Fusion of Stokes vector imagery using simple logical operators: application to the problem of surface land mine detection

Abstract

Efficient surface and subsurface land mine detection is still one of the several important tasks in the civilian and military areas. Often information derived from multiple sensors are fused together to lead a net gain in probability of detection of these land mines. Some drawbacks of this approach are the different sensor output formats and the overall processing efficiency of the system. To overcome some of the aforementioned concerns we propose a simple and efficient image fusion using logical operators. The proposed technique uses logical operators to fuse multiple images derived from a single sensor. The logical fusion operation is applied to the output of the sensor to lead a logically fused image. This image serves as an input to the detection algorithm. The proposed technique is shown to improve the probability of detection of land mines. This improvement is shown using the ROC curve approach.