Extraction and analysis of sea ice leads from SAR images

An approach using the technique of skeletonization to measure lead characteristics in Arctic sea ice radar images is presented. Skeletonization has been done in other studies using a mathematical morphology approach, but is approached using a thinning algorithm in the present study. The thinning approach guarantees maximally thin and minimally eroded skeletons. Image analysis techniques are then used to distinguish individual lead skeletons and measurements are made of these skeletons, including the main diagonal length, skeletal length, total skeletal length, linearity index, branching index, area, average width, main diagonal orientation, and average branching angles. Lead population measures are then calculated which include the overall lead density, average lead orientation, average length, average area, and categorization of leads by orientation. These population measures are aggregate measures of the individual lead measures compared across the entire image