Research On The Clearness Processing Of Degraded Video Image Based On Transformation Domain

Marine traffic and road transport monitoring,factory production monitoring,public safety and intelligent navigation and other outdoor computer vision system is playing an increasingly important role in people's life.However,in the fog conditions,the images captured by the outdoor vision system will have different degeneration degrees such as low contrast.The images which contain a large number of details will become blurred or even be covered.Seriously,it will lead outdoor image equipment to abnormal operation.Therefore,in order to improve the resolution of the fog-degraded video images,the degraded video images are processed by the way of image enhancement,at the same time,combining with the characteristics of local,multi-resolution,multi-directional in the transform domain.Firstly,to overcome the shortcomings of the commonly used defog enhancement algorithm,which only enhances the whole images from the frequency domain or time domain rather than the local details of the images,a new image enhancement algorithm based on wavelet transform is proposed.Based on the analysis of the characteristics of wavelet image decomposition coefficients,it is determined that the threshold function is used to modify the wavelet coefficients and then inverse transformation is calculated in the time domain for contrast stretching,and the enhanced images are obtained.Studies have shown that the inter-frame scene has strong correlation and the scattered fog has weak correlation.Based on the characteristics of video and the single-frame wavelet defogging threshold function,a video sequence wavelet defog algorithm based on inter-frame relation is proposed.The validity and superiority of the algorithm are verified by MATLAB simulation and comparison with the parameters of peak signal to noise ratio and image entropy.Then,by using the idea of video image wavelet defogging algorithm,video image defogging algorithm based on contourlet transform is proposed to solve the problem of wavelet transformation with limited directions and edge characterization of non directional.The high frequency and low frequency coefficients are effectively separated by the contourlet transformation.The low frequency part uses the Visushrinking threshold method to remove the noise,and the processed images are relatively smooth.The high frequency part uses the adaptive threshold method to classify the coefficients,and uses the enhanced function to deal with the coefficients processing,suppressing noise,enhancing the details.Finally,the images are reconstructed to the time domain for contrast stretching.The validity of the algorithm is verified by simulation experiment and data comparison.