Research On Image Enhancement And Restoration Technology In Foggy And Marine Environment

In recent years,haze and smog around the world has become more and more serious.The haze and smog in China appear more and more frequently.The haze causes a drop in air quality.In addition to posing a great threat to physical and mental health,smog also has a huge impact on outdoor computer vision acquisition systems,which seriously hinders the development of public transportation,outdoor surveillance and aerial photography.The presence of haze smog causes the captured image to degrade.Such images are prone to low contrast and inconspicuous details,and the degraded image has blurred visual effect.There are many ocean and rivers and lakes in China.China has four sea areas,and the Bohai Sea,the Yellow Sea,the East China Sea and South China Sea are adjacent to each other on the edge of the Chinese mainland.In China,there are more than 1500 rivers with drainage area of more than 1000 square kilometers,and there are more than 24800 lakes,of which more than 2800 are natural lakes with an area of more than 1 square kilometer.These oceans,rivers and lakes are not only an important part of China's geographical environment,but also rich in natural resources.Similarly,when we are working underwater,the more complex optical environment under water directly causes the captured image to have a sever color cast effect,accompanied by low contrast and extremely blurred visual effects.It has caused great trouble to underwater research work.Compared the outdoor acquisition systems,more complex underwater environments make image restoration more difficult.In order to improve the reliability and stability of the imaging system,this paper focuses on the visual degradation of images and video captured by outdoor hazy and underwater computer vision systems.By studying and analyzing the atmospheric model and underwater attenuation model of image imaging,we conduct research on outdoor and underwater videos.The main research contents of this paper include the following parts:In order to overcome the dehazing distortion in images contain large sky bright area caused by the dark channel prior algorithm,this paper proposed an improved single image dehazing algorithm based on adaptive region partitioning and fused transmittance.First,a more accurate atmospheric light is obtained by quad-tree decomposition based on sub-block average gray value and standard deviation.Then,the image is segmented into two parts: bright area and non-bright area by adaptive threshold segmentation.The grayscale distribution of the image is used to calculate a new weighting factor.And this weighting factor is used to blend the transmittance to make the edges smoother.Finally,the haze-free image is restored by the atmospheric scattering model.This paper compares the proposed algorithm with several commonly used dehazing methods.Experiment results show that this algorithm can restore natural and clear images in most cases,and effectively improve the image quality in bright areas.Due to the presence of haze or fog,the visual effects of images collected by computer vision systems are degraded.Image dehazing based on the dark channel prior method can achieve good results.But distortion still appears on some restored images,especially images that contain large sky areas.This paper proposes an improved dehazing method for images with large sky area.First,we detect the presence of the sky area based on the distribution characteristics of the image intensity.Then,region growing is used to segment the sky area and a more accurate atmospheric light is obtained from the segmented area.And then we propose a linear correction factor to adjust the transmission map.Finally,the improved Gamma correction algorithm is used to increase the brightness of the restored image.Experimental results show that the improved algorithm has good dehazing effect especially on images and videos which contain large sky areas.Due to the special optical characteristics of light in water,the information of the red channel is seriously attenuated in collected image.This causes other colors to dominate the image.This paper proposes an underwater image enhancement algorithm based on red channel weighted compensation and gamma correction model.Firstly,by analyzing the attenuation characteristics of RGB three channels,the intensity and the edge information of red channel are compensated by weighting the attenuation coefficient ratio between different channels to correct the chromaticity.Then the gamma correction model is employed to stretch the intensity range to enhance the contrast which makes the image look clearer.The experimental results show that the proposed algorithm can correct the color cast effect and improve the contrast by nearly 2 times for the underwater images with too much red component attenuation.