Planar surface area transformation using modified homography matrix

Abstract

This research targets a problem of planar area size inconsistency after image warping to the top view. An orientation sensor attached to a camera was used to acquire the camera’s orientation in real time. Then, extrinsic parameters derived from the sensor and pre-computed intrinsic parameters were used to generate a homography matrix. The homography matrix was modified to eliminate the effect of an angle change, by avoiding the use of shift parameters. In doing so, an area size stays the same regardless of any change in such an angle. A separately required translation matrix was used to shift the top view image. In the shifted top view image, the number of target pixels can be directly counted. Finally, the number of pixel was converted to an area size in real-world units.