| Human factors related to pilots' behaviors play an important role in aviation safety.The motion state of an aircraft is the most important information that the pilots need to obtain.The cognition of the motion state of an aircraft mainly relies on pilots' visual system and vestibular system,and,more than 80% of the information of movement in external environment is recognized and processed by the visual system,so that the study of the cognition by pilot visual is an important research orientation in pilot modeling.Through the methods of image processing,do image preprocessing including graying?smoothing and threshold processing to the images outside the cabine which were collected by video camera,and then execute image edge detection based on Canny operator and Hough transform,after that we can get some information of the skyline as well as the centre-line and the sideline of the runway which we are interested in from the images outside the cabine.In this paper,the concept of fuzzy mathematics is applied to the study of pilot visual cognition modeling.The knowledge base of visual cognition model of pilots in-cabin and out-cabin in the visual approach stage is established by means of experiments.We achieved to combine the visual cognition model of pilots in-cabin and out-cabin by considering a pilot as a Finite-state machine.Finally,a complete visual cognition model of pilots in the visual approach stage is obtained.In this paper,the knowledge base of the visual cognition model out-cabin is divided into three parts:long-term memory,short-term memory and buffer zone,and the simple self-learning mechanism of the visual cognition model of pilots is realized in the form of feedback from the visual cognition model in-cabin to the visual cognition model out-cabin.In the end,the Dynamic Time Warping(DTM)algorithm is used to prove that the visual cognition model and human pilots produce the same visual cognition in flight tasks,so that the reliability of the model is verified. |
PDF Full Text Request