Quaternion‐based improved cuckoo algorithm for colour UAV image edge detection

Abstract

With the extensive application of unmanned aerial vehicles (UAVs), there is an increasing demand for fast processing of coloured UAV images. The coloured UAV image pixels are usually represented by quaternion vectors with three bands of visible light corresponding to the three imaginary parts of the pure imaginary quaternion. Accordingly, the colour image edge points can be determined based on the quaternion polar coordinating the rotation principle. Here, a quaternion-based improved cuckoo algorithm is proposed to perform fast processing for UAVs images. In particular, a novel guiding equation is used to optimize the positions of the improved cuckoo algorithm before the Levi flight. Furthermore, a novel disturbance equation is used to obtain a varied location for the next location after the Levi flight. Comprehensive experiments are conducted to evaluate the performance of the proposed solution. The experimental results showed that the proposed method significantly reduces the image processing time and remarkably improves the quality.