Efficient Fpga Architecture of Grey-Scale Soft Morphological Filters for Image and Video Restoration

Abstract

Recently, Soft Morphological Filters have shown excellent performance in restoring noisy images. To add soft morphological filters in image processing chains and also to process image sequences in real time, it became a necessity to implement soft morphological operations in hardware. In this paper, FPGA architecture of grey-scale soft morphological filters is proposed. The architecture is based on a stack filter expression with some modification. The design can calculate soft morphological operations of greyscale soft morphological filters with structuring functions of any size and shape inside the structuring function's overall support. The filter architecture guarantees fixed processing time independent of the size and shape of the structuring function of the filter. Implementation results showed that the proposed filter architecture is suitable for real time applications.