Using pivot image for the development of composite visual space based on image normalization

Digital videos have numerous applications, ranging from amateur videos on social media to complex imaging of space objects. Most of the time, different images and videos are combined to get a large field of view, including a simple panoramic image or detailed images of Mars created by stitching over 900 images. In most cases, different images are stitched linearly on a plane. However, recently researchers have stitched images and videos together to produce 360-degree views. In these 360-degree videos, there is primarily the main camera or a set of cameras that covers the scene. In contrast, another concept-the 3D view-captures depth along with the height and width of the images. The proposed work focuses on developing a composite visual space that captures a scene with different cameras and combines it accordingly. One of the essential features of the proposed work is image normalization. Images acquired from multiple sources are of various sizes, orientations, and brightness levels. A set of eight augmented normalized images are formed in a circular form where each image has its positional features. The article focuses on the normalization process of the images captured from different cameras with different specifications so that they can be used to form the proposed visual space. The results of the proposed algorithm are compared for time and space. The proposed algorithm uses 45-70% less computation. On average, this method normalizes with only 52% of computations for the selected dataset. This proposed algorithm us less computational and storage resource. In term if computational, it is faster as most of the calculations involve shifts in the integer values and the range of the values are from 0 to 255 that can fit in 8-bit integer. Most of the other methods uses complex real number equations that are computationally expansive and use more bits per pixels. Moreover, this approach requires a smaller number of shifts because of which quality is affected almost insignificantly.