Study On Local Image Reconstructing Algorithm Based On The Static Magnetic Field

Traditional research on information reconstruction and object remove,focuses more on the overall visual effect with low efficiency.Nowadays,image and video reconstruction has been developed to apply in feature extraction,recognition and other computer processing,which leads to higher reconstruction requirements on the accuracy and efficiency.The stable field model for image local texture,attempt a new view of research to meet these requirements on image reconstruction.Based on the relationship between each pixel and energy in the stable field,a new reconstruction model for image local region based on static magnetic field is proposed.Set Laplace's equation on static magnetic field and employ Green function to get the energy model for image local texture,come to get the reconstruction model for image local region based on static magnetic field: In this field,each pixel is regarded as a current source with unify resistance and the corresponding voltage value of current source,and the voltage gradient in the known region works as field source for information transfer.Pointing at each current source in the defect region,considered all current sources in the known region as the electromagnetic induction and the energy transfer function to express and calculate the supplied energy that all current sources transfer from the known region to the defect region,to hope for the accurate transfer information on image local texture,in the final,the pixel influence function with normalized superposition is proposed to get the reconstruction model.The innovation of the reconstruction model is: the structural feature(such as edge feature and detail texture)works as a driving transfer from energy information,the energy works as a pre-estimation pixel value,the pixel influence function is normalized superposition by energy transfer function,and then the reconstruction of each pixel in the defect region is achieved.In the above study,the gradient calculating schedule should be designed in a proper way to get the edge feature in defect region as exactly as possible.After single-direction calculation is selected,and the demerit about nearly principle is analyzed theoretically and verified experimentally,the single-direction calculating schedule based on nearby-axis principle,which sets the first priority as the pixel on the other axis rather than nearer target,and sets the second priority as the pixel nearer the other axis.Finally,synthesized above gradient thought and the reconstruction algorithm in last chapter,a strong robust gradient calculating schedule of image local reconstruction is proposed.In general,experimental results show that research achievements of this paper including static magnetic field based image local reconstruction model and the gradient calculating schedule,on the premise of visual effect peer previous methods,can be used to increase the efficiency and therobustness of the reconstruction process.Moreover,it is possible to be applied in the real-time processing of video frame.Specially to be noted,the paper tries to do some work in the accurate reconstruction,and get good performance in reconstructing edge feature and detail texture accuratly.