Research on visual-based motion sensing for unmanned drone-vessel system
With the rapid advancement of unmanned drone and vessel technologies,the collaborative application of un-manned drones and vessel is gradually becoming an indispensable component of future maritime military missions and smart transportation.However,considering the high dynamic characteristics of unmanned vessels under different sea conditions,traditional methods such as radar,while meeting the requirements for high accuracy and real-time performance,incur high costs.On the other hand,traditional vision-based approaches,though cost-effective and easy to implement,suffer from lower stability.This paper proposes a monocular method for unmanned drone and vessel system based on the principles of pinhole camera model.By identifying and ranging multiple landmarks in the landing area,a plane fitting reconstruction of the land-ing area for the unmanned vessel is performed,enabling the determination of the distance between the unmanned drone and vessel.The more landmarks identified,the lower the probability of algorithm failure.A numerical simulator was constructed to test the algorithm,evaluating its accuracy and stability under the large-angle rotation of the unmanned vessel.The results demonstrate that the algorithm can obtain accurate distances,with average relative errors in the X,Y and Z directions of 7.47%,6.48%and 8.04%,respectively.