Real-time automatic small target detection using saliency extraction and morphological theory

We propose a real-time target detection approach in infrared imagery. This method combined saliency detection technology and morphological theory. A local frequency-tuned based saliency extraction technology is designed for highlighting the potential targets area. With saliency extraction and thresholding processing, the potential targets map is obtained, and then morphological operation is used for noise suppression. The experimental results prove the proposed approach is robust and efficient for target detection.

► It is a real-time target detection approach for infrared image. ► The theory of local frequency-tuned based saliency extraction is proposed. ► The saliency extraction method can highlight potential targets from background clutter. ► The approach works accurately and efficiently, with low false alarm.