Automatic Boresight Alignment of Optical Sensors

Modern shipboard video sensor systems achieve full coverage of their surroundings in multiple complementary sensing modalities through a distributed network of cameras with varying capabilities.  To form an integrated picture of the local environment, it is necessary to combine information gathered across multiple discrete sensor systems.  To accurately predict where a target in one sensor FOV is located in another separate sensor FOV requires, at a minimum, precise knowledge of the relative positions and orientations of the two sensors.  Modern sensors have high resolutions and must maintain very precise measurements for boresight and related values to get accurate sensor-to-sensor mappings.  An automated system for generating measurements and computing sensor boresights will be key to maintaining operational systems.  Toyon Research Corporation proposes to apply our expertise in developing innovative video bore-sighting capabilities for maritime application.  Specifically, the proposed effort includes (1) obtaining a test data set with multiple sensor modes from a shoreline environment, (2) developing sensor boresight monitoring and estimation algorithms using maritime features, (3) developing cross-sensor and illuminator estimation monitoring and refinement methods, (4) investigating alternatives for projected feature creation, and (5) creating a recommended system design including standard-compliant input and output interfaces.