| The Bohai Sea is located in highest latitude sea areas in China. And it’s an very important economic development zone. However, sea ice and its drift often block shipping, destroy offshore structures and limit the development of marine industry. Thus, approaches about sea ice monitoring are necessary to be studied. Remote sensing images with high space resolution and short time lag are necessary to track sea ice drift variation daily in the Bohai Sea in terms of sea ice transforming easily resulting from rapid motion of the sea water. Geostationary Ocean Color Imager(GOCI) with space resolution 500m can provide eights image sequences continuously in the daytime with one hour interval, which shorten revisit period largely. A method of sea ice monitoring combining projection and phase correlation is proposed to improve the performance of tracking sea ice motion, including rotation and translation. First, eight image sequences from Geostationary Ocean Color Imager (GOCI) are selected. Then, a specified function using projection transformation of self-adaptive windows with certain features is established. A set of angel candidates are obtained by searching for the optimal solution, from which determine the best one using revised correlation coefficient. Meanwhile, sea ice translation to the sub-pixel is derived from remaining phase information. Finally, matching errors due to ignoring phase frequency of images are decreased to a certain level. The results show that the maximum/average of root-mean-square error between the rotation angels using the method proposed in this paper and manual measurements are 0.59/0.50 respectively, while the ones between traditional technique and manual vales are 1.41/0.94 respectively, and the running speed is increased by 50.6%. For the translation part, the sea ice motion field agrees well with in-situ data and records of historical materials in the Bohai Sea. The method above can track sea ice motion quite well. |
PDF Full Text Request