In infrared detection, we utilize the infrared radiation emitted from target to detect and identify the target we need as
soon as possible, so we can get more reaction time for on the battlefield. However, because of the long distance between
target and the sensor, the area of the image into target is very small. Infrared small target detection has become a difficult
area of target recognition. In order to improve the accuracy of target detection,, we should preprocessing the infrared
image before identifying target, including noise reduction, background suppression and target enhancement, after that,
we can lay the foundation for the recognition processing. Firstly, we should filter the image by using median filter.
Because the median filter not only filter noise effectively (especially the impulse noise) but also maintain the goal of
good details, and the operation of this method is simple and faster. Secondly, we want to suppress the background. In
describing the signal characteristics, mathematical morphology has unique advantages, so we suppress the image
background in this way. Thirdly, after the above treatment, we can use adaptive threshold to achieve the objectives of the
segmentation, and get the binary image obtained.