3.1 Image processing method: Image-Pro Plus and PCAS
Image-Pro Plus converts a color image into a grayscale image and then converts the grayscale image into a binary image. This software can extract more than 60 measured parameters by measuring the attributes of objects, including area and perimeter, measuring the lengths of lines and short axis, focusing on the center point of a z-stack, as well as reporting on the fractal dimension of an object, etc. These could be done both manually and automatically. The measured parameters are selected based on the requirements of the user. The target parameters are then shown and the targets automatically numbered. The main function of PCAS is to identification, geometric quantification and statistical analysis of microscopic images of particles, cracks and pores. PCAS can automatically remove noise from binary images, automatically segment and accurately identify soil pores (Chun L. et al,2011). SEM image processing steps are shown in Fig. 2.
(1) Pre-processing of SEM Images
The SEM image of undisturbed soil was selected as an example to illustrate the SEM image preprocessing method. First, Image-Pro Plus was used to correct distortions of the image background, and eliminate uneven brightness. Then, the background corrected image was processed to enhance the contrast of image intensity by adjusting the brightness, contrast and gamma correction. Finally, filter was used to improve the measurement accuracy and image quality.
(2) Thresholding of SEM Images
In carrying out the image segmentation of processed images in Image-Pro Plus, the range limit bar was dragged into the image segmentation dialog box to change the threshold pixel value to a value that is between 0 to 255 and the original image was continuously compared to the image with changing pixel values. Visual segmentation was adopted and carried out multiple times to reduce errors until there was optimal segmentation.
(3) Morphological Processing of SEM Images
After binarizing the threshold image, some isolated black spots and bright spots were found, and this study dealt with the isolated spots with the help of automatic analysis of the pore system of PCAS. Through the corrosion operation on the image, the fine connections between pores were eliminated, so as to identify and separate the particles that contact and overlap each other in the image. Image-Pro Plus was also used to quickly obtain parameters such as the diameter, area, perimeter, fractal dimension, and roundness of the contaminated clay particles and pores.
3.2 Morphology of Microstructure of Contaminated Clay
The control sample and seepage municipal waste contaminated soil samples (taken at column depths of 3.1 cm, 17.3 cm, 36.8 cm, 48.5 cm, 56.3 cm and 75.8 cm) and uncontaminated soil and soaking municipal waste contaminated soil samples (concentration were 0.2mol/L, 0.4mol/L, 0.6mol/L CaCO3 and CH3COOH) were scanned at different magnifications (1200, 1500, 1800, and 2100 x). The research results of Tang et al. (2008) were incorporated, it is not the case that the larger the magnification is, the better the observation effect is, and the magnification is selected in the range of 1500 ± 300 as the best. Therefore, SEM images magnified 1500 x were selected for the control sample and seepage municipal waste contaminated soil samples, and SEM images magnified 1200 x were selected for uncontaminated soil and soaking municipal waste contaminated soil samples. Before the observation, the sample table was cleaned by the cleaning instrument, and then a small amount of conductive adhesive was dipped with a wooden stick, and the sample was firmly plated with gold on the clay surface. The SEM images of the control sample and seepage municipal waste contaminated soil samples at depths of 17.3 cm, 56.3 cm and 75.8 cm and the binary images after treatment are shown in Fig. 3. The SEM images of uncontaminated soil and CaCO3 and CH3COOH contaminated soil samples at concentration of 0.2 mol/L, 0.4mol/L and 0.6mol/L and the binary images after treatment are shown in Fig. 4.
3.3 Effective threshold range of image pixel values based on apparent porosity ratio of contaminated clay
The pore size, pore area and fractal dimensions were used as the measured parameters of the pores, and particle size, area and quantity, fractal dimension of the particles, and other parameters were applied as the measured parameters of the clay particles. Using the data collected by Image-Pro Plus, the apparent porosity ratios and image pixel values for the seepage municipal waste contaminated soil are plotted in Fig. 5, the apparent porosity ratios and image pixel values for the soaking municipal waste contaminated soil are plotted in Fig. 6.
Table 1 shows that the porosity ratio of the whole soaking municipal waste contaminated soils is 0.97. Table 3 shows that the porosity ratio of the seepage municipal waste contaminated clay at column depths of 3.1 cm, 17.3 cm, 36.8 cm, 48.5 cm, 56.3 cm, and 75.8 cm is 0.34, 0.64, 0.73, 0.79, 0.81, and 0.83, respectively. At a pixel value of 130, the apparent porosity ratio of the clay at a column depth of 3.1 cm is about 0.31 which approximates the measured porosity ratio of 0.34. Therefore, a pixel value of 130 is recommended for the clay sample at a column depth of 3.1 cm. Similarly, the pixel value for the clay samples based on the proximity of the apparent porosity ratio to the measured porosity ratio at column depths of 17.3 cm, 36.8 cm, 48.5 cm, 56.3 cm, and 75.8 cm is 130, 130, 130, 110, and 110, respectively. The reasonable pixel value of CaCO3 contaminated clay with concentrations of 0.2mol/L, 0.4mol/L and 0.6mol/L is 120, 130 and 140, respectively. The reasonable pixel value of CH3COOH contaminated clay with concentrations of 0.2mol/L, 0.4mol/L and 0.6mol/L is 130, 130 and 110, respectively.
3.4 Effective threshold range of image pixel values based on fractal dimension of contaminated clay particles
Fractal theory uses fractal dimension to describe the shape of particles (Pi,Z. et al., 2021). The fractal dimension is determined by image analysis to quantify the soil structure (Nakatsuka et al., 2016). Many previous studies have shown that fractal dimension is a commonly used parameter to study the microstructure of soil (Hong Sun et al., 2020; Risović, D. et al., 2008; Zou X. et al, 2021; Munoz et al., 2014).
The relationship between fractal dimension and the image pixel threshold value of the seepage municipal waste contaminated soil and the soaking municipal waste contaminated soil is plotted in Fig. 7 and Fig. 8.
With increases in the pixel value, the fractal dimension of the particles increases at first, then tends to a constant value, and then decreases substantially. Tang et al. (2008) indicated that the fractal dimension of clay particles should be examined within the constant interval. The fractal dimension of the contaminated clay particles at a column depth of 3.1 cm has a relatively constant value at a pixel value interval of 110 ~ 170. Similarly, the pixel value interval for the fractal dimension of the seepage municipal waste contaminated clay particles at a column depth of 17.3 cm, 36.8 cm, 48.5 cm, 56.3 cm and 75.8 cm is 110 ~ 190, 110 ~ 190, 110 ~ 210, 110 ~ 170, and 90 ~ 150, respectively; the pixel value interval for the fractal dimension of CaCO3 contaminated clay with concentrations of 0.2mol/L, 0.4mol/L and 0.6mol/L is 80 ~ 150, 110 ~ 140 and 90 ~ 180, respectively; the pixel value interval for the fractal dimension of CH3COOH contaminated clay with concentrations of 0.2mol/L, 0.4mol/L and 0.6mol/L is 100 ~ 140,110 ~ 150 and 100 ~ 180, respectively. In summary, the optimal pixel value interval is between 110 and 140 based on the morphological structure of the contaminated clay particles, and calculations based on this range are more reliable.
3.5 Determining effective threshold range of image pixel values for contaminated clay
The focus is on the clay pores and particles when the clay microstructure is examined. The optimal threshold range of image pixel values needs to take the pores and particles of clay into consideration. In this study, the effective threshold range of image pixel value was found to be 110–140. When the fractal dimension of the contaminated clay particles was also taken into consideration, the effective threshold range of the pixel value was 110–140. Therefore, the optimal threshold range of the pixel value based on the microscopic structure of municipal waste contaminated clay was 110–140.