ECE 5760: Final Projects

Real-time "Photoshop"

An FPGA-based Real-time Morphological Image Processor.

Xiaofan Bao (xb46@cornell.edu)

Jiayuan Wang (jw937@cornell.edu)

 

 

"An FPGA based real-time morphological image processor to show Edge-detection, Grayscale, Dilation, Erosion, Opening and Closing."

Project Soundbite

Introduction    top

Mathematical morphology is a theory and technique for the analysis and processing of geometrical structures, based on set theory, lattice theory, topology, and random functions. It is most commonly applied to digital images, but it can be employed as well on graphs, surface meshes,solids, and many other spatial structures. Often times, people would apply dilation, erosion, opening, closing and edge detection when they determine to use mathematical morphology to address pictures.

Our project is a FPGA based real-time Morphological Image Processor. Projects that deals with image processing usually can't address real time image calculation. But with FPGA we may make this happen due to its fast calculation. We use M4K blocks instead of normal registers to store pixels which might exceeds the total number of registers on board, we also utilize shift registers to address data in pipeline so that we can fasten the speed when it's processing.

We aims to realize dilation, erosion, opening, closing and edge detection using Cyclone DE-2 and Terasic TRDB-DC2. Video signal is sent to the board’s decoder chip from a camcorder via a composite cable. Once decoded to RGB, the signal will pass through edge detection, binaryzation that we created in hardware. Then the pixel data will be sent to the board’s VGA driver and outputted to a monitor. The user is able to select both binary threshold, exposure threshold and also the sobel threshold by using the DE2’s switches. There are also options to display in grayscale image, binary image and sobel schemes selectable by the switches as well.