Research On Infrared Image Enhancement Algorithm Based On Human Visual Characteristics

Infrared thermal imaging technology plays a key role in many fields such as medical,military and civilian.But due to the influence of multiple factors such as the special imaging process and the device itself,infrared images have problems such as blurred details,low contrast,and poor visual effects.Infrared image enhancement processing has become a necessary means.The traditional algorithm only considers the mathematical indicators in the infrared image,ignores the visual mechanism,which is not conducive to visual observation.Therefore,it is a great research significance to enhance infrared images from the human visual angle.In this paper,in order to more accurately reflect the characteristics of the image and consider the visual mechanism more comprehensively.For the first time,multiple visual characteristics are introduced into the enhancement algorithm,and the infrared image is processed by a combination of multiple algorithms.First,analyze the infrared thermal imaging mechanism and infrared image characteristics,study the common infrared image enhancement algorithms,pointing out the advantages and disadvantages of these algorithms in image enhancement.Secondly,analyze the human visual system,study visual theory and visual characteristics,focus on the application of visual characteristics in image salient regions and image contrast.Point out the correlation between infrared images and visual characteristics,and propose infrared image enhancement algorithm based on human visual characteristics.This paper applies the principle of visual lateral inhibition,human salient characteristics and human perception characteristics to infrared image enhancement technology,and proposes two infrared image enhancement algorithms.Firstly,for the problem of blurred infrared image and poor resolution,an infrared image enhancement algorithm based on adaptive multi-scale side suppression is proposed.The algorithm uses the visual lateral inhibition principle to establish a network model,changes the bimodal Gaussian function to control the lateral inhibition coefficient distribution,and uses the image pixel grayscale to adaptively adjust the size of the suppression field,which effectively improves the image contour enhancement effect and increases the image gray contrast.Then,in view of the problem that the infrared image has poor ability to highlight details and low contrast,an infrar ed image enhancement algorithm based on visual salience is proposed.The algorithm first uses the local linear characteristics of visual perception to establish the human visual system(HVS)model,highlight the visual Weber region in the image,and uses bilateral filtering to establish a layered model of the image to achieve the layering of the image.The HVS model and layered model is combined to obtain a new image detail region and background region.The improved adaptively increasing the value of histogram equalization(IAIVHE)algorithm is used to process the detail region to improve the image contrast and enhance the detailed texture.The adaptive multi-scale lateral inhibition algorithm is applied to the background region to suppresses image noise,increases gray contrast,highlights the target outline,and merges the enhanced background region with the detail region to obtain a new infrared image.Finally,on the basis of fully studying the theory of the algorithm,the algorithm of this paper is emulated experimentally.The subjective and objective evaluation indicators are used to analyze the enhancement effect of this algorithm.The experiment is compared with the common algorithms.Experiments show that the proposed lateral inhibition algorithm can effectively suppress the noise and highlight the image contour information;the proposed layered algorithm can effectively improve the image contrast,enhance the image detail texture,which have a good visual effect.