Research On Image Enhancement Technology Of Cable Tunnel

In the cable tunnel,due to uneven lighting,there will be problems such as poor image quality,too low contrast,too low illumination,and overexposure to backlight,which will bring serious troubles to the cable tunnel monitoring system.Image enhancement has the characteristics of significantly improving image quality,so it is widely used in machine vision,atmosphere,optics and many other fields,and has great research value.Due to the complex lighting environment of cable tunnels,this paper compares multiple traditional image enhancement methods and finds that the MSRCR algorithm based on Retinex theory has a good enhancement effect in cable tunnel scenarios.Since the MSRCR(Multi-Scale Retinex with Color Restore)algorithm is not specifically proposed for cable tunnel scenarios,it is not fully applicable in cable tunneling,such as poor image enhancement for backlit shots,and not ideal for flickering and dark image enhancement effects.In view of its unsuitability in cable tunneling,this paper proposes an image enhancement algorithm for cable tunnels based on Retinex theory,named MSRCR-CT(Multi-Scale Retinex with Color Restore of Cable Tunnel).The specific work of this article is as follows:Firstly,the basic theory of image enhancement algorithm,including histogram equalization and traditional algorithm based on Retinex theory,is introduced,which lays the foundation for subsequent research.Enhancing the color restoration of the image is also very important,so the MSRCR algorithm with color recovery factor is used as the basic algorithm to improve,and the improvement work is as follows:(1)Aiming at the problem of large difference in brightness between light and dark areas in cable tunnels,under the action of Gaussian factor,the guided filter introduced into the Canny operator is added to better estimate the illumination,effectively reduce the edge blur and the halo of the excessive area,so that the enhanced image retains more edge information.(2)In view of the loss of detail in the brightness area of the enhanced image,the logarithmic processing inverse transformation LIPS(Logarithmic Image Processing Subtraction)is added on the basis of the MSRCR algorithm to enhance the brightness region,and the brightness region is effectively stretched,so as to retain more details.(3)The introduction of Gamma correction on the basis of the existing algorithm has three main functions: first,effectively enhance the contrast of the image;Second,inhibit backlight overexposure;Third,adjust the average value of the RGB channel to make the image color more balanced.Then,due to the incomplete lighting environment of the images captured in the cable tunnel,the algorithm was validated by establishing a test experimental dataset.The test experimental dataset is based on the existing dataset,which filters the pixel distribution in the grayscale histogram of each image to obtain a specific image.In addition,the obtained images are classified according to the lighting environment,so that they correspond to cable tunnel images under different lighting environments,such as low illumination,overexposure,etc.The on-site experimental data was directly captured in the cable tunnel.Finally,NIQE(Natural Image Quality Evaluator)was used to objectively analyze the original image,the enhanced image of the existing algorithm and the enhanced image of the improved algorithm MSRCR-CT.This method enables the quality of the enhanced image to be analyzed without any reference image.In addition,it has a good correlation with the subjective perception of human vision,which can prove the feasibility of the algorithm from an objective perspective.Experiments show that the improved algorithm has a good enhancement effect on low-light images,backlight overexposed images,flickering light and dark images,etc.in cable tunnel scenes,and also improves the color distortion of the enhanced images.