Monocular Drogue Measurement Based on SiamMask Tracking Model and Feature Point Matching

To alleviate the pressure load on pilots during in-flight refueling, it is essential to automate the aerial refueling mission in a highly safe and reliable manner.In real refueling scenarios, factors such as strong light interference, partial obscuration of the drogue, and changes in orientation pose significant challenges. Traditional tracking methods are difficult to apply and cannot guarantee the required measurement accuracy.Furthermore, the pinhole imaging model is unable to calculate the spatial position of the black center in the drogue.To enhance the robustness and anti-interference capability of monocular drogue measurements, we designs a method for solving the spatial position of the drogue based on feature point matching, utilizing the binary mask output by the fully convolutional Siamese network SiamMask tracking model.