Automatic Detection Method Of Infrared Ship Target Under Complex Sea-Sky Background

Infrared technology,with the advantage of long distance imaging, good concealment, all-day working, has been widely used in missile guidance, military reconnaissance and other fields. However, in the anti-ship guidance, the sky and sea backgrounds are susceptible to season, climate, temperature and other factors, especially in the early long-range imaging the ship target is lack of the features such as size, shape, texture and so on, so the ship target detection in the context of complex sky and sea is very difficult. However, the ship target is always located in the sea-sky area, which can narrow the search range and suppress the unnecessary noise outside the area.In this paper, we first analyze the infrared image characteristics and background suppression algorithms, then we focus on the methods of sea-sky-line extraction and small target detection. According to the gray-scale transition character of sea-sky area, we propose a sea-sky-line locating method based on the anticlimax of horizontal mean curve of the background. After comparing and analyzing with other three sea-sky-line extraction algorithm, we find that our method has the advantage of locating the sea-sky-line quickly and accurately in the environment of fuzzy sea-sky-line, and full of ups and downs sea. Allowing for the "bright spot",as well as stability and continuity in the adjacent frames of the small ship target, we combine our segmentation algorithm self-adaptive threshold method upon neighborhood mean value with the track association to detect the small ship target, which ensures the reliability of detect results. Significantly, the self-adaptive threshold method upon neighborhood mean value can not only suppress the clutter of waves, clouds effectively, but also detect ship target with the high accuracy rate of95%.