Research On Fusion Method Of Water Target Image Based On Infrared And Visible

The infrared sensor imaging principle is thermal radiation imaging,which is not affected by fog,cloudy weather and so on,but the images obtained are relatively fuzzy,and the spatial resolution and image contrast are low;the visible light sensor uses visible light reflection imaging,and the obtained image has clear texture,but it is easily disturbed by the environment such as fog and rain.When a smart boat or rescue boat is sailing on the surface of the water,the specific location of the person in need of assistance in the water cannot be identified when fog or visible light imaging is not clear.Infrared sensors can recognize the heat radiation but cannot clearly know the position of the person.Therefore,the infrared and visible light image fusion methods of the images collected by the surface vessel are studied in depth.Based on the above background,this paper uses infrared and visible light sensors to fuse the collected water image information.This article mainly includes the following four aspects:1.First of all,this article mainly introduces the three levels of image fusion(pixel level,feature level,and decision level).It also introduces the principles of visible light and infrared imaging,and several commonly used image fusion methods.Mature image fusion method are introduced in the spatial domain and transform domain.It also shows the subjective evaluation and objective evaluation indicators of the fusion picture,in order to evaluate the fusion picture later.2.In addition,this article mainly introduces the basic principle of NSCT which is consist of NSPFB and NSDFB.What is more,this article expounds the principles of image fusion based on NSCT and fusion as an example,and by changing the threshold to find the best fusion threshold.At the same time,the basic principle and simplified model of Pulse coupled neural networks(PCNN)are introduced.3.Next,based on the FT algorithm,this chapter proposes a saliency detection algorithm suitable for infrared images and applies it to the fusion of infrared and visible light images.The saliency map normalized values and the low-frequency coefficients of the NSCT were weighted and averaged to achieve optimization,and compared with other mature methods in the subjective and objectively to evaluate the images after the image fusion.4.Finally,based on the PCNN-based algorithm,we select the regional energy that can reflect the local characteristics of the image as the link strength ? of the neuron.After the source image is transformed by NSCT,the regional variance is used as the low frequency;PCNN model is used as a rule for selecting the subband coefficients in the bandpass direction to realize the algorithm fusion of PCNN and NSCT.Compiled by MATLAB to display the fused image and compared with other mature methods in the subjective and objectively to evaluate the images after the image fusion.