Robust long range target detection algorithm using adaptive selective top-hat transform

Airborne threats like aircrafts and missiles at long ranges appear as point or small targets in visible and infra-red image sequences. Size of the target in input image increases as it approaches towards the imaging sensor. Mathematical morphology is used to detect such targets but its performance is highly dependent on shape and size of structuring element. In this paper, we propose a novel robust algorithm for detection of long range approaching targets in visible and infra-red image sequences. The proposed detection algorithm is based on the novel adaptive selective double structuring element top-hat transform (Adapt-Sel-DSTHT) and maximum correlation criterion. A switching logic based on variance of the input image is used to apply the additional rule to modify the filtered image for highly cluttered images. Results demonstrate high probability of detection and low false alarm even for highly clouded scenario. Reduced number of operations led to suitability for real time implementation.