The Research On The Dehazing Algorithm Based On The Ellipsoid Model In RGB Color Space

The light transmitting in the atmosphere is usually absorbed and scattered by the suspended particles in the air,thus resulting in the blurred images taken in the atmospheric scattering environment and the difficulty to identify the objects.The image dehazing technology for the images taken in the atmospheric scattering environment has been widely used in the fields of target detection and identification,road traffic monitoring and so on.Aiming at solving the problem of the series degradation of the images taken in the atmospheric scattering environment,this thesis proposes a dehazing algorithm based on the ellipsoid model in the RGB color space.By establishing an ellipsoid model in RGB color space to describe the aggregation state of the pixels' intensities from a hazy image,the transmission can be estimated,and then the target image for this hazy image can be restored.The imaging system is established to carry out the data acquisition experiments,and then the experimental data is used to verify the effectiveness and correctness of the algorithm proposed by this thesis.This main research contents are followed as:(1)The transmission estimated by some traditional methods such as filtering could result in the distortion in recovery results for image dehazing.This thesis analyzes the fundamental principle of the aggregation state of the pixels' intensities from a hazy image in RGB color space,and grasps the statistical characteristics of the pixels' intensities in general.The ellipsoid model in RGB color space is used to describe the aggregation state of the pixels' intensities in a rectangular window from a hazy image,and then the ellipsoid model is solved to estimate the transmission for each pixel in hazy image.The algorithm based on the ellipsoid model in RGB color space has more advantages in detecting targets and improving details.(2)The imaging experiments of the targets in the atmospheric scattering environment and dehazing experiments are carried out.Single hazy images and polarization hazy images are obtained by ordinary industrial camera and polarization camera respectively.The algorithm introduced in this thesis is used to process the single hazy images to obtain the clear target images through haze,and also the algorithm can be combined with the polarization imaging technology to dehaze the polarization hazy images.The experiments for single image dehazing and polarization image dehazing are carried out,and then the corresponding dehazing results are obtained.(3)As for the dehazing results of the single hazy images and polarization hazy images,the restoration results of the algorithm proposed by this thesis are compared with those of other algorithms through the combination of subjective evaluation and objective evaluation,so as to analyze both advantages and disadvantages of the restoration results of the proposed algorithm in terms of contrast and clarity.