8
In summary, the pairplot shows that each feature can be
used to separate some of the data, and among all these fea-
tures, correlation, and contrast (dissimilarity) are the most
important. When all of these features are combined for the
classification, a good performance can be expected.
Influence of Resolution on Coal Lithotype
Classification
The GLCM is a texture analysis method that depends on
the spatial relationships between pixel values in the image.
When the coal images of different resolutions are used,
the spatial relationships between pixel values change and
different GLCM features can be obtained. The resolution
of an image refers to the amount of detail it contains, typi-
cally measured in pixels per unit area (e.g., pixels per inch
or pixels per centimeter). In order to understand how the
image resolution can affect the GLCM features and the
lithotype classification, the change of image features with
varying resolution is investigated in this section. The change
of image resolution was achieved by resizing the images,
which allows the change of image dimensions (width and
height) to achieve a different scale or resolutions. It is
important to note that resizing an image is not a perfect
substitute for capturing an image from different distances,
Figure 6. Pairplot of the GLCM features
Previous Page Next Page