Cockpit Crew Task Analysis for Civil Aircraft Based on Image Recognition

Whether the cockpit crew task design of civil aircraft is accurate and reasonable directly affects the preparation y requirements of civil aircraft crew and flight safety. Civil aircraft cockpit minimum flight crew provisions is also an important provision in the process of airworthiness certification of new models, at this stage, the airworthiness certification of this provision is mainly analysed through the simulator test and test flight test through the pilot's subjective evaluation, there is a strong subjectivity, not able to objectively evaluate the workload of the crew, which has brought a lot of challenges to the airworthiness certification. This paper proposes a method based on image recognition to carry out the minimum crew task analysis in the cockpit of a civil aircraft. According to the route operation scenario of a civil aircraft, we carry out the simulation test of the man-in-the-loop, use the camera to track and record the spatial movement trajectory of the pilot's hand, judge the crew's operation task in the cockpit by identifying the crew's gesture, and make use of the cockpit's interface for displaying some information to carry out the analysis of the crew's task in the cockpit of a civil aircraft. The method helps to improve the accuracy and efficiency of the monitoring of crew tasks in the cockpit of civil aircraft, and provides a basis for optimising crew resource management and flight safety.