Target detection approach for UAV via Hybrid Model

Aiming at the problem of poor target detection of unmanned aerial vehicles (UAVs) in complex environments, the hybrid model of Edge Potential Function (EPF) and Cauchy Gaussian Simulated Annealing Pigeon-inspired Optimization (CGSAPIO)algorithm is proposed, which realizes the target detection task of UAVs in low altitude flight aerial image. First, the EPF is calculated from the edge map of the original image, and a template specifying the target under different affine transformations is provided, and then the template matching is implemented using an optimization algorithm. The Pigeon-inspired Optimization (PIO) algorithm based on bionic intelligence is a new optimization algorithm. Based on this, this paper proposes an improved algorithm to achieve better matching effect. A series of experimental results show that the proposed algorithm is more robust and effective than basic PIO and other improved algorithms. At the same time, it also can ensure the accuracy of target matching.