Research Of Restoration Algorithm Of Images Collected In Adverse Weather

Currently vision systems are more and more widely used.Outdoor vision systems play a key role in outdoor surveillance,and traffic management,and satellite remote sensing,and even military reconnaissance and other fields.While in adverse weather conditions of haze,fog,rain,snow,images got by systems have seriously degraded.Images are vague,and contrast of images is low,and color of images is distorted.The value in use of images is greatly affected,thereby it reduces the robustness and applicability of visual systems.So it is not conducive to feature extraction and target recognition for images.Thereby increasing the definition of images got by outdoor vision systems in adverse weather conditions becomes inevitable.In this paper,the problem of restoration of degraded images in adverse weather conditions is deeply researched through studying domestic and abroad related information and literature.The content of research and innovation are the following aspects:First of all,for cloudy degraded images,according to degradation model of cloudy image,the paper uses MSR(multi-scale Retinex)algorithm for restoration.But MSR algorithm can't effectively restore the details and color of images in processing cloudy images,so the paper proposes a new MSR improvement algorithm to process cloudy images.The three scales of traditional MSR is changed into four scales,that is,a Gaussian function with middle scale keeping the details and color of images is added.And information fusion strategy based on wavelet transform domain is substituted for linear weighted strategy of multi-scale reflection images for MSR algorithm.The basic idea of fusion is that: Firstly,the images to be fused are for two layers of wavelet decomposition.Then high-frequency component are taken the local standard deviation method to stress the details in image.And local energy method is used for low frequency component to adjust the background and color,which realizes the effect of fidelity.Lastly,subjective observation and objective evaluation indicates that compared with traditional MSR algorithm,the algorithm of the paper has better effect on details restoration and color fidelity for restoring cloudy images.Secondly,for foggy degraded images,the paper proposes an improved defogging algorithm of single image which can defog the foggy images rapidly,based on dark channel priority.The algorithm in the paper applies the method combining adaptive median filter and bilateral filter to figure out clear dark channel on the edge.And the algorithm is based on the physical model of foggy images to estimate transmission.Compared with the traditional algorithm,the estimated transmission is detailed and clear,and has no need to be optimized,which not only overcomes the disadvantages of traditional algorithm using plenty of time to optimize transmission,but also reduces the complexity of the algorithm.The experimental results indicate that the algorithm realizes rapid and high-quality defogging on single image.Next,for rainy or snowy degraded images,the paper proposes a rain or snow removal algorithm of multiple images based on rain or snow noise pollution.Traditional rain or snow removal algorithm is restricted with the intensity,so the effect is not ideal.According to the characteristic of vision system acquiring multiple different degraded images in a short time,the paper processes multiple images for realizing restoration.And snow and rain has the dynamic characteristic that the direction,intensity and shape of rain and snow are unfixed,which makes it difficult to establish unified physical model in the spatial domain.But analyzing them in the frequency domain doesn't affected by the dynamic characteristic.From the perspective of frequency domain,the paper uses the method of wavelet multi-level decomposition and wavelet fusion to determine the number of layers of rain or snow noise,formulates a fusion rule based on rain or snow noise pollution,and makes wavelet fusion on specific layer of multiple continuous degraded images for achieving the objective of rain or snow removal.Simulation results indicated that the algorithm in the paper not only has ideal restoration results,but also is not restricted by noise intensity.Finally,for the less of evaluation methods of image qualities of restored images in adverse weather and limitations of existing evaluation methods,in this paper,selecting the appropriate method of image quality objective evaluation to analyze the recovery results of degraded images of each adverse weather based on specific characters of each adverse weather and simulations.In this paper mean gradients,edge intensities and structural similarities are used to evaluate the restoration results of cloudy images.Relative contrasts and edge similarities are used to evaluate the restoration results of foggy images.Means and signal-to-noise ratios are used to evaluate the restoration results of rainy or snowy images.The results of image quality objective evaluation in this paper are consistent with subjective feelings of the human eyes,and compared with the traditional evaluation methods,they have better adaptability and reliability.So it can prove that the proposed restoration algorithms in this paper are superior to traditional algorithms.