OPTIMAL 2D DFT WITH SLIDING WINDOW FOR IMAGE EDGE DETECTION

Abstract

An edge detection technique based on two dimensional sliding window Discrete Fourier transform (2D SDFT) and thresholding is proposed in this work. The 2 2 sliding window DFT with bin indices (k1; k2 = 0; 1) for horizontal edge detection and (k1; k2 = 1; 0) for vertical edge detection has been proposed. In 2D SDFT, the DFT bins at the current position of window are directly computed from the already computed bins of the previous position of window. These computed DFT bins are thresholded against a threshold value to obtain the edge map of input image. The output edge map of the proposed technique is equivalent to that of the traditional techniques. In the presence of noise and various signal to noise ratio conditions, the horizontal and vertical edges have been e ciently recovered with good Pratt gure of merit (PFOM) without any application of pre-processing and post processing techniques. 2D HDFT is also derived in this dissertation work. In 2D HDFT, the window will hope with hopping distance, L. Output of 2D HDFT and 2D SDFT with window size 4 4 and hopping distance L = 2 are compared. Output od both techniques are similar. The system-on-chip implementation of the 2D SDFT/HDFT edge detector on cyclone IV FPGA is also carried out.