Comparison of Approaches of Distributed Satellite Image Edge Detection on Hadoop

In the world of bigdata, Hadoop has become the major platform for storage and processing. Due to advancement in technology in the area of remote sensing, tremendous growth in data has been witnessed. In this paper, we have conducted experiments and compared two methods in which edge detection of satellite images is performed on Hadoop. Since, edge detection is one of the prime steps in the field of image processing and is being used for object detection in the image, we have targeted this basic algorithm of image processing for our experiments. In earlier research for edge detection on Hadoop, SequenceFiles were used to store and process images. In our experiments, we have leveraged distributed processing of Hadoop by logically splitting the file on HDFS and performing edge detection in distributed manner. The experiments were performed on Amazon AWS Elastic MapReduce (EMR) cluster using different satellite images varying from 10MB-200MB. The paper describes the comparison of the two approaches.