Infrared Small Moving Target Detection, Recognition Technology

With the development of infrared imaging, infrared imaging system has been widely used for the fields of civil and military use, such as optical remote sensing, night navigation, target detection, fire control and guidance. Infrared small moving target detection technology, one of the key techniques of infrared imaging system, adopts imaging processing algorithm whose performance is critical for action distance and intelligence of infrared imaging system, to detect target automatically which is under random background and heavy noise. Especially in the field of military, infrared small moving target detection technology holds an important position in advanced weapon system of Occident, including forewarn system of aeroplane carrier, infrared searching and tracing system of varies airplanes, infrared imaging guidance missile and warning system of some ground military facilities. It can be said that automatic detection of remote airplane and missile has been the key techniques for imaging guidance, warning system and photoelectricity resist.The detection of small target in infrared image is a problem involving time and space. Supposed that presenting an image, it just refers to the problem of space, namely, confirming that whether the target exists or not in two-dimensional space and confirming its position. Time detection often exists in real detection system, namely, detect target out within presented time. It can also be considered as sequential images detection problem which differs from single frame detection. So sequential images detection and single image detection are two different problems of small target detection algorithm which should be with corresponding algorithms. Therefore, this paper researches: detects small target through suppressing background based the characters of small target in infrared image. In terms of single image, we adopt mathematic morphology Top-hat transform to suppress background, then get image which holds small target;in terms of sequential images, adopt neighboring region verdict to suppress background in order to insure the reliability of target detection, then get moving track of small target.When the sensor has a certain distance from target, we construct image character which has inflexibility after extracting target image to identify. And we compare thecharacter of normalized moment of inertia of extracting target image and stencil image to ensure the target kind. All the analysis is combined with real infrared image to carry out simulation experiments and experiment results verify the champion characters of various algorithms and conclusion.