A Method for Infrared Sea-Sky Condition Judgment and Search System: Robust Target Detection via PLS and CEDoG

It has always been a serious challenge to efficiently detect infrared targets from remote sea-sky background without any prior knowledge. This is especially true when target features of different areas are irrelevant in the same image. The main contribution of this paper is to design an algorithm of judge sea-sky condition and detect infrared target, in which a two-direction local maximum (TLM) method and peak local singularity (PLS) is proposed, constant excitatory difference of Gaussians (CEDoG) operator is applied by analyzing the characteristics of the target. First and foremost, the TLM method is adopted to refine seasky area to extract the suspected sea-sky line, and design the strategy of “de-false”to realize the accurate determination of the sea-sky line. Secondly, small targets in the sea-sky area (pixels 2 × 2 to 9 × 9) are detected by applying the PLS method. Finally, CEDoG filtering method is used to suppress the background and improve the significance of the target, and high threshold OTSU method (H-OTSU) is applied to find the most significant area and self-designed area growth rule to ensure the accuracy and integrity of the maritime area target detection. Comparing with the other state-of-the-art methods in the experiments, our strategy has a robust and effective performance in terms of recall, precision, elapsed time, complexity, detection and false alarm rate.