height reference. An evaluation of the reliability and limits of the proposed method based on the observation of the results regarding manually selected moving and non-moving targets is included.