Authors : K. Sai Venu Prathap, S.A.K. Jilani and P. Ramana Reddy
Abstract: An image mosaic is a technique of combining partial images that refer to the same scene to obtain a single large image that contains each information of a partial image. In this study, image mosaic is constructed with a single board Intel Compute Stick hardware device. The performance of various algorithms is evaluated in terms of criteria such as execution time, accuracy, Consistency of Corner Numbers (CCN) and a number of detected features with a rotation, horizontal shear, vertical shear transformations and uniform scaling. The image acquisition sensor of single-board hardware captures the remote sensing areas and scenes to construct the mosaic image. The image mosaic is executed in many phases such as feature detection, feature description, geometric computation, image warping and blending. Steerable filters of the 1st derivative is used for an orientation selective. The proposed algorithm is a combination of steerable filter and Features from Accelerated Segment Test (FAST) detector. Experimental results prove that steerable filters with FAST corner detector and Fast Retina Keypoint (FREAK) binary descriptor play vital roles in real-time applications than other detectors and descriptors.
K. Sai Venu Prathap, S.A.K. Jilani and P. Ramana Reddy, 2019. A Real-Time Image Mosaicing using Single Board Computer. Journal of Engineering and Applied Sciences, 14: 3150-3157.