Enterprise Crowdsourcing Platform for Geospatial Image Analysis

Although today's computing systems present powerful solutions to process big data with the help of recent advances in cloud computing technologies, many problems remain unsolved due to lack of acceptable algorithms. For instance, searching for a lost plane or damage assessment after earthquake on a high-resolution remote sensing satellite image are unsolved problems. In recent years, in order to solve such unsolved problems, a group of expert and nonexpert people called crowd is utilized. An expert in the group is not expected to solve whole problem. Instead, geospatial image is partitioned in space and each expert in the pool studies the partition assigned to him/her. The solution to the original problem is obtained by merging the partial solutions. There are two open issues: i) How to partition the space and how to distribute the partitions to the crowd ii) How to merge the partial solutions. In this study, we devise several algorithms to address these issues, introduce our web-based platform, and crowd-sourcing implementation.