Research of Centroid Extraction for Feature Point Based on FPGA

Abstract

This paper focuses on the centroid extraction algorithm of feature point. We present a recognition algorithm to identify the feature point and extract centroid. This algorithm can extract the centroid of the feature point from the complex background by scanning the original image only one time. We design a hardware architecture and implement it based on FPGA. Experimental results show that it can extract the centroid coordinates exactly from the complex background in real time with low-cost hardware resources.