Image Segmentation Based on Height Maps.pdf

Image Segmentation Based on Height Maps Gabriele Peters1 and Jochen Kerdels2 1 University of Dortmund Department of Computer Science - Computer Graphics Otto-Hahn-Str. 16, D-44221 Dortmund, Germany 2 DFKI - German Research Center for Artificial Intelligence Robotics Lab Robert Hooke Str. 5, D-28359 Bremen, Germany Abstract. In this paper we introduce a new method for image segmen- tation. It is based on a height map generated from the input image. The height map characterizes the image content in such a way that the application of the watershed concept provides a proper segmentation of the image. The height map enables the watershed method to provide better segmentation results on difficult images, e.g., images of natural objects, than without the intermediate height map generation. Markers used for the watershed concept are generated automatically from the in- put data holding the advantage of a more autonomous segmentation. In addition, we introduce a new edge detector which has some advantages over the Canny edge detector. We demonstrate our methods by means of a number of segmentation examples. Key words: segmentation, edge detection, watershed, height maps 1 Introduction In this paper we introduce a new method for image segmentation. It is based on a height map which is generated from the gray values of the input image. Among the different methods for image segmentation morphological watersheds have some advantages [1]. They yield more stable results in comparison to other segmentation concepts such as detection of discontinuities, thresholding, or re- gion processing. But they also have a drawback. Watersheds work on height level images. The association with height maps refers directly to the input image. If an image is interpretable as topographic image, such as images of cells under a microscope, watersheds perform well. To apply watershed segmentation to arbi- trary images such as photographs of natural objects we propose to generate a height map which char