Adaptive Self-Calibration of Fisheye and Wide-Angle Cameras

Abstract

In this paper, a plumb line method of highly adaptable self-calibration has been proposed and tested on multiple fisheye and wide-angle camera systems. In an industrial perspective, the calibration of heavy distortion lens (fisheye, wide-angle) is a scuffle of bottleneck factors such as specificity, cost and precision. To resolve this problem, an adaptive and automatic estimation of distortion parameters by employing parameter optimization techniques on scene geometrical attributes such as straight lines and edges was proposed. Experimental results demonstrate the quantitative comparison of the proposed method on simulated ground-truth data and real data with previous self-calibration techniques and OpenCV traditional checkerboard methods in terms of Peak-Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Metric (SSIM). Additionally, the proposed method can preserve almost all the image pixels which is the primary reason behind employing wide-angle and fisheye cameras. Finally, a practical qualitative analysis has been carried out by estimating distortion parameters under various scenarios such as indoors, outdoors, day, dawn, traffic conditions etc.