Identification of image angle using projective transformation: Application to banana images

Abstract

This paper presents the determination of the captured image angle and correcting the image distortion using Projective Transformation. A 5-Megapixel camera is used to capture 90 images including one reference image. The images are captured from various angles, distances and directions. An image registration technique is applied to align the captured image with the reference image. Projective transformation is one of the approaches used to map the pixel locations from the input image (moving data) to the new locations in an output image by referring to the reference image (fixed data). Trigonometric function (cosine theorem) is then used to calculate the rotation angle of the captured image and image scaling value. Scaling value shows the scale of expansion and contraction of the captured image after applying image registration process. For this project, all manipulations are performed using MATLAB. © 2014 IEEE.