In situ 3-dimensional measurement of ice shape based on binocular vision

Ice accumulation poses a significant threat to the flight safety of aircraft. In response to the demand for icing measurement in the fields of ice prevention and de-icing, this paper establishes an in situ measurement method for ice shape in foggy field based on binocular vision, employing multi-view imaging with a binocular camera to achieve three-dimensional (3D) reconstruction of various ice shapes. Considering the high transparency of clear ice and the influence of clouds on ice imaging, this study adopts dark-field illumination to obtain clear ice shape images and uses a moving binocular camera to acquire multi-view images. In addition, an improved Scale-Invariant Feature Transform (SIFT) feature extraction algorithm based on image enhancement is adopted to improve the accuracy of ice reconstruction in response to the problem of excessive brightness in localized areas of the ice image. The proposed method is used to experimentally measure the icing process of the NACA0012 wing model in a foggy field. Imaging and 3D reconstruction of three different types of ice shapes, namely clear ice, rime ice, and mixed ice, verify that the method established in this paper can realize high-fidelity measurements of 3D ice morphology and can be used for continuous monitoring of the ice shape during the icing process in foggy environments.