engleski

A*-Search Image Segmentation Algorithm for Mine-like Objects Segmentation in SONAR Images

This paper addresses a SONAR image segmentation method employing a novel A*-Search Image Segmentation (ASSIS) algorithm. ASSIS is applied on Mine-Like Objects (MLO) in SONAR images, where an object is defined by highlight and shadow regions, i.e. regions of high and low pixel intensities in a SONAR image. Usually, the A*-Search is used in mobile robotics for finding the shortest path, where the environment map is predefined and the start/goal locations are known. The ASSIS algorithm represents the image segmentation problem as a path-finding problem. Main modification concerning the original A*-Search is in the cost function that takes pixel intensities in order to navigate the 2D segmentation contour. Firstly, the ASSIS algorithm uses a pixel intensity thresholding technique to build a goal distance map with the numerical navigation function (NNF). The NNF uses the L2 norm to estimate the Euclidean distance between the current pixel and a goal pixel. The estimated distance is then recalculated as predicted cost h to the goal. So far made cost g is calculated according to the current path length. Finally, the sum of g and h defines the cost function f that determines the next contour point around an MLO.

A*-Search; Image Segmentation; Path Finding

nije evidentirano