Enhanced Georeferencing by Adapting Improvised Lens Distortion Correction in Stationary Drone Video Frames

Accurate lens distortion correction is crucial for maintaining geometric precision in various imaging applications. This abstract presents a novel approach to radial lens distortion correction that aims to enhance image analysis accuracy. This paper introduces an advanced distortion model that incorporates higher-order coefficients (k3 and beyond), surpassing the limitations of traditional models. The proposed method is integrated into a homography-based georeferencing framework, effectively rectifying distorted images captured by stationary drones. Through a meticulous calibration process involving known coordinates and a centroid point, it determines distortion coefficients with heightened accuracy. The proposed model not only rectifies distortion but also improves the precision of speed calculation for stationary drone video frames. Computational efficiency remains a focus, accommodating wide-angle lenses, and complex distortion patterns. In contrast to traditional models, this approach captures intricate distortion behaviors, enabling precise geometric measurements crucial for photogrammetry and computer vision applications. This substantiates the effectiveness of proposed model through comparative analyses, highlighting its superior distortion correction and enhanced accuracy in speed calculations for stationary drone imagery. The technique significantly improves the measurement of vehicle speed by accounting for subtle horizontal or vertical movements, offering valuable insights for transportation analysis and related applications.