Specialized software that supports existing approaches to processing images of the crystal structure of materials for analyzing transmission electron microscopy images have a lot of different digital image processing methods, but major part of it are weakly automated. In some tasks automated algorithms of image processing have been developed, e.g. in task of estimation of the width of a layer of material from a raster image. The paper considers the problem of automated processing of diffraction images obtained by transmission electron microscopy. A number of modifications, such as Watershed algorithm, binarization and Fast Fourier Transform, are proposed for existing image processing algorithms. These modifications can help automate the processing of the diffraction pattern of a material sample from an image of transmission electron microscopy. The given examples of image processing of particular cases of diffraction patterns have shown the prospects for the development of algorithm based on combination of the proposed modifications of considered algorithms. Adaptive binarization with Watershed segmentation would be useful in automated distance estimation in transmission electron microscopy images.
Methods and algorithms of computer vision associated
with the processing and analysis of raster images are
widely used in various fields of science [
Transmission electron microscopy (TEM) involves the study of thin images using a beam of electrons passing through them and interacting with them. The electrons that pass through the sample are focused on the imaging device: a fluorescent screen, a photographic plate or a camera sensor. Using TEM, it is possible to study objects even at the atomic level. At relatively low magnifications, the contrast on the TEM arises due to the absorption of electrons by the material of the test sample. At high magnifications, the complex interaction of waves forms an image that requires a more complex interpretation. Modern TEMs have operating modes that allow one to study the elemental composition of samples, crystal orientation, phase shift of electrons, etc.
The TEM method is used to assess the structure of the material, both in the volume of the sample and in its surface region. TEM is one of the most highly informative research methods used in materials science, solid state physics, biology, and other sciences.
The ability to observe the electron diffraction pattern (or, in other words, to observe cross sections of the reciprocal lattice of the sample) simultaneously with the image of the microstructure of the sample makes it possible to obtain valuable information on the symmetry of the crystal lattice and structural defects of the material under study. Comparison of microphotography with electron diffraction patterns makes it possible to correlate the microstructure elements with a particular crystalline or amorphous phase identified on the electron diffraction pattern. The TEM method can be applied to solve problems such as: • characterization of the structure of the sample in volume and on the surface; • determination of the qualitative phase composition of the sample; • determination of orientational relations between the elements of the structure of the sample.
The development of information technology and
computing devices contributes to progress in solving such
problems. However, nowadays, existing methods for
identifying and evaluating microobjects from a raster
image do not have sufficient universality that would make
it easy to automate them. In addition, these methods and
algorithms are often part of proprietary software that is
closely associated with electronic microscopy equipment
and is protected by the copyright of the manufacturers of
this equipment [
At the same time, systems for analyzing such images can be widely demanded by many scientific organizations as well as enterprises in the manufacturing sector, making it possible to carry out operational control and analysis of the composition and structure of materials and products. Cheap analogues of existing software for critical areas of enterprise activity can stimulate the development of technologies for the synthesis, analysis and production of micro- and nanomaterials. Ultimately, improving the characteristics of materials for various purposes and products based on them can both have a positive impact on the level of safety in many areas of human activity, and contribute to the development of new technologies.
In [
Thus, the interdisciplinary topic under consideration seems relevant and promising. There is a need to systematize existing methods and algorithms for the automated evaluation and processing of images and to develop software for the evaluation and analysis of images of electron microscopy based on them. In the future, this could contribute to the development of software systems for analyzing images of electron microscopes with the possibility of mass application and implementation.
Earlier in [
With the known relation between the image points and the actual physical quantities, it is possible to obtain the layer width estimated by the physical quantity.
The proposed algorithm is quite simple from the point of view of implementation, and does not require large computational resources.
The described algorithm was tested on several images shoot in direct resolution mode; crystalline planes in the composition of the crystalline material are visible in such images (figure 1).
As a result of comparing the values obtained in the framework of the developed algorithm with the values obtained by specialized software and used as reference, insignificant discrepancies of the order of 0.005 nm were obtained, which indicates the successful development of an analogue to existing algorithms. 3.
Within the framework of increasing the level of automation of TEM algorithms, determining of the crystal lattice parameters of material samples was chosen as the next task.
The phase composition and orientation of the crystals can be determined from the diffraction patterns of the samples.
Diffraction contrast is formed as a result of elastic scattering of transmitted electrons by regularly located atoms in the crystal lattice.
Contrast is used to determine the crystallinity of the film and the lattice parameter (figure 3).
The results were compared with the results of
calculations of specialized software (Gatan Microscopy
Suite v.1.8) [
Specialized software processes these images in a
semiautomatic mode, requiring the selection of starting points
for calculations. An example of such processing is
presented in Figure 4 [
As can be seen from figure 4, the software function consists of determination of brightness in a circular radius from the center of the diffraction pattern, as well as calculation of the number of points on concentric circles and the distance between these circles, which characterize the material sample. The problem with this approach is in the necessity of accurate manual selection of the center of the image and the distance to one of the concentric circles, consisting of bright points.
In addition, not all images of the diffraction pattern have the same high quality, which allows to determine important characteristics. Images with excessive brightness of the central area can be the most frequent example of low quality images (Figure 5). The situation with reflection of points is also often can be found (Figure 6). These specific cases make difficulties in developing a universal automatic algorithm that would be able to parse all particular cases with the same efficiency.
Evaluation of the effectiveness processing algorithms in the automating the determination lattice parameters of of image task of crystal
There are a number of algorithms that can be useful in automating the processing of diffraction patterns. A number of well-known image processing algorithms and the possibilities of its applying to this task were considered.
For images with a sufficiently high contrast, the determination of distances can be carried out by transferring the entire image to the frequency domain, for example, using a variation of the Fourier transform (DFT). However, in case of irregular brightness or if the image does not fit the conditions for such a conversion, it will be impossible to determine anything from it.
The Watershed method of automatic image
segmentation [
The advantages of using this algorithm are in automatic selection of all areas with large differences in the brightness of the image and in determination of their size and position. The disadvantage of this algorithm is that it is very sensitive to the choice of the brightness threshold and to its irregular distribution in the processed image. Figure 7 shows that, despite the successful selection of points remote from the center, the entire area in the center almost merges into one continuous segment.
Given the specifics of the images of the diffraction pattern of materials, it can be assumed that a modification of this algorithm that dynamically changes the sensitivity threshold from the center of the image to its borders can become quite effective for the task of evaluating these images.
Another, simpler automation option may be the calculation of a binary image according to a similar principle, with a dynamic sensitivity threshold, taking into account the decrease in brightness of images from the center to the borders (Figure 8). Thus, it can be assumed that for the task of determining the characteristics of the diffraction pattern, the most effective direction will be a complex algorithm combining binarization and segmentation to automate the search for the image center and bright points distant from it with the algorithm for calculating the brightness of concentric circles traditional for specialized software with previously known positions of the center and distance to these circles.
Further work in automation of the processing of images of diffraction patterns will be carried out in the direction of developing specific modifications of the considered algorithms.
The review of the existing approach to processing images of the crystal structure of materials using specialized software for analyzing TEM data is carried out.
The analysis of the problem of estimating the parameters of the crystal structure from the image of the diffraction pattern of a material sample is carried out; image processing methods suitable for automating this task are identified. A number of possible modifications of existing algorithms are proposed, which allow to process such images in a partially automatic mode.
The development of these modifications will make it able to solve the problem of estimating the parameters of the crystal structure without involving specialized software.
The work was supported by RFBR, Grant № 18-41700001.