Next Generation 4 -D Distributed Modeling and Visualization

Next Generation 4 -D Distributed Modeling and Visualization of Battlefield Avideh Zakhor UC Berkeley June, 2001

Participants y. Avideh Zakhor, (UC Berkeley) y. Bill Ribarsky, (Georgia Tech) y. Ulrich Neumann (USC) y. Pramod Varshney (Syracuse) y. Suresh Lodha (UC Santa Cruz)

Battlefield Visualization z. Detailed, timely and accurate picture of the modern battlefield vital to military z. Many sources of info to build “picture”: y. Archival data, roadmaps, GIS and databases: static y. Sensor information from mobile agents at different times and location: dynamic. y. Multiple modalities: fusion z. How to make sense of all these without information overload?

Major Challenges z Disparate/conflicting sources of information resulting in large volumes of data. z Data sources inherently uncertain: resulting models also uncertain. z Data and models together with associated uncertainty need to be visualized on mobiles with limited capability. z All of this is time varying, time dependent and dynamic.

Mobile AR Visualization Laser-------Lidar-------Radar------Camera-----GPS----Maps-------Gyroscope-- Visualization Database 3 D model extraction UNCERTAINTY Model update Decision Making Mobiles with augmented reality sensors

Research Agenda z Model construction and update: y Radar, Laser, Cameras, Roadmaps, Satellite photos, GPS, Gyros, Speedometer, INS. y Augmented reality sensor tracking and registration required z Mobile, real time visualization, interaction and navigation within the database : y Augmented reality sensor tracking and registration required; z Uncertainty processing and visualization: y Sensors, algorithms, models. z Fusion used in all of the above.

Visualization Pentagon 4 D Modeling/ Update Tracking/ Registration Information Fusion Visualization Database Uncertainty Processing/ Visualization

Model Construction z Develop a framework for fast, automatic and accurate 3 D model construction for urban areas. z Models must be: y Easy, fast, automated, to compute; y Compact to represent; y Easy to render; y Amenable to photo realistic flythroughs. z Strategy: y Fusion of multiple data sources: intensity, range, heading, speedometer, panoramic cameras. y Incorporate apriori models, e. g. aerial photos, and digital roadmaps.

3 D Modeling: z Close-range modeling: y ground based vehicle with multiple sensors z Far-range modeling: y. Aerial/satellite imagery z Fusion of close range and far range info at multiple levels: y. Data and models.

Ground Level Based Model Construction • Laser range scanners • Digital roadmaps • Aerial photos

Ground level based 3 D model construction • 3 D point cloud from 2 D scans • Mesh generation • Texture Mapping

Far Range Modeling Example of an aerial image and extracted disparity map of manmade structures

Tracking, Registration, & Autocalibration z Needed for Augmentation or Visualization z Mobile people with sensors provide textures and data for visualizations… y. Where are they? Where are they looking? y. What are they doing? z Tracking provides position and orientation y. Tracking is possible with landmarks in view y. Autocalibration allows tracking in regions beyond/near landmarks – also can improve/refine models y. Sensor fusion (GPS, vision, gyroscopes) adds more tracking information for wider areas

Mobile Testbed

Registration Using Mutual Information Original IRS and RADAR image pair Registered RADAR image

Temporal Aspects z Time critical applications: y. Sequential methods for information fusion y. Tradeoff between speed and accuracy of results(Qo. S) z Fuse information arriving at different times: y. Perform independent time-based calculations, followed by fusion/integration

Fusion of Asynchronous Information t 1 tne w Source 1 Source 2 t 2 Decision maker Source 2 tnew Projection Fusion

Mobile Visualization z Synchronized data bases of dynamic data in collaborative environments y Peer to peer mode y Sharing geo-spatial data, annotations, triggering events z Mobility: y Development of mobile visualization capability using laptop version of global geospatial system y Gesture recognition capability z New hierarchical volume rendering technique to weather and uncertainty Real-time acquisition and insertion Dynamic, Synchronized Databases

Uncertainty Computation and Visualization

Uncertainty in Visualizations of Terrains z Terrain visualization is important in a battlespace environment z Terrain compression needed due to -- real-time interaction -- limited screen space -- progressive transmission requirements z How much uncertainty is introduced due to data compression? z Can we trust compressed terrain images?

Monterey Bay original unconstrained, 20% unconstrained, 70% constrained, 80%

Target Uncertainty Visualization in 2 D Galaxy Diffuse Color Speed: Mean 1. 0; Std Dev 0. 2 Direction: Mean 0. 0; Std Dev 45

Transitions (1) z Established close relationships with both government agencies and industry: y. Government agencies x. AFRL Information Directorate, TRADOC Analysis Center x. Office of Naval Research (ONR) x. NASA (Houston Space Center) x. DARPA x. CERTIP (Center for Emergency Response Technology, Instrumentation, and Policy), y. Industry: x. HRL, Boeing, Raytheon, General Electric, HP x. Olympus America x. Niagara Mohawk Power Corp, ICT x. Carrier Corp, Andro consulting, Kitware

Publications (1) z Hassan Foroosh, “Conservative Optical Flow Fields”, submitted to International Journal of Computer Vision, October 2000. z Hassan Foroosh, “A General Motion Model based on Theory of Conservative Fields”, submitted to International Conference on Computer Vision and Pattern Recognition, May 2001. z Hassan Foroosh, “A Closed-Form Solution For Optical Flow By Imposing Temporal Constraints”, accepted in International Conference on Image Processing, January 2001. z Hassan Foroosh and Avideh Zakhor, “Virtual Cities From Aerial Images By Fusion of Segmentation And Stereo Vision in a Bayesian Framework”, submitted to International Conference on Computer Vision and Pattern Recognition , May 2001. z Hassan Foroosh and Avideh Zakhor, “Projective Rectification Using a Single Homography”, submitted to International Conference on Computer Vision and Pattern Recognition, May 2001. z Christian Frueh and Avideh Zakhor, “Fast 3 D Model Generation in Urban Environments”, accepted in Conf. on Multisensor Fusion and Integration for Intelligent Systems, February 2001.

Publications (2) z Christian Frueh and Avideh Zakhor, “ 3 D model generation for cities using aerial photographs and ground level laser scans”, submitted to International Conference on Computer Vision and Pattern Recognition, May 2001. z N. L. Chang and A. Zakhor, “A Multivalued Representation for View Synthesis”, in Proceedings of the International Conference on Image Processing, Kobe, Japan, October 1999, vol. 2, pp. 505 -509; Also submitted to International Journal of Computer Vision, January 2000 z T. Y. Jiang, William Ribarsky, Tony Wasilewski, Nickolas Faust, Brendan Hannigan, and Mitchell Parry, “Acquisition and Display of Real-Time Atmospheric Data on Terrain, ” to be published, Proceedings of Eurographics-IEEE Visualization Symposium 2001. z Mitchell Parry, Brendan Hannigan, William Ribarsky, T. Y. Jiang, and Nickolas Faust, “Hierarchical Storage and Visualization of Real-Time 3 D Data, ” to be published, Proceedings of SPIE Aerosense 2001. z Mitchell Parry, William Ribarsky, Chris Shaw, Tony Wasilewski, Nickolas Faust, T. Y. Jiang, and David Krum, “Acquisition, Management, and Use of Large-Scale, Dynamic Geospatial Data, ” submitted to IEEE Computer Graphics & Applications

Publications (3) z J. W. Lee, S. You, and U. Neumann. “Large Motion Estimation for Omnidirectional Vision, ” IEEE Workshop on Omnidirectional Vision, with CVPR, June 2000 z J. W. Lee, U. Neumann. "Motion Estimation with Incomplete Information Using Omni. Directional Vision, " IEEE International Conference on Image Processing (ICIP'00), Vancouver Canada, September 2000 z S. You, U. Neumann. "Fusion of Vision and Gyro Tracking for Robust Augmented Reality Registration, " IEEE VR 2001, pp. 71 -78, Yokahama Japan, March 2001 z B. Jiang, U. Neumann, “Extendible Tracking by Line Auto-Calibration, ” submitted to ISAR 2001 z Antonio Hasbun, Pramod K. Varshney, Kishan G. Mehrotra and C. K. Mohan, “Uncertainty Introduced by Quantization in Image Processing Applications”, Under preparation. z Kishan G. Mehrotra, C. K. Mohan and Pramod K. Varshney, “Theory and Applications of Random Sets: A Review”, Under preparation.

Publications (4) z Suresh K. Lodha, Krishna M. Roskin, and Jose Renteria, ``Hierarchical Topology Preserving Compression of 2 D Terrains'', Submitted for publication to Computer Graphics Forum. z Suresh K. Lodha and Arvind Verma, ``Spatio-Temporal Visualization of Urban Crimes on a GIS Grid'', Proceedings of the ACM GIS Conference, November 2000, ACM Press, pages 174 --179. z Suresh K. Lodha, Jose Renteria and Krishna M. Roskin, ``Topology Preserving Compression of 2 D Vector Fields'', Proceedings of IEEE Visualization '2000, October 2000, pp. 343 --350. z Christopher Campbell, Michael Shafae, Suresh K. Lodha and D. Massaro, ``Multimodal Representations for the Exploration of Multidimensional Fuzzy Data", Submitted for publication to Behavior Research Methods, Instruments, and Computers.

Publication (5) z z z Q. Zhang and P. K. Varshney, “Decentralized M-ary Detection via Hierarchical Binary Decision Fusion”, Information Fusion, Vol. 2, pp. 3 -16, March 2001. Hua-mei Chen and Pramod K. Varshney, “A pyramid approach for multimodality image registration based on mutual information”, Proceedings of 3 rd international conference on information fusion, vol. I, pp. Mo. D 3 0 -15, Paris, July 2000. Hua-mei Chen and Pramod K. Varshney, “Generalized partial volume interpolation for image registration based on mutual information”, presented at 2000 WESTERN NEW YORK, IMAGE PROCESSING WORKSHOP, University of Rochester, October 13, 2000. Hua-mei Chen and Pramod K. Varshney, ”A cooperative search algorithm for mutual information based image registration, ” Proc. of Aerosense 2001, Orlando, April 2001. Andrew L. Drozd et al. , “ Towards the development of multisensor integrated display systems, ” To be presented at Fusion 2001, Montreal, Aug. 2001. David Krum, William Ribarsky, Chris Shaw, Larry Hodges, and Nickolas Faust, “Situational Visualization, ” submitted to ACM Virtual Reality Software and Technology.

Cross Collaboration UCB USC G. T. SYR Model const. & update X x x x Tracking & reg. x X x x Mobile visual. database x x X x x Uncertain. processing x x X x x X Uncertain. Visualizati on. x UCSC

Outline of Talks z U. C. Berkeley, "3 D model construction for urban environments“ z USC, " Tracking and Visualization with Models and Images" z Georgia Tech, "Mobile Visualization in a Dynamic, Augmented Battlespace" z Syracuse, "Uncertainty Processing and Information Fusion for Visualization" z U. C. Santa Cruz, "Uncertainty Quantification and Visualization: Terrains and Targets"

3 D Model Construction for Urban Environments z Close-range modeling: y ground based vehicle with multiple sensors z Far-range modeling: y. Aerial/satellite imagery z Fusion of close range and far range info at multiple levels: y. Data and models.