Approach to Enhancing Target Discrimination via 3D Visualization without 3D Glasses

Intelligent Automation, Inc. (IAI) and Arizona State University (ASU) present a systematic approach to stereo image creation, depth estimation, free-viewpoint 3D visualization, and 3D target modeling and visualization. The approach consists of automatic feature point extraction, correspondence and tracking, and stereo image generation from uncalibrated cameras. Additional postprocessing algorithms such as 3D image/free-viewpoint reconstruction, depth estimation, super-resolution image generation, etc. will be customized from our past algorithms. The Phase 1 development of the approach has proven high accuracy and obvious advantages over any existing state-of-the-art algorithm. We have achieved a lot of results in Phase 1. Extensive tests including in-door/out-door scenarios were performed to improve and optimize our proposed algorithm. 3D video tests for moving target inspection were also performed. Current testing results demonstrate the robustness and accuracy of our algorithm on stereo image creation with uncalibrated cameras. In Phase 2, we will further enhanced and expanded the Phase 1 results. It is envisioned that a prototype system will generate more rich and highly accurate visual information for the human operator to perform better in target detection and discrimination. Real-time demonstrations of our prototype system for different scenarios will be performed at MDA.