Physics-Based Radiometric Signature Modeling and Detection Algorithms of Landmines Using Electro-Optical Sensors
OHIO STATE UNIV COLUMBUS
Pagination or Media Count:
This work, supported by the USA RDECOM CERDEC Night Vision and Electronic Sensors Directorate, focuses on signature modeling and detection algorithms. Signature modeling helps to provide insight for sensor deployment. The model addresses relevant issues in sources, targets, and sensors. Natural sources such as thermal emission, solar radiation, and solar scattering were considered and incorporated using empirical models. A BRDF model that defines scattering and emission from rough surfaces was developed that integrates geometric relations with intrinsic surfaces properties. Stokes vectors are used throughout this work to describe incident and scattered radiances, which permits a polarimetric study of the signatures. The simulated signatures are compared with several measured data sets from different scenarios and exhibit strong quantitative agreement. The advancements described throughout the document will serve to improve real-time mine detection.
- Test Facilities, Equipment and Methods