A Multistage Process Method for Detecting Small Targets in Infrared Image Sequences

Yao Xun,Li Dehua,Sun Xianbin,Huang Fei
DOI: https://doi.org/10.3963/j.issn.2095-3844.2008.06.044
2008-01-01
Abstract:This paper proposes a method for detection of small targets in infrared image sequences.The proposed method consists of two serial processing stages-the predetection processing stage and clutter rejection processing stage.The predetection algorithm processes the entire image,in a low computational cost manner,to extract the potential targets based on the saliency of visual feature in each segmented region of image.Subsequently,the clutter rejection operator processes those potential targets passed from first stage using double average gray value filter method to discard spurious targets and identify true targets.A statistic analysis is given to justify the validity and efficiency of the method.The computation complexity of the algorithm is analyzed and compared with other methods.Experimental results suggest that the proposed combining computation architecture of predetection and clutter rejection modules exhibits excellent performance for detecting small target in infrared image sequences,especially in process speed aspect.
What problem does this paper attempt to address?