УДК 004.932.4
Аппаратная
реализация сжатия динамического диапазона цифровых изображений на ПЛИС Xilinx
Д.
А. Гаврилов 1, А. В. Павлов 2, Д. Н. Щелкунов 1
Ключевые слова: обработка изображения, билатеральный фильтр, управляемый
фильтр, HDR-изображение, сжатие HDR-изображения, интегральное
представление изображения, оптимизация ресурсов.
Abstract. Introduction: technologies of computer
vision are widely used in many fields of science and industry. The tasks of
vision systems are the acquisition, transmission, reproduction and processing
of visual information without loss of information. Algorithms and filters used
for preprocessing are designed to suppress noise, retrieve, and store
information necessary for further processing. Purpose: the development of a digital image
processing module with hardware implementation on the Xillerx XC7Z020-2CLG400I
crystal XC7Z020-2CLG400I chip Xilinx, which allows the HDR image to be
compressed without loss of image detail and high processing speed. Results: In the presented work, the image
processing system based on computer vision technology was developed. The
analysis of filters for image processing is carried out, drawbacks and
advantages of the bilateral and guided filters are considered, on the basis of
which the filter is selected. The algorithm of the selected controlled filter
is presented. The hardware implementation of the selected filter on the
programmable logic is performed. The authors conducted an effective method for
optimizing the guided filter using the integrated image representation and
reducing its bit depth, which allowed to increase the efficiency of the filter
and to compress the HDR image without loss of quality and with high processing
speed. Practical relevance: The developed algorithm provides noise suppression and storage of
information necessary for further processing.
Keywords: image processing,
bilaterial filter, guided filter, HDR image, HDR image compression, integrated
image representation, resource optimization.
Для цитирования:
DOI
10.30898/1684-1719.2018.10.6