Aerial image stitching based on group control

Aerial image stitching is a significant technology in the field of UAV applications. The state-of-art methods for aerial image stitching are always difficult to simultaneously guarantee the accuracy and efficiency. Aiming at the bottleneck, a new aerial image stitching algorithm based on group control is proposed. It mainly provides a hierarchical aerial image stitching framework, which is constructed by a dichotomous grouping and stitching process. Meanwhile, the direction of stitching is also controlled to prevent error accumulation. Additionally, the time consuming for stitching can be further cut down without losing accuracy, which attributes the success to the multithreaded concurrent implementation. Experimental results validate that a more attractive aerial image stitching performance can be achieved with the proposed algorithm.