Fast image matching via multi-scale Gaussian mutation pigeon-inspired optimization for low cost quadrotor

Abstract

Purpose The purpose of this paper is to propose an improved optimization method for image matching problem, which is based on multi-scale Gaussian mutation pigeon-inspired optimization (MGMPIO) algorithm, with the objective of accomplishing the complicated image matching quickly. Design/methodology/approach The hybrid model of multi-scale Gaussian mutation (MGM) mechanism and pigeon-inspired optimization (PIO) algorithm is established for image matching problem. The MGM mechanism is a nonlinear model, which can adjust the position of pigeons by mutation operation. In addition, the variable parameter (VP) mechanism is exploited to adjust the map and compass factor of the original PIO. Low-cost quadrotor, a type of electric multiple rotorcraft, is used as a carrier of binocular camera to obtain the images. Findings This work improved the PIO algorithm by modifying the search strategy and adding some limits, so that it can have better performance when applied to the image matching problem. Experimental results show that the proposed method demonstrates satisfying performance in convergence speed, robustness and stability. Practical implications The proposed MGMPIO algorithm can be easily applied to solve practical problems and accelerate convergence speed of the original PIO, and thus enhancing the speed of matching process, which will considerably increase the effectiveness of algorithm. Originality/value A hybrid model of the MGM mechanism and PIO algorithm is proposed for image matching problem. The VP mechanism and low-cost quadrotor is also utilized in image matching problem.