Research On Infrared Video Enhancement Algorithm Based On Retinex Theory

Infrared imaging technology is not affected by light conditions,and can work all day.And the anti-interference ability is strong.Therefore,today's infrared imaging technology has been widely used,such as military,medical,agricultural and industrial fields.But the current infrared imaging equipment,often 14 or 16 bits,is not easy to display on the current display device.Secondly,because the temperature difference between objects is small,the gray value range of all infrared images is narrow,resulting in low contrast of infrared images;because the temperature difference between different parts of the same object is small,resulting in unclear details.The purpose of this paper is to enhance the contrast and detail of infrared image and infrared video,focusing on the local adaptive processing infrared image enhancement algorithm based on Retinex decomposition.Based on this algorithm,the optical flow method is improved to adapt to the enhancement of infrared video.The main innovation and work of this paper are as follows:(1)Infrared image has low contrast,fuzzy details and low signal-to-noise ratio.In order to obtain infrared image with good visual effect,this paper proposes a local adaptive processing infrared image enhancement algorithm based on Retinex decomposition.Firstly,in order to enhance the contrast of the infrared image,this paper modifies the descending histogram of the image through the geometric distribution,and updates the local mapping function adaptively based on the global mapping function to compress the dynamic range and enhance the contrast of the original infrared image.Secondly,in order to enhance the details of the infrared image,this paper decomposes the contrast enhanced infrared image into illumination component and reflection component through the joint inner and outer prior model jiep of Retinex.The reflection component reflects the essence of the object,which contains a lot of details and noise.Then the reflection component is input into the boundary preserving filter mugif to obtain the base layer image and detail layer image.By defining an adaptive weight parameter,the detail of the reflection component is enhanced and the noise amplification in the process of detail enhancement is suppressed.Then the detail enhanced reflection component is combined with the illumination component decomposed by Retinex.Experimental results show that the algorithm can effectively enhance the details and contrast of infrared image,and suppress the noise amplification.(2)For infrared video,this paper improves the local adaptive processing infrared image enhancement algorithm based on Retinex decomposition,and proposes an infrared video enhancement algorithm based on optical flow method,which makes it more suitable for infrared video enhancement.First of all,in order to solve the problem that the background of the image block with the foreground moving in is over enhanced,and the background of the image block with the foreground moving out is not enhanced enough.The local adaptive contrast enhancement algorithm is changed to the global contrast enhancement algorithm,which effectively solves the above phenomenon.Secondly,in order to enhance each frame of infrared video adaptively according to the characteristics of each frame,the proportion of moving target pixels in the infrared image is estimated by optical flow method to control the combination proportion of geometric distribution histogram and global descending histogram.Finally,in order to enhance the details of the moving target,the distribution of the moving target pixels in the infrared image is estimated by the optical flow method,and the moving target in the infrared video is further enhanced on the basis of enhancing the background details.Experiments show that the algorithm can effectively enhance each frame of infrared video.