V.A. Kulikov, S.A. Popov, and V.S. Kirichuk (Russia)
X-ray, image processing, CLAHE, software, liner filtering
The paper presents the application of different filtering algorithms to simplify the search for suspicious objects on the X-ray image obtained by SRC "Sibscan”. The basic idea of x-ray image filtering is an image contrast enhancement with minimal loss of information. We have proposed a new method of obtaining image luminance differences, allowing to identify 12 directions in the 35 * O (N) operations on the image. We had shown that the combination of the logarithm of the image and the image gradients can improve the visual quality of the latter. A study was conducted with existing methods of histogram equalization, in that study the method of adaptive histogram equalization had shown the best results. Combination of the adaptive equalized image and the image gradient corrects deficiencies of the equalization. Filtering algorithms were implemented in a software system, written on programming language C+ +. Image processing takes 3 seconds on a standard PC with image size of 1000 by 1024 pixels, and can be accelerated on parallel computation using multiple cores in accordance to the technology OpenMP.
Important Links:
Go Back
