Research On Infrared Small Target Detection Method Under Complex Background

Infrared small target detection technology is an important part of infrared early warning system,but the application of this technology in complex backgrounds still faces many challenges.These challenges include the difficulty of suppressing background clutter under strong ground clutter or strong cloud clutter,and the problems of intersection and splitting during the movement of small targets.Therefore,this paper has conducted in-depth research on the above issues in this field,and the specific content is as follows:The image surface distribution model of small targets is derived based on the infrared imaging process,and the validity of the model is verified by fitting experiments.A background clutter quantization model based on gradient mean is proposed through multi-scale difference analysis of background clutter.The validity of the model is verified by comparison with the traditional statistical variance model.Based on the above experimental analysis,the gradient comparison of the small target and the background clutter in four directions is performed,and the distribution difference between the two is obtained.Aiming at the shortcomings of the traditional detection algorithm's insufficient background suppression ability,which is prone to false alarms in complex scenes,this paper proposes a background suppression algorithm based on the local derivative and neighborhood difference based on the difference in gradient distribution of small targets and background clutter in multiple directions.In the algorithm,the local derivative graph of the original image is obtained by constructing the structure element,and then the extreme value of the center point of the structure element is represented by the neighborhood range.Finally,the range graph is fused to obtain the small target saliency graph.The experimental results show that the algorithm proposed in this paper have strong background suppression ability,and have excellent performance in quantitative indicators such as signal-to-clutter ratio,background suppression factor and ROC curve.Aiming at the insufficient background suppression ability of traditional detection algorithms,this paper proposes a background suppression algorithm based on the local derivative and the neighborhood extreme difference based on the differences in the gradient distribution of small targets and background clutter in multiple directions.Firstly,the algorithm uses multiple characteristics of the small infrared target to match the track with the suspected target,and completes the operations of track generation,update,and deletion through the track update mechanism.Then obtain the true target information according to the track judgment.Finally,simulation experiments are performed on the algorithm using multiple sets of image sequences.The experimental results show that the algorithm can control the amount of calculation by adjusting the threshold,and at the same time,it can solve the problem of target detection failure due to intersection and splitting during the movement of the target,and control the detection accuracy within one pixel.