Vision-Language Modeling with Regularized Spatial Transformer Networks for All Weather Crosswind Landing of Aircraft

arXiv:2405.05574v1 Announce Type: new
Abstract: The intrinsic capability to perceive depth of field and extract salient information by the Human Vision System (HVS) stimulates a pilot to perform manual landing over an autoland approach. However, harsh weather creates visibility hindrances, and a pilot must have a clear view of runway elements before the minimum decision altitude. To help a pilot in manual landing, a vision-based system tailored to localize runway elements likewise gets affected, especially during crosswind due to the …

abstract aircraft arxiv capability cs.cv depth of field extract however human information intrinsic landing language modeling networks pilot spatial transformer type visibility vision vision-language weather