Multi-processor denoising of weak video signals in strong noise

This paper presents multi-processing implementation of video de-noising algorithms based on Least Square Estimation (LSE) and Best Linear Unbiased Estimation (BLUE). Real-time processing was achieved using multiple Digital Signal Processors (DSPs) and tested on video images with low Signal-to-Noise ratio (SNR). The imaging device used to perform the vision-processing tasks is the Matrox Genesis, which consists of on-board processing, frame grabber and display modules. Texas Instruments´ TMS320C80 Multimedia Video Processor (MVP) that runs at 50 MHz is employed to perform the signal processing task.