10 December 2014 A new method for mesoscale eddy detection based on watershed segmentation algorithm
Author Affiliations +
Abstract
Mesoscale eddies are widely found in the ocean. They play important roles in heat transport, momentum transport, ocean circulation and so on. The automatic detection of mesoscale eddies based on satellite remote sensing images is an important research topic. Some image processing methods have been applied to identify mesoscale eddies such as Canny operator, Hough transform and so forth, but the accuracy of detection was not very ideal. This paper described a new algorithm based on watershed segmentation algorithm for automatic detection of mesoscale eddies from sea level anomaly(SLA) image. Watershed segmentation algorithm has the disadvantage of over-segmentation. It is important to select appropriate markers. In this study, markers were selected from the reconstructed SLA image, which were used to modify the gradient image. Then two parameters, radius and amplitude of eddy, were used to filter the segmentation results. The method was tested on the Northwest Pacific using TOPEX/Poseidon altimeter data. The results are encouraging, showing that this algorithm is applicable for mesoscale eddies and has a good accuracy. This algorithm has a good response to weak edges and extracted eddies have complete and continuous boundaries. The eddy boundaries generally coincide with closed contours of SSH.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Lijuan Qin, Qing Dong, Cunjin Xue, Xueyan Hou, Wanjiao Song, "A new method for mesoscale eddy detection based on watershed segmentation algorithm", Proc. SPIE 9261, Ocean Remote Sensing and Monitoring from Space, 92610A (10 December 2014); doi: 10.1117/12.2069167; https://doi.org/10.1117/12.2069167
PROCEEDINGS
7 PAGES


SHARE
Back to Top