*5.2. Absolute Contrast*

Contrast-based methods are the simplest image processing techniques that can be applied to enhance the thermograms quality. Even though a requirement of selection of sound area is the disadvantage of these methods, they are still the most common techniques utilized to preprocess image sequences [18,28]. In this study, in order to evaluate the detectability of delaminations with different WTDRs using passive IRT, the absolute contrast (Δ*T(t)*), which is defined as the temperature difference above a defect (*Tde(t)*) and its neighborhood (*Tso(t)*) at the same time, as shown in Equation (3), is employed [28].

$$
\Delta T(t) = T\_{dc}(t) - T\_{so}(t), \tag{3}
$$

The selection of areas of region of interest (ROI) used to compute *Tde(t)* and *Tso(t)* is depicted in Figure 13, where *Tde(t)* is the average surface temperature of a ROI above a defect, called as the "delaminated area" whereas *Tso(t)* is determined from of a ROI nearby the delamination denoted as the "sound area". The average surface temperature is employed as mentioned by Vaghefi in 2013 that is more effective compared to only one pixel or a group of three pixels [26,52,53]. The size of delaminated areas is similar with the dimension of respective delamination while the size of all sound areas is same (around of 8 cm × 8 cm). Nine sound areas at the middle position of defects are selected for twelve delaminations. The level of upper edges of delaminated and sound area for each delamination are similar. There are three couples of delaminations at the middle region of specimen (D2 and D3, D6 and D7, D10 and D11) in which each pair refers to the same sound area to avoid the misdetection phenomenon. As per the ASTM Standards, a delamination is considered as detectable if the amplitude of the absolute contrast (Δ*T(t)*) known as "temperature difference" is 0.5 ◦C or higher [51]. Thus, this threshold is used in our present study. The higher the absolute contrast, the greater the certainty that delamination may appear.

**Figure 13.** Selection of delaminated and sound areas on the structure surface.

In the experiment, delaminations with depth of 5, 6, and 7 cm (F-D1 to F-D12) were not detected at any given time during the test (day 3 and day 4). Then, the absolute contrast is analyzed focusing on the back face that delaminations (B-D1 to B-D12) are located at depths equal to or less than 4 cm (day 1 and day 2). Figure 14 shows the absolute contrast of all delaminations during the sunny days. In each graph, delaminations with the same depth but different sizes are depicted. The absolute contrast profiles have a similar tendency in comparison with the surface temperature and ambient temperature. In general, the absolute contrast increases quickly and reaches a positive peak value around noon. After that, it reduces and attains the maximum negative value before increasing again during early morning in the next day.

Owing to the formation of the volume of trapped heat under the daytime heating effect, the surface above a delamination becomes warmer than its surrounding while it is cooler at night because of the nighttime cooling effect. It is shown in the graphs in Figure 14 that there are two interchange periods between positive values (daytime heating effect) and negative values (nighttime cooling effect) of the absolute contrast. The first interchange period occurred due to the shift from cooling effect during the nighttime to heating effect during the daytime, while the second period is caused by the change from the daytime heating effect to the nighttime cooling effect. At the interchange period, delaminations cannot be observed by the IR camera because the temperature difference between the delaminated area and sound area is small (≤0.5 ◦C). The interchange period lasts for approximately 2 h during both the morning time (from 6:00 to 8:00) and afternoon time (from 16:30 to 18:30) under the given experimental conditions.

**Figure 14.** Absolute contrast of delaminations with different depths and sizes on day 1: (**a**) at the depth of 2 cm; (**b**) at the depth of 3 cm; (**c**) at the depth of 4 cm.

The effect of the size of delaminations on the temperature difference is studied as well. Under the heating effect during daytime and cooling effect during nighttime, a larger delamination produces a higher temperature difference than a smaller delamination. This phenomenon is caused by the intensity of the trapped heat volume and effect of the diffusion heat. For example, at 11:30, the temperature difference of delaminations B-D1, B-D2, B-D3, and B-D4, which have the same depth of 3 cm and different sizes of 13.5, 12.0, 9.0, and 6.0 cm, respectively, corresponds to 6.86, 5.32, 4.80, and 2.48 ◦C. This implies that delamination B-D1 can be observed more clearly than delaminations B-D2, B-D3, and B-D4 on the thermal image, as illustrated in Figure 15.

**Figure 15.** Delaminations (B-D1 to B-D8) on thermal image at 11:30 on day 1.

In terms of WTDR, it is indicated that a larger WTDR delamination can produce a higher temperature difference than a smaller WTDR one at the same depth. In addition, the conclusion can be given from the present study that a delamination, which has a larger WTDR, can produce a higher temperature difference even though it is located at greater depth than a smaller delamination as shown in Figure 16. For example, in Figure 16, delamination B-D5 and B-D6 obtained temperature differences of 3.84 ◦C and 2.94 ◦C whereas delamination B-D4 attained a temperature difference of 2.48 ◦C at 11:30. Moreover, the temperature differences are 1.73 ◦C, 1.28 ◦C, and 0.82 ◦C corresponding to delaminations B-D5, B-D6, and B-D4 at 20:00. Therefore, delamination B-D5 and B-D6 can appear more certainly than B-D4 during both the daytime and nighttime (Figures 11 and 15). Furthermore, a delamination placed at a relatively greater depth experiences a time delay to obtain the maximum temperature difference than a delamination located at a relatively shallower depth from the surface as demonstrated in Figure 16. In detail, delaminations B-D5 and B-D6 achieve the maximum temperature difference at 11:30 whereas delamination B-D4 obtains the maximum temperature difference at 11:00.

**Figure 16.** Absolute contrasts of delaminations with different WTDRs on day 1.

The maximum absolute contrasts of all delaminations composed in this study are graphically depicted in Figure 17. The red solid and red dashed line represent the maximum absolute contrast during daytime and nighttime on a sunny day, respectively. It can be concluded that a higher WTDR delamination obtains a greater maximum temperature difference than a smaller WTDR delamination on a sunny day. The maximum temperature difference rises from 0.49 ◦C to 6.95 ◦C (positive value) and from 0.32 ◦C to 3.00 ◦C (negative value) corresponding to the daytime and nighttime when the WTDR increases from 1.25 to 7.9. However, under the effect of rain, the above-mentioned trend is not observed clearly although the maximum temperature difference may tend go up with the rise of WTDR, as indicated in Figure 17. In addition, almost all delaminations obtain the maximum temperature difference smaller than 0.5 ◦C. Thus, it is recommended that the concrete bridge deck inspection must not be conducted while it is raining or after rain.

**Figure 17.** Maximum absolute contrast of delaminations during daytime and nighttime on both days 1 and 2.
