Application of SIFT algorithm in the sea ice image matching

摘要

In order to meet the needs of the marine development and maritime disaster reduction work, it is necessary to strengthen the monitoring and forecasting of sea ice. Doing sea ice drift detection can provide accurate field data for the numerical prediction of North Sea and medium-range forecasts, which is an effective way to improve forecast accuracy and reducing ice disasters. Use feature point extraction algorithm and image matching algorithm to study the ice drift speed and direction extraction method, which is the preparation of sea ice drift monitoring technology research. In many image matching algorithms, SIFT algorithm has good scale invariance of rotation, spatial rotation invariance and strong matching ability. SIFT and its extension algorithm has been proved to be the robust in the same descriptor, which is currently a hot research at home and abroad. This paper introduces four main steps of the SIFT algorithm and its application in the sea ice image matching feature points.