Showing papers on "Dark-frame subtraction published in 1995"

Video signal noise reduction system using time-varying filter coefficients.

Abstract: A video signal noise reduction system is disclosed in which corresponding image pixels from a current frame and image pixels from a noise reduced previous frame are averaged to produce respective noise reduced pixel values. In the disclosed exemplary embodiment of the invention, a processed frame memory stores a noise reduced image and a motion estimator identifies a target block of pixels in this processed frame memory to be used in calculating a noise reduced video signal. A motion compensated noise reduction processor combines the block of the noise reduced image pixels stored in the processed frame memory with a current block of pixels generated from the video signal to produce a motion block of noise reduced pixels. The relative weighting applied to the stored pixel and the current pixel depends on the number of pixels that were used to generate the stored pixel. A breakdown detector discards pixels in the motion block if a measure of noise reduction achieved for the one component divided by a measure of noise in the motion block is greater than a threshold value. Sparkle reduction is used to replace any original component of the motion block with its corresponding noise reduced component if all of the components adjacent to the original component have been replaced by their corresponding noise reduced components.

Effects of image noise on submicroscan interpolation

Abstract: The submicroscan interpolation image-processing technique is analyzed to determine the effects of additive noise on the quality of the output imagery. Both temporal and fixed-pattern spatial noise are evaluated assuming a white Gaussian noise model. Closed-form solutions for the power spectral density of the output noise are derived for both one- and two-dimensional submicroscanning. It is found that temporal noise degrades the output imagery and determines the minimum usable submicroscan image shift. For the case of fixed-pattern noise, submicroscan interpolation causes a spectral redistribution of the noise power spectrum that tends to improve image quality.

Impulse noise correction in TV transmission

Abstract: We describe a real-time digital image processing approach for correcting the impulse noise impairment in television transmission. The described algorithm identifies the pixels impaired with impulse noise and subsequently corrects them. A real-time processing board is constructed to demonstrate the algorithm. >

Noise limitations in dynamic infrared scene projection

Abstract: The spatial and temporal noise factors that set the ultimate performance limits on dynamic infrared scene projectors are assessed. Fixed pattern noise and the needs for non-uniformity correction procedures are both discussed, followed by a detailed analysis of the temporal and residual spatial noise contributions within the projector/imaging unit-under-test (UUT) composite system. It is found that the system is more tolerant to projector spatial noise by a factor of about four compared to UUT spatial noise. The temporal and spatial noise contributions are combined appropriately in order to provide guidance as to the levels of array uniformity that will need to be achieved if the expected demands for fine temperature resolution are to be met.

Method of suppressing burst noise interference in images using smearing-desmearing transform

Abstract: To realize high-quality image transmission and storage, it is important to reduce the effects of the impulsive noise and the image data erasure produced in the transmission and storage. This paper proposes a new method for image noise suppression using the smearing-desmearing transform. The smearing-desmearing transform retains the power spectrum of the image signal and can be applied to the highly efficient coding. the basic principle for the image noise suppression using the smearing-desmearing transform is shown and the construction of the suitable smearing-desmearing transform is presented. Then, it is shown that the noise suppression performance of the image noise suppression method based on the smearing-desmearing transform depends on the variance of the noise, even if there exist the same number of noises. Based on such a property, a method to suppress the burst noise in the image based on the interleaving-deinter-leaving and smearing-desmearing transform is proposed where the smearing-desmearing transform is applied after spreading the burst noise to fit to the image noise suppression process. Finally, by a computer experiment, it is shown that the proposed method has an excellent effect in image noise suppression.