Real Time Edge Detection using Three Dimensional Systolic Array

Ö. Tamer, L.A. Rønningen, and M. Panggabean (Norway)


Systolic array, edge detection, video processing, 2-D convolution


Real time video processing is gaining importance, while the size of the video sequences are also getting larger with the improvement of new multimedia applications. Even though software based video and image processing systems have significant advantages like flexibility and cost effec tivity, they are lacking performance, which makes hard ware based systems more preferable for real time high de finition applications. This paper presents a row wide edge detection algorithm based on a three dimensional systolic array structure. The row wide approach improves the per formance of the algorithm when compared with the pixel wide algorithms presented in previous works. The pre sented results show significant improvement in processing time.

Important Links:

Go Back