Research On The Rapid Detection Of Infrared Small Targets Under Complex Marine Backgrounds

The technology of infrared small targets detection is the basis of surface guidance application.It can effectively improve the distance and accuracy of target detection.However,there are many problems that have led to great challenges in the actual practical detection,such as the low resolution,fuzzy geometric information,weak gray-scale response.Aiming at the key problems to the detection subject on the sea surface in a complex background,the following work is proposed in this paper.In view of the low signal-to-noise ratio and the weak contrast of the observed target in the image,this paper proposed a target-oriented adaptive contrast enhancement strategy based on Gaussian fitting.The method exploits the high-frequency distribution of an image to estimate an intensity-weighting matrix,which is then used to control the Gaussian fitting curve and shape the distribution of the contrast gain.As such,the curve can be easily designed to enhance the important details hidden in noteworthy regions.Subsequently,the proposed gray-scale transformation is obtained from the Gaussian fitting can rationally express contrast distribution.Considering the variations in the distribution of gray levels,a single-peak Gaussian function sometimes cannot adequately express the scattering of these gray levels in an image.Therefore,we propose the use of piecewise Gaussian fittings,also called MGFT,to overcome this shortcoming.Unlike previous contrast enhancement methods,our technique is fully automatic and target-oriented in the sense that it can be directly applied to any input image without any parameter adjustment.The experimental results for some of the widely accepted criterions show the effectiveness and broad application prospects of the proposed algorithm.By studying the theory of "Dark Channel",we found out the limitation of the contrast enhancement performance in the infrared image dehaze method of sea surface.The differences between visible and infrared image are analyzed and compared.Then,the reason for the contrast enhancement limitation in the defogging of sea infrared image of the dark channel theory is analyzed.Aiming at the limited enhancement performance of the long-range small target,we propose a regional enhancement method based on the target-oriented gain distribution in transmittance matrix.Based on the method,we construct an optimization model,and use the model to verify the performance of defogging method on the sea surface for infrared target.In this paper,the phase encoding method is proposed for the sea-sky line detection.To solve the problem of complex background,the bilateral filter is used to preserve the edges of image and filter out the high frequency noise.The edge response is enhanced by the Laplacian of Gaussian,then edge phase encoding is realized by the constructing of directional filter,then,competitive strategy is used to construct phase components to extract the scan lines.According to the analysis of phase components,the adjacent scan lines are fused together.In the end,the sea-sky line is detected by the comparison of scan lines'corresponding phase response intensity.The experimental results show that this proposed method can effectively detect the sea-sky line in different complex background.In this paper,the gray morphological filtering theory is used for the target contrast enhancement.The principle of peak detection based on second derivative function is analyzed,and an anisotropic filter with sound mathematical derivation is constructed by the subtraction of anisotropic Gaussian function.We analyze and verify the selectivity of anisotropic Gaussian difference filter to the scale and direction of a target.Furthermore,we designed the multi-directional and multi-scale filter groups to achieve directional filtering for targets with specific size and geometric shape.Experiments show that the strong anti-noise ability of the proposed filter.The strategy of signal-to-noise ratio improving to small targets is studied based on the fusion of spatial and time.We statistically analyze the principle of dynamic programming strategy in small target detection.Then,an energy accumulation enhancement method based on target correlation is designed in the time domain.Firstly,the direction probability distribution of the target is extracted by using anisotropic Gaussian differential filter.Then,by similarity evaluation function of inter-frame targets,we propose a method that combines the local inter-frame motion of targets in spatial domain and the probabilistic characteristics of target direction.This method combines the spatial location information and the weak surface gray features of the target,and effectively overcomes the matching constraints caused by the randomness and motion.The result of experiments proved the effectiveness of the energy accumulation strategy for the small targets detection in complex marine backgrounds.