A Detecting Algorithm of Infrared Armor Target under Complex Ground Background Based on Morphological Wavelet

摘要

Detecting of infrared armor target is the important technology in watching ground target and technological reconnaissance field. In most cases, the target is supposed either darker or brighter than its immediate adjacent background thus a possibility is provided to detect infrared target. How to enhance the character of target area becomes a hot research area because of its complex ground background. This paper advances a novel target detection method based on morphological wavelet decomposition through calculating a global threshold. Its main idea is explained as follow: First we set a global threshold value according to analyzing the distribution character of the image's histogram; second, we decompose the image using the two-dimensional morphological haar wavelet according to this value: when the maximum value is greater than it we use the maximum value of four pixels, and else we use the minimum; third, we carry out differential operation about the detailed image including the horizontal and vertical image and make them into binary image by a threshold value; at last we decompose the binary images and use their low frequency part to find the target area by calculate the bolck's weight. Through all of the steps above, the fast detection on infrared armor target under complex ground background can be realized successfully. The algorithm has the next advantages: strong ability of detecting objects, small operation quantity and good real-time performance etc. So it can be used widely and effectively in actual work.