FPGA-based Floating Point Fractional Order Image Edge Detection

In this paper, the hardware implementation of an image edge detector based on fractional order filters is presented. Fractional order filters excel over their integer order counterparts in their noise robustness and in detecting the high frequency details of images, which is very useful for biomedical image processing. The fractional order parameter also offers an extra degree of freedom in designing the fractional order filter to achieve higher efficiency than integer ones. The hardware design proposed exploits the single precision floating point arithmetic, IEEE 754 representation, to have precision in implementation. The filters were applied to standard images as well as an MRI scan as a biomedical image processing application. The hardware results obtained are the same as the Matlab simulations. The design is implemented on VIRTEX 5 XC5VSX50T-2FF1136 FPGA kit achieving a clock frequency of 170.1 MHz.