Piecewise Planar Region Matching for High-Resolution Aerial Video Tracking

Abstract

In order to tracking moving objects of aerial images, the frames and the scene is kept space consistency through image registration at the background. Due to high image resolution and large geographic deformation between different frames of aerial video, complicating the image registration. An piecewise planar region matching based image registration is introduced that can subdivide large frame into planar region, Image subdivision reduces the geographic distortions between aerial video, as it is usually the case of high-resolution aerial images. Then we can use select the most “useful” matching points that best satisfy the affine invariant space constraints are used to estimate the transformation model and register the images in a piecewise manner. Experiment result illustrate that the proposed method can register the high-resolution images and track the moving object in an aerial video.