Zagrebnyuk V. Methods of image processing and encoding for application television systems

Object of research is the process of encoding, compression and processing of digital television images in the application television systems. Subject of research is system of encoding and processing of television images in the application television systems. Methods of research are mathematical methods of functional and statistical analysis, linear algebra, information theory, computer simulation and modern programming technologies are used to develop a linear transformation of digital television signals and to develop the methods of encoding and processing of television images. Theoretical and practical results are as follows. The linear transformation with basis and properties determined by the statistical distribution of colors on television picture was obtained. This transformation allows to encode color components using adaptive uniform code with the length that depends on the palette size. Due to it also the method for converting color images to grayscale has been developed. The method of compressing linear mapping with a variable compression ratio was developed. This method is invariant with respect to the bit representations of television signals and gives the possibility to encode the color components using adaptive uniform code with less than one byte length. For the images of the text documents the methods of quality assessment and adaptive brightness and contrast increasing were developed. The method for forming and analyzing the Hough accumulator array was developed. The codec of ХХ3 format for television image encoding with adaptive uniform code is implemented using compressing linear mapping сВХ2Х3. This codec also uses chroma and achromatic components sampling and recovery using bilinear and bicubic interpolation. The software for converting color images to grayscale with maximum preservation of color contrasts, for automatic binarization of images of text documents, for detecting the line segments of arbitrary thickness was developed. The application for detecting and recognition of handwritten or printed numbers on television image was developed.