Research On Infrared Maritime Small Target Detection

Infrared maritime small target detection is of great scientific research and practical significance for safeguarding national sovereignty,protecting marine rights and promoting the construction of a powerful country.However,the sea background is complex which usually contains a large number of dynamic and static background clutter,such as sea-sky line,island,fish scale light,and etc,which brings great challenge to target detection.In the single frame detection,it is difficult to suppress the background clutter effectively only based on the image basic gray and gradient characteristics,especially the dynamic clutter with similar basic characteristics to the target;in the multi frame detection,due to the unavoidable inter frame jitter in the sea infrared image sequence,the traditional pipeline filtering method has low accuracy and poor robustness.In view of the above problems,this paper analyzes the characteristics of sea infrared image in detail,and designs single frame detection algorithm and multi frame detection algorithm of infrared small target in the sea scenary according to the analysis results.The single frame detection algorithm of infrared maritime small target in this paper can detect the target according to the difference of the structural characteristics between the target and the background clutter,which can suppress a large number of static background and some dynamic clutter.The algorithm first designs a similarity measurement method CLSM,and then searches the most similar image block in the neighborhood according to the measurement method,carries on the weighted filtering to the current pixel,which can effectively use the structural characteristics difference to realize the static background estimation;then uses the same measurement method CLSM to measure the compressed contour feature similarity between image block and its neighborhood,and suppresses the similar structures.Finally,target detection is realized.A large number of experimental results show that the algorithm can suppress the dynamic and static background in a variety of scenes,while retaining the characteristics of the target,which can effectively improve the detection rate and reduce the false alarm rate,but the dynamic clutters can not be suppressed utterly.On the basis of single frame detection,the multi frame detection algorithm of infrared maritime small target in this paper can detect the target according to the behavior difference between the target and the dynamic clutter,which can further suppress the dynamic clutter with similar structure characteristics to reduce the false alarm.Firstly,a grid line structure which is compatible with the above similarity measurement method CLSM and can expand the receptive field is designed,and then the inter frame jitter is compensated according to the sea-sky line coarse tuning strategy and the grid line fine tuning strategy;and then,based on the time-domain behavior characteristics of the target and dynamic clutter,the spatiotemporal saliency map is generated according to the detection results of multiple single frames;finally,the target detection is realized.A large number of experimental results show that the algorithm can effectively detect the target in a variety of scenes,and can further suppress the dynamic clutter and reduce the false alarm rate compared with single frame detection,but its efficiency should be enhanced furtherly.