Multi-Target Track and ID with Persistent Hyperspectral Data
Small Business Information
4850 Hahns Peak Drive, Suite 200, Loveland, CO, 80538
AbstractMilitary operations in urban warfare provide an added emphasis to effectively detect, track, and ID ground targets in challenging environments. Given the high dynamic nature of ground targets and the ambiguity that may result from closely spaced targets, incorporation of feature data from sensors such as hyperpsectral imagery (HSI) cameras provides a means to disambiguate the tracking of these targets. The benefit of using these sensors in multi-target tracking is the ability to build feature models for target tracks based on the spectral information over multiple wavelengths. The video-based tracking community has demonstrated the ability to resolve closely spaced targets by incorporating color features versus intensity data alone. The addition of 20 to 200 additional wavelengths has the potential to significantly improve a target tracker's ability to distinguish among targets. The Phase I effort will demonstrate the ability of our algorithms to (i) simulate an urban scene and embedded targets along with models to generate sensor measurement data (ii) generate detections from HSI imagery based on a combination of motion and feature segmentation (iii) mitigate the effects of spectral smearing (iv) perform HSI feature-aided tracking of multiple targets in an urban setting using both real and simulated data.
* information listed above is at the time of submission.