Fast morphology algorithm with parallel processing structures

Morphology, including erosion and dilation, is important for many image processing applications, such as shape analysis, pattern recognition, denoising, and segmentation. In this paper, an efficient way to perform erosion and dilation operations is proposed. Since the proposed algorithm deals with each row and column separately, it can be implemented in a parallel processing structure, which can significantly reduce the computation time. Moreover, recursive structures can also be adopted in the proposed algorithm. In addition to less computation time, the memory requirement of the proposed algorithm is also very less. The proposed algorithm is suitable for any structuring element of the symmetric and convex form. When using different structuring element, only a very small part of the structure should be adjusted. The proposed algorithm can much improve the efficiency of morphology and is suitable for both software and hardware implementation.