Simulation of Complex Camera Motion Paths for Image Stabilization

Abstract

Image stabilization (IS) is a family of techniques used to reduce blurring like effects associated with the motion of a camera during exposure. Specifically, it compensates for pan and tilt (angular movement, equivalent to yaw and pitch) of a camera. These camera motions are combinations of basic movements like linear horizontal or vertical motions, circular motion or sometimes arbitrary motions. In true sense motion blur generated by these motions is the result of the relative motion between the camera and the scene during the integration time of the image, which are generally observed in amateur photography. Different motion paths by taking up a form of point spread function (PSF) kernel, produces different effect on image captured. Camera motion kernel and the real image are mathematically related via convolution operation and hence the deconvolution is used to recover the original image. Thus the work is intended to simulate different camera motion paths and obtain image stabilization for camera shaken images.

Keywords: PSF, blur kernel, image degradation, blind deconvolution,  camera motion paths