Improving Test Operations Through Scalable Video Processing on Computer Clusters

In today?s propulsion testing environment, the number of simultaneously captured video streams, the number and variety of video derived data products, and quality expectations are increasing, while available manpower and the desired turnaround time between tests is decreasing. This project will take advantage of recently developed super computer clustering technologies to develop new scalable data processing techniques to improve the quality, efficiency and turnaround time of video based data products. New and innovative scalable video processing software will be developed that can be applied to multiple video streams simultaneously while utilizing any number of available processors. This new software will also allow a facility to create multiple and complex data products, with an order of magnitude less operator setup and intervention because the video streams can be selected individually or as a group and assigned to processes prior to, during or after data acquisition. The processors assigned to the tasks may be dedicated or ?borrowed? from other users on the network on an as available basis. With enough processors applied to the work, the time required for completing the data processing tasks could be reduced by several orders of magnitude.