Implementing a scalable image processing workflow on Microsoft Windows Azure platform

OData support
Supervisor:
Albert István
Department of Automation and Applied Informatics

Nowadays cloud computing applications are becoming more and more popular, and play a very important role on the internet. One of their main benefits is, that you have to pay only after the consumed ammount of services. This means, that the initial costs are low, and that it is suitable solution in those situation when then computing or storage needs of the application can rapidly change, or when there are regular, predictable peeks int the the service usage.

My task, processing the Student Feedback Surveys, belongs to the latter group of applications. Computational capacity is needed only twice a year for a few days. But as the speed of the processing is an important requirement, we need during these periods a larger ammount of computational capacity.

In the first part of my work I present the general characteristics cloud computing and the possible ways to classify the cloud computing services. After that the content of the paper will show the services provided by the Windows Azure cloud platform, compared to two other popular cloud solutions the Google App Engine and Amason Web Services.

I will introduce the planned steps of the survey processing togeother with the algorithms used during the identification. The identification of the surveys is accomplished with the SURF algorithm and the marks are found with Hough transformation and edge detection.

The application runs in the Windows Azure Cloud, and is based on the Windows Azure SDK. Beside the processing service I created a Silverlight based client application to administer the processing steps, and data, togeother with an ASP.NET MVC web application to display the results of the surveys.

Downloads

Please sign in to download the files of this thesis.