KR20200106763A - Method for analyzing electron microscopy image - Google Patents

Abstract

Provided is an electron microscope image analysis method. The electron microscope image analysis method comprises the steps of: obtaining an electron microscope image; processing the electron microscope image to form a binarized image; and analyzing the binarized image.

Description

Electron microscope image analysis method {METHOD FOR ANALYZING ELECTRON MICROSCOPY IMAGE}

The present invention relates to an electron microscope image analysis method.

The growth of nanoparticles involves several unique mechanisms, including nucleation, growth, adhesion, and maturation. Due to this complexity, the growth of nanoparticles must be experimentally investigated in terms of both individual growth trajectories and kinetics in ensemble mean for a comprehensive understanding of growth mechanisms. Liquid transmission electron microscopy can directly observe the growth of nanoparticles, but is limited to a limited number of growth trajectories.

In order to solve the above problems, the present invention provides an electron microscope image analysis method.

An electron microscope image analysis method according to embodiments of the present invention includes obtaining an electron microscope image, processing the electron microscope image to form a binarized image, and analyzing the binarized image.

The electron microscope image may be binarized by adaptive binarization using an adaptive binarization algorithm. Background noise of the electron microscope image may be corrected by the binarization.

Processing the electron microscope image may include removing noise from the electron microscope image and increasing an edge contrast of the electron microscope image. Removing noise from the electron microscope image may be performed using a biner filter and a Gaussian filter, and increasing the edge contrast of the electron microscope image may be performed using a Laprasian filter. Processing the electron microscope image may further include adjusting a gamma value of the electron microscope image using gamma correction.

The binarized image can be analyzed using a computer.

The electron microscope image may include a nanoparticle image, and analyzing the binarization image may include analyzing the size and position of the nanoparticle and analyzing the growth rate of the nanoparticle. The size and position of the nanoparticles may be analyzed using an ensemble analysis method.

The electron microscope image may include a liquid transmission electron microscope image.

According to embodiments of the present invention, an electron microscope image can be accurately and effectively analyzed using a binarized image. Since the binarized image is composed of only 0s and 1s, it can be easily calculated by a computer, so that a lot of data can be systematically and accurately analyzed in a short time. As a result, the effectiveness of electron microscopy measurements including transmission electron microscopy and the like can be increased.

1 schematically shows a process of forming a binarized image according to an electron microscope image analysis method according to an embodiment of the present invention.
2 and 3 show ensemble analysis of the mechanism of formation of individual Pt nanoparticles.
4 to 11 are diagrams for explaining the correlation between adhesion growth and monomer growth in a specific nanoparticle size.
12 to 15 are diagrams for explaining a narrow size distribution through monomer growth in an initial stage of nanoparticle growth.
16 to 19 are views for explaining Ostwald ripening in a later stage of nanoparticle growth.

Hereinafter, the present invention will be described in detail through examples. Objects, features, and advantages of the present invention will be easily understood through the following embodiments. The present invention is not limited to the embodiments described herein and may be embodied in other forms. The embodiments introduced herein are provided so that the disclosed content may be thorough and complete, and the spirit of the present invention may be sufficiently transmitted to those of ordinary skill in the art to which the present invention pertains. Therefore, the present invention should not be limited by the following examples.

An electron microscope image analysis method according to embodiments of the present invention includes acquiring an electron microscope image, processing the electron microscope image to form a binarized image, and analyzing the binarized image.

1. Acquire an electron microscope image.

The electron microscope may be, for example, a liquid phase transmission electron microscopy (TEM), and the electron microscope image may be a liquid transmission electron microscope image. In addition, the electron microscope image analysis method is not limited to the liquid transmission electron microscope and can be applied to various electron microscopes.

2. The electron microscope image is processed to form a binarized image.

Processing the electron microscope image includes removing noise from the electron microscope image, adjusting the gamma value of the electron microscope image using gamma correction, and/or increasing the edge contrast of the electron microscope image. Can include. Removing noise from the electron microscope image may be performed using a biner filter and a Gaussian filter, and increasing the edge contrast of the electron microscope image may be performed using a Laprasian filter.

The noise of the electron microscope image may be removed by using a biner filter, and a dark portion of the electron microscope image may be further emphasized by adjusting a gamma value. Characteristic Gaussian noise can be removed from the transmission electron microscope by using a Gaussian filter to minimize a portion without image information, that is, a signal of a background. Using a local laplacian filter, a portion of the electron microscope image with a weak boundary signal may be weaker and a portion with a high contrast may be more emphasized.

The electron microscope image may be binarized by adaptive binarization using an adaptive binarization algorithm. Background noise of the electron microscope image may be corrected by the binarization.

1 schematically shows a process of forming a binarized image according to an electron microscope image analysis method according to an embodiment of the present invention.

Referring to FIG. 1, an electron microscope image is subjected to binar filtering, gamma correction, Gaussian filtering, local Laplacian filtering, and binarization to form a binarized image.

3. Analyze the binarized image.

Since the binarized image consists of only 0s and 1s, it is suitable for computer analysis, and can be easily calculated on a computer, so that a lot of data can be systematically and accurately analyzed in a short time.

The electron microscope image may include a nanoparticle image, and analyzing the binarization image may include analyzing the size and position of the nanoparticle and analyzing the growth rate of the nanoparticle. The size and position of the nanoparticles may be analyzed using an ensemble analysis method.

The electron microscope image may include a liquid transmission electron microscope image.

Hereinafter, an electron microscope image analysis method will be described by taking nanoparticle growth mechanism analysis as an example.

Understanding the mechanism of nanoparticle growth is essential in the field of colloidal nanoscience. The formation of nanoparticles is generally achieved by nucleation and subsequent growth through the bonding of monomers, which are the smallest units of nanoparticles. Thermodynamic and kinetic factors, including the surface energy of the growing nanoparticles, the diffusion constant and monomer concentration of the monomer in the mixture of ligand and solvent, influence the growth trajectory.

Under diffusion-controlled growth conditions with a high monomer concentration, the growth rate of small nanoparticles is faster than that of large nanoparticles, resulting in a narrow size distribution. When the monomer concentration is low, Ostwald ripening becomes dominant because the frequency of monomer separation from small particles increases, as is typically observed in later stages of nanoparticle synthesis. In order to elucidate the origin of the change in size distribution during nanoparticle synthesis, the growth rate and path of individual nanoparticles must be directly observed.

Liquid TEM is suitable for real-time analysis of individual nanoparticle behavior and helps to uncover growth mechanisms. For example, liquid TEM can be used to confirm the presence of frequent adhesions between small particles driven by a large reaction cross-sectional area and their contribution to a narrow size distribution.

Statistical analysis of the overall growth trajectory of a large number of nanoparticles via liquid-in situ TEM can provide information on size-dependent motion and thermodynamic factors associated with adhesion development. In addition, the comprehensive investigation of the interactions between different types of growth pathways enables the synthesis of monodisperse nanoparticles with precisely controlled morphology.

An electron microscope image analysis method according to embodiments of the present invention provides an ensemble analysis of a large number of growth trajectories of Pt nanoparticles obtained from an in situ liquid TEM using an image binarization method and computational analysis. The inherent properties and contributions of monomer growth and adhesion occurrence resulting in a narrow particle size distribution can be understood by separating the growth trajectory types in the ensemble. In addition, the critical size of the nanoparticles is important in determining the adhesion frequency and growth rate, and ensemble analysis shows that the growth rate of nanoparticles along the monomer pathway is determined by the monodispersity due to Ostwald ripening when the monomer is depleted in the later stages ) Is inversely proportional to the size of the nanoparticles before it begins to decrease.

2 and 3 show ensemble analysis of the mechanism of formation of individual Pt nanoparticles.

2 and 3, a silicon nitride liquid cell was prepared to directly observe the growth of nanoparticles in solution and applied in situ TEM. A series of TEM images with viewing angles showing the numerous growth trajectories of Pt nanoparticles in the presence of oleylamine ligands are obtained. Overall, burst nucleation of Pt nanoparticles is observed in the early stages, and they continue to grow to large sizes.

To efficiently and accurately track hundreds of individual growth trajectories, TEM images are processed with noise removal, edge contrast optimization, inconspicuous pixel removal and adaptive binarisation. The size and position of the nanoparticles tracked in the binarized image agree well with the original TEM image.

The trajectories of more than 300 individual nanoparticles from the binarized in-situ TEM images were successfully tracked over 200 s of growth and used for statistical analysis of the parameters describing the growth mechanism.

Full growth trajectory

The average size, size distribution and number of particles of each frame are extracted from the calculated analysis. The radius of a particle is determined by counting the number of pixels in each particle. The average radius continues to increase as the reaction proceeds, and the growth rate decreases at a later stage as the monomer is consumed. The size distribution narrows to 120 seconds and then expands so that the total growth is divided into two stages: size focusing and size defocusing.

Due to the dynamic nucleation and dissolution of small particles in the initial stage, the number of particles fluctuating during the first 30 seconds reaches a maximum number of 325 nanoparticles in 25 seconds. While the number of particles decreases to 30 seconds after this maximum, dissolution of some of the small particles is continuously observed with adhesion. Another dramatic decrease in the number of nanoparticles from 65 seconds to 95 seconds is due to the occurrence of most multiple adhesions that occur with monomer growth. The number of particles becomes almost constant after 95 seconds, which means that adhesion is no longer dominant.

Classification of growth trajectories

Since adhesion growth and monomer growth can affect the size distribution of nanoparticles, independent analyzes must be performed to differentiate the two pathways and to comprehensively understand the growth mechanism. By tracking individual nanoparticles, the ensemble growth trajectory can be classified into two categories: growth with adhesions and growth without adhesions.

Excluding nanoparticles that exist for a short period of time in the early stages of growth, 186 nanoparticle growth trajectories are classified. The 113 nanoparticles grown only through monomer bonding without the occurrence of adhesion showed a continuous increase in size. On the other hand, the size of 73 nanoparticles experiencing adhesion increases suddenly at the time of adhesion and eventually becomes 34 nanoparticles.

All classified growth trajectories are analyzed statistically. The average size and size distribution of groups of nanoparticles grown through monomer bonding show similar trends in overall growth. Specifically, the group size distribution follows the focusing and defocusing scheme, while the average size continues to increase during the whole process. The average size of nanoparticles adhering from numerous orbits shows several distinct time intervals with a rapid increase in size. This shows that most of the adhesions tend to occur simultaneously in a short time. The size distribution of nanoparticles for monomer growth is narrower than that of nanoparticles for adhesion growth. This indicates that the synthetic procedure can be designed such that most individual growth trajectories follow the monomer pathway to obtain improved monodispersion of the nanoparticles.

Adhesion occurs

The adhesion between nanoparticles is one of the important growth pathways. However, since conventional liquid-phase TEM studies rely on a hand-countable number of trajectories, the critical factors determining adhesion and interference during the entire growth process are not well understood. The analysis method according to the embodiments of the present invention may examine a group of trajectories including the occurrence of adhesions for quantitative ensemble analysis.

The number of nanoparticles decreases significantly in the short period between 74 and 97 seconds, which can be defined as the coalescing period. The presence of this period indicates that the occurrence of adhesions between nanoparticles is not random. This means that the adhesion phenomenon can be influenced by certain thermodynamic factors.

There appears to be an important relationship between the size of the growing nanoparticles and the frequency of adhesion occurrence. The size of the nanoparticles before adhesion is similar regardless of the time at which adhesion occurs. 4 and 5 show TEM images of three representative nanoparticles and their corresponding growth trajectories. Nanoparticles growing before adhesion delay growth at a radius of 1.5 nm and maintain this size for several seconds.

After the delay time, nanoparticles with a radius of 1.5 nm are adhered to form larger nanoparticles. This critical size is consistently observed in the three coalesced nanoparticles at different times (Fig. 5). When the number of occurrences of all adhesions are calculated after 47 seconds including the adhesion period, the radius histogram of the size before adhesion shows a narrow distribution with a center of 1.5 nm (FIG. 6). The histogram shows an FWHM of 0.36 nm, which is lower than that measured for all nanoparticles at the same period, i.e. 0.51 nm.

While the average size continuously increases as the reaction proceeds, the radii of the nanoparticles adhering at different times are similar (FIG. 7). In the adhesion period when the average size of all nanoparticles increases to 1.5 nm or more, the fraction of nanoparticles with critical size is high. Therefore, adhesion phenomenon is dominant. This reduces the number of nanoparticles in the adhesion period.

Relationship between adhesion growth and monomer growth

The presence of a critical size in the coalescence orbit assumes that this critical size is likely to play an important role in the kinetics of monomer growth. In individual growth trajectories of Pt nanoparticles, a slow growth rate close to the critical size is consistently observed, regardless of the path type. Statistical analysis of this behavior is performed. First, after smoothing, the radius-time plot obtained for each particle is differentiated to determine the growth rate of individual nanoparticles.

The growth rates of all particles along the monomer path in one time frame are aligned with the particle radius and averaged over a similar radius range. This step is repeated every time frame. The growth rate of nanoparticles having a radius of about 1.5 nm is lower than that of nanoparticles of other sizes (FIG. 8). This delay at a certain size results in a distorted size distribution in each frame (Fig. 9).

For example, at 53 seconds before the start of the adhesion period, slow monomer growth at this size accumulates particles with a radius of 1.5 nm, resulting in a negatively distorted size distribution with a Gaussian distribution tailored to a lower value (Fig. 9 and 10). As the reaction proceeds and the average size increases, since the nanoparticles grow larger than the critical size, the growth rate of the nanoparticles is restored, and the distortion is reversed (FIGS. 9 and 11 ). The slow growth and distortion changes of the 1.5 nm radius nanoparticles indicate that these nanoparticles are more stable than other sized nanoparticles.

Nanoparticles with a critical radius of 1.5 nm exhibit a high frequency of adhesion and a low growth rate in monomer growth, indicating a correlation between the two pathways. As the monomer growth is delayed near the critical size, the nanoparticles grow through an alternative route, ie adhesion, to reach their final size, as determined by the thermodynamic factor in the solution including the monomer and ligand concentration. This behavior is also observed in the discrete growth profile of nanoparticles incorporating the magic size. The magic-sized nanoparticles have a unique thermodynamically stable structure, resulting in low growth rate and high density. The magic-sized nanoparticles are easy to form polycrystalline nanoparticles due to adhesion. The similarity between the formation of the nanoparticles with the critical size and the nanoparticles of the magic size means that the Pt nanoparticles with the critical size temporarily form a thermodynamically stable ligand complex before adhesion.

size Focusing Monomer growth of nanoparticles during

Growing Pt nanoparticles gradually achieve monodispersion as the particle size increases up to 120 seconds. A time series TEM image of a region dominated by monomer growth, a corresponding binarization image, and a size bar graph are shown in FIG. 12. During this period, the number of particles remains the same.

The size distribution varies from 21.1% to 15.4% in the first region, ie from 55 to 75 seconds, and from 13.8% to 12.1% in the second region, from 95 to 115 seconds. The size change of five representative nanoparticles growing without adhesion is shown in FIG. 13. The size of the nanoparticles increases from 1.4 nm in 80 seconds to 2.1 nm in 120 seconds (green dot in FIG. 13), while the size of other nanoparticles increases from 1.8 nm to 2.3 nm in the same period (blue dot in FIG. 13). This shows that smaller nanoparticles grow faster than larger nanoparticles.

To get more statistical information about this growth pattern, we measure the growth rate of 110 individual nanoparticles after monomer growth. As shown in FIGS. 8, 14, and 15, the growth rate and radius have a negative correlation except for the critical size, that is, the delayed growth rate near 1.5 nm. This trend is consistent with the mechanisms for controlling the size of nanoparticles established on the basis of classical crystallization theory. The growth rate (dr/dt) of a nanoparticle depends on its radius (r) as expressed in Equation 1.

[Equation 1]

In Equation 1, [M] is the monomer concentration in the bulk of the solution, C ⁰ _flat is the solubility of the monomer in equilibrium at a flat interface, γ is the surface free energy per unit area, and Vm is the molar volume of the monomer in the crystal. , R is the general gas constant, T is the temperature, D is the diffusion constant, k is the reaction rate constant, and α is the transfer constant, usually set to 0.5.

In Equation 1, the ratio of the diffusion constant to the reaction rate coefficient is an important factor in determining the nanoparticle growth rate. In the case of diffusion-controlled growth (D >> k) with a large amount of monomer ([M] >> C ⁰ _flat ), the growth rate and radius have a negative relationship, and the smaller nanoparticles have a higher growth rate than the larger one. Represents. As a result, the nanoparticles exhibit monodisperse over time. In addition, by applying the Lifshitz-Slyozov-Wagner model to the size change as a function of time, it can be seen that the monomer growth follows a diffusion control path that includes a rate delay around the critical size.

In the early stages of observation, where the size distribution of the growing Pt nanoparticles gradually narrows, the monomer concentration is high and the growth conditions can be regarded as diffusion control. Pt nanoparticles are capped with long chain oleylamine ligands that slow the diffusion of the monomer to the surface of the growing particles.

Ostwald Of nanoparticles by aging Defocusing

After the late stage of nanoparticle growth, ie, about 120 seconds, the growth of Pt nanoparticles enters the defocusing stage, and the size distribution increases (Fig. 16). This can be explained by Ostwald ripening based on Equation 1. In Ostwald ripening, the monomer supply is limited, so that the dissolution of small nanoparticles with large surface energy is remarkable, and the dissolved species can be consumed for the growth of large nanoparticles with low surface energy. As a result, larger nanoparticles exhibit faster growth rates than smaller nanoparticles, and the size distribution exhibits polydispersity at lower monomer concentrations. Time-series TEM images, corresponding binarized images, and their histograms show that nanoparticles become polydisperse after about 120 seconds. The standard deviation of the nanoparticle size increases from 10.7% in 122 seconds to 13.4% in 180 seconds (Fig. 16).

The defocusing step is systematically investigated by tracking individual nanoparticles. 17 shows five representative nanoparticles. The size of the nanoparticles with a radius of 2.4 nm increases to 2.7 nm (yellow in Fig. 17), while the size of the nanoparticles with a radius of 2.2 nm increases to only 2.3 nm (green in Fig. 17). This is the overall trend of measuring growth rate.

Growth rate according to size shows a positive correlation, which is a characteristic of defocusing in classical theory (Fig. 18). Also, as the reaction proceeds to the final stage, the overall growth rate decreases. During this period, the occurrence of adhesion is remarkably suppressed as indicated by the constant number of nanoparticles. Even when the nanoparticles are within a very short distance of about 1 to 2 nm, they do not adhere for a long time as shown in FIG. 19. This is because the larger nanoparticles inherently have a low surface energy and a high activation energy, which separates a significant number of surface ligands for adhesion. These results show that Ostwald ripening becomes dominant when the growing nanoparticles enter the defocusing stage and the monomers are consumed in the reaction.

The growth mechanism of Pt nanoparticles can be investigated by computationally analyzing the ensemble-in-situ orbitals obtained from liquid TEM. The growth trajectories of the enormous number of individual nanoparticles are successfully classified into cohesive growth and monomer growth pathways based on their intrinsic basic properties. Critical size plays an important role in both adhesion and monomer pathways, and nanoparticles with critical size adhere frequently and show a low growth rate. In addition, nanoparticles along the monomeric pathway exhibit a typical negative correlation between predicted growth rate and size under diffusion controlled growth conditions. Different mechanistic features in each category of branching growth pathways contribute broadly to the overall development of size monovariance. In addition, Ostwald ripening becomes dominant as a pathway to widen the size distribution when monomers are consumed in the later stages. Thus, understanding the mechanisms of nanoparticle growth requires experimental confirmation at different levels of complexity, individual and ensemble of nanoparticles. Determining and controlling the kinetic and thermodynamic factors obtained from this approach can improve synthesis control, thereby obtaining nanoparticles with high monodispersity.

[An example of an electron microscope image analysis method]

Liquid TEM Measurement (Liquid cell TEM measurement)

For in situ TEM experiments, prepare a liquid cell compatible with the TEM holder. The liquid cell consists of two 100 μm thick cell bodies, two 25 nm thick Si ₃ N ₄ windows and 100 nm thick spacers. A precursor solution is prepared by mixing 4 mg of Pt(acac) ₂ and oleylamine in dichlorobenzene, and 0.5 μL of the precursor solution is loaded into a cell reservoir. After absorbing the excess solution with filter paper, the liquid cell is sealed with a 600 μm pore grid using vacuum grease. Monitor the growth of nanoparticles. In-situ TEM images are acquired at a frame rate of 1 fps.

Image processing

In the TEM image, noise leads to random pixels with high contrast, hindering the generation of binarized images. Accordingly, noise is removed using a biner filter, gamma value of the TEM image is adjusted using gamma correction, and Gaussian noise is removed using Gaussian.

Local Laplacian filtering of the TEM image improves the edge contrast of the particles of the TEM image. Laplacian filtering improves the accuracy of particle identification by removing minor contrast changes and highlighting high contrast changes. Finally, the TEM image is binarized with adaptive binarization to correct the background noise of the liquid TEM image. After binarization, small objects with a size smaller than 4 pixels are removed.

Ensemble analysis

Nanoparticle tracking is performed by comparing the correlation coefficient defined by the location and size between the nanoparticles in a series of TEM images. The correlation factor (F) is defined as follows.

은 n번째 프레임에서 나노입자의 면적이다. 연속 프레임에서 최소 상관 인자를 갖는 나노입자는 동일한 궤도로 그룹화된다. 이미지 드리프트는 모든 추적된 입자의 질량 변화의 중심을 사용하여 보정된다. 상기 드리프트로 인해 인 시츄 TEM 무비(movie)에서 고정 시야 내외로 움직이는 입자는 성장 메커니즘에 대한 추가 분석을 위해 제외된다. 유역(watershed) 알고리즘은 인접한 입자를 분리하는데 사용된다. 정렬을 위해 전자빔이 강하게 조사되는 영역은 비정상적으로 높은 단량체 농도를 포함하여 다른 화학적 조건을 가질 수 있기 때문에 앙상블 분석에서 제외된다.Here, x _n and y _n are the x and y positions of the nanoparticles in the nth frame,

Is the area of the nanoparticle in the nth frame. Nanoparticles with the least correlation factor in successive frames are grouped into the same trajectory. Image drift is corrected using the center of mass change of all traced particles. Particles moving in and out of the fixed field of view in the in-situ TEM movie due to the drift are excluded for further analysis of the growth mechanism. The watershed algorithm is used to separate adjacent particles. The area to which the electron beam is strongly irradiated for alignment is excluded from the ensemble analysis because it may have other chemical conditions including abnormally high monomer concentration.

Analysis of growth rate

의 관계를 사용하여 구한다. 여기서 A는 개별 나노 입자가 차지하는 면적이다. TEM 이미지의 아티팩트에서 비롯되는 입자 반경의 시간 프로파일에서 노이즈를 부드럽게 하기 위해 단계 번호가 10인 고속 푸리에 변환(FFT) 필터가 적용된다. 시간의 함수로서 성장 속도는 평활화된 반경 프로파일을 시간에 대해 미분함으로써 계산된다. 각 시간 프레임에 대한 성장 속도 대 크기 플롯은 0.2nm의 윈도우 크기를 갖는 박스-카-이동 평균법을 사용하여 얻어진다. 작은 데이터 포인트에서 발생하는 통계적 오류를 줄이기 위해 4포인트 미만으로부터 평균화된 데이터 포인트는 제거된다.Quantitative analysis of the size of individual particles is performed by applying a computational algorithm to the binarized image. The size information of individual nanoparticles is obtained by calculating the number of pixels in the region representing the individual particles in the binarized image. The effective radius (r) is

Find using the relationship of. Where A is the area occupied by individual nanoparticles. A Fast Fourier Transform (FFT) filter with step number 10 is applied to smooth out the noise in the temporal profile of the particle radius resulting from the artifacts of the TEM image. The growth rate as a function of time is calculated by differentiating the smoothed radius profile over time. The growth rate versus size plot for each time frame is obtained using a box-car-moving averaging method with a window size of 0.2 nm. Data points averaged from less than 4 points are removed to reduce statistical errors that occur with small data points.

So far, specific examples of the present invention have been described. Those of ordinary skill in the art to which the present invention pertains will be able to understand that the present invention can be implemented in a modified form without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments should be considered from an illustrative point of view rather than a limiting point of view. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope equivalent thereto should be construed as being included in the present invention.

Claims (10)

Obtaining an electron microscope image;
Processing the electron microscope image to form a binarized image; And
Electron microscope image analysis method comprising the step of analyzing the binarized image. The method of claim 1,
The electron microscope image analysis method, wherein the electron microscope image is binarized by adaptive binarization using an adaptive binarization algorithm. The method of claim 2,
An electron microscope image analysis method, characterized in that the background noise of the electron microscope image is corrected by the binarization. The method of claim 1,
Processing the electron microscope image,
Removing noise from the electron microscope image and
And increasing the edge contrast of the electron microscope image. The method of claim 4,
Removing noise from the electron microscope image is performed using a biner filter and a Gaussian filter,
The electron microscope image analysis method, characterized in that increasing the edge contrast of the electron microscope image is performed using a Laprasian filter. The method of claim 4,
Processing the electron microscope image,
The electron microscope image analysis method further comprising adjusting a gamma value of the electron microscope image using gamma correction. The method of claim 1,
The electron microscope image analysis method, characterized in that the binarized image is analyzed using a computer. The method of claim 1,
The electron microscope image includes an image of nanoparticles,
Analyzing the binarized image,
Analyzing the size and location of the nanoparticles and
Electron microscope image analysis method comprising analyzing the growth rate of the nanoparticles. The method of claim 8,
Electron microscope image analysis method, characterized in that the size and position of the nanoparticles are analyzed using an ensemble analysis method. The method of claim 1,
The electron microscope image analysis method of an electron microscope image, characterized in that it comprises a liquid transmission electron microscope image.

Images