|
Automated Video Screening Techniques for Operator Workload Reduction
Navy SBIR 2012.1 - Topic N121-051 NAVSEA - Mr. Dean Putnam - [email protected] Opens: December 12, 2011 - Closes: January 11, 2012 N121-051 TITLE: Automated Video Screening Techniques for Operator Workload Reduction TECHNOLOGY AREAS: Information Systems, Sensors, Electronics ACQUISITION PROGRAM: PEO IWS 5.0, Aircraft Carrier Tactical Support Center (CV-TSC), ACAT III OBJECTIVE: The Navy seeks to design, develop, and evaluate innovative algorithms, techniques, or system approaches to rapidly and automatically screen large volumes of various types of streaming video and to provide appropriate tagging of events of interest in order to reduce operator workload and complexity of interaction, and to focus attention on events that increase situational awareness. DESCRIPTION: With the evolution of Navy systems, significant new or additional data sources are providing high volumes of additional information to more highly integrated systems. At the same time operational manning levels are expected to remain at current levels or be reduced in the future. Operators can often become overwhelmed during their vigilant analysis of large quantities of FLIR (forward looking infrared) and ISAR (Inverse synthetic aperture radar) video recordings. Overwhelmed operators can delay the decision-making process since critical data is not properly analyzed. For instance, a system like the AN/SQQ-34C Aircraft Carrier � Tactical Support Center (CV-TSC) has added processing from several heterogeneous sensor streams, such as ISAR and FLIR. These new interfaces introduce significant volumes of streaming video data when helos are deployed. The Navy desires to have these multiple sources analyzed and cue (bell ring) the operator to high interest events. This topic is different from current "smart video" systems since it must incorporate analysis of the various feeds and not just the transmission of the data from the helicopter back to the carrier. The Navy seeks innovative concepts to reduce operator workload relative to data screening activities for the many sources and modes of video data, thus reducing the volume of data the operator has to sift through to find the important information. Potential solutions might focus on automatically identifying, detecting important features and events within the video streams, potentially tagging those events, and/or presenting events in prioritized fashion for operator review and concurrence. There are many video analytics and smart video systems approaches that have been applied in the commercial security industry to detect change, provide object identification and recognition, and even perform object behavior discrimination. Most have been focused on human features, vehicle tracking, and perimeter or space security. Algorithmic approaches to object detection, tracking, and identification from the video sources as well as innovative display and presentation solutions for rapid operator detection, discrimination, review, and assessment are all potential solutions. Products from the video review stream should be sufficient to support automated association or automated tagging operations to support downstream fusion and significantly reduce operator interaction time, time to detect, and/or increase holding time, and improve tracking or classification performance. Successful technology transition will provide improved performance and reduced cost for ASW command and control systems on surface combatants, carriers, and shore command facilities. PHASE I: This research will develop algorithms and/or approaches for rapid screening of streaming videos for ISAR and FLIR. It will provide object recognition and detection on video clips. An analysis of the feasibility of implementing the proposed concept using simulated or real data will be conducted. A Phase II development plan with performance goals and key technical milestones will be created. PHASE II: Implement the plan developed in Phase I by developing, evaluating, and validating the Phase I solution in a prototype software baseline suitable for real-time operation. A proof of concept will be conducted with simulated and pre-recorded test data. The performance will be assessed using quantitative measures of performance. An innovative data visualization approach will be identified and defined that will enable the migration of compliant software into a single geospatial framework without loss of functionality. The approach should facilitate interoperability between geospatial objects (GO) and diverse client applications. System performance will be demonstrated through prototype evaluation and modeling. The evaluation should show the prototype meets Navy requirements. A Phase III development plan will be prepared to show how the technology will transition into fielded Navy systems. PHASE III: If a Phase III contract is awarded, the company will integrate and test their prototype software in a real-time environment via the CV-TSC. The company will support the Navy for test and validation in a real-time setting to certify and qualify the system for Navy use and support the production and installation of their technology into fielded Navy Systems. PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: Rapid video review and assessment technologies have a wide array of commercial and federal uses in video security, forensics, product quality monitoring, traffic and pedestrian tracking and behavior monitoring, access and perimeter control systems, wide area surveillance, and intelligence gathering. REFERENCES: 2. T. Lindeberg (2008/2009). "Scale-space". Encyclopedia of Computer Science and Engineering (Benjamin Wah, ed), John Wiley and Sons IV: 2495�2504. 3. "The Use of Semantic Human Description as a Soft Biometric", Samangooei S, Guo B, Nixon MS. Arlington, VA, USA, 2nd IEEE International Conference on Biometrics: Theory, Applications and Systems, 2008, page(s) 1-7. 4. Zhai, L., Dong, S., and Ma, H. 2008. Recent Methods and Applications on Image Edge Detection 2008 international Workshop on Geoscience and Remote Sensing - Volume 01 (December 21�22, 2008). KEYWORDS: Automated Video Monitoring; Video Analytics; Command and Control; Video Display; Video Object Detection; Video Image Detection
|