Research on Visual Measurement Method of Aircraft Pose Parameters Based on Prior 3D Model Matching

To address the dynamic measurement requirements of aircraft pose parameters during near-field and takeoff-landing processes, we propose a visual measurement method based on prior 3D model matching. This method utilizes an intelligent electro-optical tracking measurement system equipped with multiband variable-focus imaging components to achieve extensive target tracking and imaging under complex weather conditions throughout the day. The system’s parallel computing and control platform receives a sequence of target images, employing a relative pose estimation algorithm based on the uncertainty of spatial reference points to obtain the initial pose parameters of the aircraft. Subsequently, it integrates an existing 3D CAD model of the aircraft for spatial online projection registration. This process iteratively optimizes conditions to minimize segmentation loss and dense feature matching loss, yielding optimal measurements of the aircraft’s pose. Experimental validation using actual C919 flight data demonstrates that the proposed method achieves a position measurement RMS error of better than 1.2m and an attitude measurement RMS error of better than 0.5°.