Monocular Cone Sleeve Measurement Based on Kernelized Correlation Filter Tracking and Pinhole Imaging Model

To achieve accurate detection and position measurement of the refueling aircraft’s cone sleeve target during autonomous aerial refueling docking, this paper proposes a method based on Kernelized Correlation Filter tracking and the pinhole imaging model. For cone sleeve tracking, the fusion of multi-scale features from both HOG and convolutional layers is used to improve tracking accuracy in scenarios with complex backgrounds and varying cone sleeve scales. For cone sleeve measurement, the pinhole imaging model is employed to compute the spatial coordinates of the cone sleeve in real time, ensuring fast and precise docking during aerial refueling. The proposed method has been validated on both real refueling video datasets and simulated experimental data, demonstrating its effectiveness and feasibility.