An Improved Surf Algorithm Image Feature Point Extraction Based on Hadoop Cluster

摘要

This paper discusses the method of extracting image feature points on a Hadoop cluster using an improved Surf algorithm. Edge detection is performed on the pictures with Prewitt operator. The feature points are extracted at the edge of the target image by Surf algorithm, which remarkably decreases the amount offeature points in the image. The parallel processing of images on Hadoop cluster platform greatly improve the processing speed of massive images.