In-frame and inter-frame information based infrared moving small target detection under complex cloud backgrounds

Infrared moving small target detection under complex cloud backgrounds is one of the key techniques of infrared search and track (IRST) systems. This paper proposes a novel method based on in-frame inter-frame information to detect infrared moving small targets accurately. For a single frame, in the spatial domain, a directional max-median filter is developed to make a pre-processing and a background suppression filtering template is utilized on the denoised image to highlight target. Then, targets in cloud regions and non-cloud regions are extracted by different thresholds according to a cloud discrimination method so that a spatial domain map (SDM) is acquired. In the frequency domain, we design an α-DoB band-pass filter to conduct coarse saliency detection and make an amplitude transformation with smoothing processing which is the so-called elaborate saliency detection. Furthermore, a frequency domain map (FDM) is acquired by an adaptive binary segmentation method. Lastly, candidate targets in single frame are extracted by a discrimination based on intensity and spatial distance criteria. For consecutive frames, a false alarm suppression is conducted on account of differences of motion features between moving target and false alarms to improve detection accuracy again. Large numbers of experiments demonstrate that the proposed method has satisfying detection effectiveness and robustness for infrared moving small target detection under complex cloud backgrounds.