Efficient Implementation of Smoothing Filters Using FPGA

Abstract

Image smoothing filters are applied to reduce the amount of noise in digital images by applying the convolution process on image pixels. However, the implementation of these filters in real time systems involves the complexity of implementation as floating numbers calculation and division. In this paper, the implementation of two smoothing filters, Mean and Gaussian is achieved using FPGA Spartan 3E device. The implementation was performed in a simplified and way that does not require division, and resultant images were displayed on LCD using configured VGA ports. Two filters were effectively implemented by utilizing 880 slices (5%) of Spartan 3E- XC3S160E device with maximum frequency of 56.6Mhz.