Software for inclusions recognition and analisys

Abstract

For the past years much attention has been paid to inclusion elimination in CdZnTe crystals used for radiation detectors. One of the important parts of this is accurate determinations of inclusion size and concentration. Usually researchers create their own software to calculate these parameters from IR transmission images. Therefore it is important to determine the validity of these programs. For this purpose we developed a software environment that can be used to model the real inclusion distribution in the sample volume and then analyze it. It includes two parts: (1) virtual crystal creation with a subsystem of different structural defect parameters such as types, sizes and distribution of defects (dot, linear and plane defects are supported; distributions: uniform, random, Poisson and Gaussian), and (2) IR images recognition to visualize inclusions and other defects created during scanning of actual crystals or generated by the first program. Also, it is possible to view all created defects in the sample in three ways: in the three-dimensional image, in the set of test photos, and in the crystals description file where coordinates and the sizes of all created defects are stored. In the last case we can use the generated set of images for testing of the recognition quality of defects, a threshold under noise characteristics of the pictures, brightness and contrast. In addition the virtual crystal creation program is able to generate the defects with shadow from IR lantern. The recognition program is able to cut off the shadows and helps us to choose the optimal scanning step to minimize the shadows influence. Using the developed software we tested the correctness of our recognition algorithm. Furthermore, it allows for testing the recognition programs developed by other authors, and helps to choose the optimal IR s