| Summary: | Correlation between 3D surface roughness and characteristic features extracted from laser speckle pattern was done using an inexpensive laser pointer and a digital single lens reflex (DSLR) camera in previous research work. There had been no comparison work done between the experimental setup which uses a laser pointer, which has a diode laser as the laser source, and the experimental setup, which uses a He-Ne laser as the laser source. As such, in the current work, a comparison study between two experimental setups was carried out. One experimental setup was using a He-Ne laser, spatial filter, and charged coupled device (CCD) camera, while another experimental setup was using a laser pointer and DSLR camera. The laser beam was illuminated at angles of 30°, 45°, and 60° from the horizontal. When a laser beam falls on the surface, the beam gets scattered, and the scattered beam undergoes interference and produces speckle patterns which are captured using a camera. Using a Matlab program, the gray level co-occurrence matrix (GLCM) characteristic features, such as contrast (GLCM), correlation (GLCM), energy (GLCM), entropy (GLCM), homogeneity (GLCM), and maximum probability, and non-GLCM characteristic features, such as mean, standard deviation (STD), uniformity, entropy, normalized R, and white-to-black ratio (W/B), were extracted and correlated with 3D surface roughness parameters. The coefficient of determination (R<sup>2</sup>) was determined for each case. Compared to the setup using a laser pointer, the setup using a He-Ne laser gave better results. In the setup using the He-Ne laser, there were correlations with a coefficient of determination R<sup>2</sup> ≥ 0.7 at illumination angles of 30°, 45°, and 60°, whereas in the setup using a laser pointer, there were correlations with R<sup>2</sup> ≥ 0.7 at illumination angles of 30° and 45°. Mean characteristic features had more correlations with R<sup>2</sup> ≥ 0.7 in the case of the angle of illumination of 45° (7 out of 36 correlations) and 60° (11 out of 82 correlations), while R-normalized characteristic features had more correlations with R<sup>2</sup> ≥ 0.7 in the case of the angle of illumination of 30° (9 out of 38 correlations) for the setup using the He-Ne laser. Correlation (GLCM) had more correlations with R<sup>2</sup> ≥ 0.7 in the case of the setup using a laser pointer (2 out of 2 correlations for illumination angle of 30°, and 4 out of 19 correlations for an illumination angle of 45°). Roughness parameters <i>S<sub>a</sub></i> and <i>S<sub>q</sub></i> had more correlations with R<sup>2</sup> ≥ 0.7 for an illumination angle of 30° (1 out of 2 correlations each), and <i>S<sub>p</sub></i> and <i>S<sub>z</sub></i> had more correlations with R<sup>2</sup> ≥ 0.7 for an illumination angle of 45° (4 out of 19 correlations each) in the case of the setup using a laser pointer. The novelty of this work is (1) being a correlation study between 3D surface roughness and speckle pattern using a He-Ne laser and spatial filter, and (2) being a comparison study between two experimental setups on the correlation between 3D surface roughness and speckle pattern.
|
|---|