Most defense and security sensors perform search or monitoring functions. They may key on weapons or other threat objects, on backpacks or clothing articles for search & rescue, or on custom or covert tags for asset tracking. While a diversity of machine-vision algorithms have been developed to find targets of interest efficiently, and in many cases automatically, machine-learning algorithms are computationally-intensive and usually work well only when the scenario is relatively well-controlled, like in a manufacturing environment. Scenarios where high-performance computing or cloud computing are not available or not reliable, or where the imaging conditions are not well-controlled, must rely on simpler, more robust classification algorithms. Simpler algorithms can achieve comparable performance if the sensor provides MATERIAL-BASED features. Rather than providing just a high-resolution image, the sensor can provide an image that highlights only materials of interest; it does so by measuring material signatures known a priori from lab measurements and/or physical models. Some sensors that can do this are multi-spectral or hyperspectral imagers and passive imaging polarimeters, which rely on the sun as their light source.