Infrared Dim Small Target Detection Based On Signal Singularity Analysis

Infrared small target detection problem had been the focus and research focus in recent years at home and abroad.In the passive infrared detection system occupies an important position.As far as possible to improve the detection distance of the target,ahead of the target is to determine the key factors in the outcome of the war.Due to the distance,the weak radiation energy and other characteristics of infrared target imaging caused by weak and vulnerable to strong clutter.Therefore,how to achieve reliable detection of the target in complex background environment has important theoretical and military significance.Despite the complex background dim target exhibits a low signal to noise ratio and other characteristics.But the target relative to the local background will appear signal amplitude mutation,showing a singularity.According to the small target in the space-time domain to show this singularity can achieve the target signal detection.Therefore,this paper studied the singularity detection theory of non-stationary signals from the non-stationarity and singularity of the weak signal itself,and found a method to detect singularity of weak signal.Which was applied to the detection of weak infrared target with low SNR.A new method based on signal singularity detection was proposed.The main work is as follows:1.Summed up the status of small target detection research.Based on the airspace and time domain of infrared image,the characteristics of infrared image were analyzed and the corresponding mathematical model was established.The theory of signal singularity was analyzed.The singularity of the small target was analyzed,and the singularity of the small infrared target was analyzed and the model is established.2.Several methods of signal preprocessing are studied,including morphological filtering,local background prediction filtering,median filtering and improved median Wiener filtering.Which provides a pretreatment means for small target detection of subsequent infrared sequence images.3.According to the study of the characteristics of the time domain signal of the sequence image and the singularity characteristics of the target in the time domain signal.We proposed a new algorithm for detecting the singularity of the infrared small target,which was based on the new improved median Wiener filter.4.Complex signals can be decomposed into intrinsic modal functions at different frequency scales,and it is possible to detect significant changes such as nonstationary signals and abrupt signals.The empirical modal decomposition method was applied to the small target detection problem of infrared image.The components of different frequencies were obtained by empirical modal decomposition of rows and columns of single frame infrared images.And these components were multi-level joint to achieve the purpose of detecting targets.Finally,the method was applied to the time domain signal of the infrared sequence image.The algorithm of component synthesis was used to detect the infrared weak target.Simulation and experimental results show that the proposed method was effective and had good detection performance,and had great application prospect.