OMNIDIRECTIONAL MIRROR GRADIENT DISSIMILARITY FOR INFRARED SMALL TARGET DETECTION

Infrared small target detection (IRSTD) is a challenging task in the remote sensing field, due to the low signal-to-clutter ratio (SCR) and intricate background. Although local contrast measure (LCM) has been widely applied for IRSTD, the existing LCM-based methods suffer from high sensitiveness to heavy clutters. To resolve this problem, this paper proposes a novel LCM named omnidirectional mirror gradient dissimilarity measure (O-MGDM) for small target enhancement. A concept of mirror gradient dissimilarity (MGD) is first presented to evaluate the pixel gradient inconsistency of opposite directions. It is analyzed that small target pixels exhibit considerable MGD in all directions, while heavy clutters hold large MGD in certain directions. Then, the novel O-MGDM is proposed by uniting the omnidirectional MGDs, which achieves target enhancement and heavy clutter suppression simultaneously. After that, an adaptive threshold is used to segment the targets readily. Extensive experimental results on practical data sets demonstrate that the proposed O-MGDM achieves advanced detection performance.