1 Introduction

The stability of rock slopes is mainly controlled by the jointing of rock mass along with the condition of its discontinuities. Geostructural survey is a necessary procedure for the slope modelling and the stability analysis, because it allows the direct evaluation of both the degree of fracturing of the rock and its peculiar characteristics. Nevertheless, according to the growth of technology, satisfactory outcomes are nowadays achieved also through remote surveys, which are often considered important support methodologies to the field campaign. In this light, InfraRed Thermography (IRT), whose utility is known in several fields of science, has recently been employed in the rock mechanics for various purposes. In laboratory, it is a useful tool for the indirect estimation of the porosity of artificially heated rock specimens based on their cooling rate and for the identification of persistent voids [1,2,3]. In the field, it supported the detection of specific items along the unstable slopes, such as crevasses, caves and open fractures [4,5,6] and it was also used as a complementary methodology during the assessment of rockslide susceptibility scenarios [7, 8].

In this paper, a review on the most recent outcomes about the application of IRT for the survey of unstable rock slopes is reported, with the aim of presenting a general overview on some applicative cases of such pioneering practice. In particular, three cases will be reported and commented, such as the survey of heavily jointed rock masses, presented by Mineo et al. [9] and Pappalardo et al. [10]. They proposed IRT for the study of the discontinuity systems, with particular reference to those characterized by highly persistence, in order to understand what kind of satisfactory information such methodology could bring in a rock slope modeling. Moreover, they analysed the cooling behaviour of rock masses introducing the Cooling Rate Index, which well correlates to the main geomechanical parameters of the rock.

A further application of IRT discussed in this paper concerns slopes affected by landslides at a highly loosened and weathered rock masses, where Mineo et al. [11] and Mineo and Pappalardo [12] highlighted potentially unstable spots characterized by a different thermal behaviour.

The last case is focused on the analysis peculiar unstable areas of rock slopes, where an IRT survey allowed highlighting portions of the rock that can be involved in future rockfalls.

2 The InfraRed Thermography Technique

InfraRed Thermography (IRT) is a technique allowing the estimation of the surface temperature of an object by exploiting its emitted infrared radiation. This is linked to a property that all forms of matter with temperature above the absolute zero have, i.e., the thermal emissivity, whose intensity is function of the temperature of the material itself. As all the electromagnetic radiations, thermal emissivity falls within a precise portion of the electromagnetic spectrum, mostly comprised between the Ultra Violet and Infrared bands [e.g. 13, 14] (Fig. 5.1); therefore it is mostly invisible for the human eye.

Fig. 5.1
figure 1

Representation of the electromagnetic spectrum with emphasis on the infrared band [10]

Full size image

According to the known optical principles, the radiation can be deflected, focused with a lens, or reflected from surfaces [15]. Based on the amount of energy reflected or absorbed by a body, emissivity ranges from 0 (when all the energy is reflected) to 1 (when all the energy is absorbed). The last condition is related the “black bodies”, which are capable to re-emit 100% of the absorbed energy. However, these cases are purely theoretical, since bodies capable to reflect or absorb all the radiation do not exist in nature. All the materials have intermediate emissivity values and are known as “grey bodies” [9].

From the mathematical point of view, emissivity is the ratio between the infrared radiation emitted by a body and the radiation emitted by a black body at the same temperature. According to the Stefan-Boltzman law (Eq. 5.1), there is a proportionality between the whole energy emitted by an object (J) and the Stefan-Boltzman constant (σ), to the surface temperature of the body (T) to the fourth power and to its emissivity (ε) [15].

$$ \mathrm{J}={\upvarepsilon \upsigma \mathrm{T}}^4 $$
(5.1)

A thermal camera can build images using the infrared radiation, as it is a device operating in the range of wavelengths as long as 13–14 μm. It provides colour-scaled pictures showing the variation of temperature within the shot area. Such images are called “thermograms” and consist of a matrix of pixels labelled with a surface temperature value.

The radiation captured by the camera does not depend only on the surface temperature of the shot body, but it is also influenced by the surrounding ambient (i.e. parasite radiation). Moreover, it can be both partially absorbed/dispersed by the atmosphere and reflected by the object [e.g. 16, 17].

3 Application of IRT in the Geomechanical Study of Heavily Jointed Rock Slopes

This section provides a concise description of IRT shooting campaigns carried out by Mineo et al. [9] and Pappalardo et al. [10] with the aim of collecting the most recent outcomes on the applicability of such methodology to the survey of heavily fractured rock masses. Due to the pioneering aspect of this research, most of the attention is focused on what kind of information can be retrieved by the analysis of thermograms acquired both at different stages of a day and under different seasonal conditions.

3.1 Method

The campaign described herein was carried out by using a thermal camera with the following features: temperature accuracy ±2 °C or ±2%; range of assessable temperature comprised between −20° and +120 °C; spectral range of 7.5–13 μm; Field of view 25 × 19°; focal length of 18 mm; Noise Equivalent Temperature Difference (NETD) of 50 mK. Sensing device was an uncooled Focal Plane Array with IR resolution of 320 × 240 pixels.

Thermal images of rock masses (Fig. 5.2), belonging to different lithologies (dolostones, limestones and porphyroids), were taken in the dry season with the help of a tripod and an incorporated laser pointer to preserve the constant framing (Fig. 5.3a). The different lithologies have in common the tectonic history and, therefore, the presence of numerous discontinuity systems, which strongly affect the geomechanical quality of rock masses (e.g. [18]).

Fig. 5.2
figure 2

Representative rock masses of selected lithologies with corresponding geostructural stereonet

Full size image
Fig. 5.3
figure 3

Field operation during (a) summer and (b) winter campaigns [10]

Full size image

The geostructural setting of all the outcrops was previously studied, according to ISRM [19], to provide a reliable reference dataset in terms of orientation of discontinuities and main geostructural parameters [20,21,22]. To ensure the lowest number of variables, all the rock slopes chosen for this test face NE, therefore they are usually hit by the sun radiation during the morning, while they are in shadow from midday and during afternoon.

Based on the daylight condition, the set up methodology consisted in four different measurement stages during the same day:

  1. 1.

    t1: morning shooting, when the slope faces were directly hit by sun radiation.

  2. 2.

    t2: images shot at noon, when the slope had just gone under shadow.

  3. 3.

    t3: images taken in the afternoon, when the cooling phase of the slope had already started for some hours.

  4. 4.

    t4: images shot at night, when the impact of parasite radiation is minimum, since no artificial illumination was lighting the slope face.

    The same campaign was repeated in the cold season, to compare thermal images and to study differences occurring between outcomes (Fig. 5.3b).

3.2 Analysis of Thermograms

The analysis of thermograms shows that t1 images are strongly influenced by the sun radiation, which, due to the orientation of the slope face, was right hitting the indented rock masses. Under this condition the thermal output is “disturbed” by the uneven warming of the slope face, as jutting sector of the outcrop are radiated by the sunrays (resulting in positive anomalies/warm areas), thus gaining heat and shadowing the hollow portions (negative anomalies/cold areas).

During the subsequent shooting sessions (t2, t3, t4), shadowed rock masses slowly release heat through their discontinuity systems and the IRT images provide interesting hints on the degree of fracturing of the rock, with an increasing quality as time passes.

At these stages, warm areas in thermograms retrace fractures and hollow portions of the rock, while colder sectors are proper of low fractured planes, weathered portions and jutting rocks. It is underlined that t2 and t3 were carried out in daylight condition, therefore the influence of parasite radiations cannot be excluded although rock masses were already shadowed. At nighttime (darkness condition), the best definition among the IRT images is achieved (Fig. 5.4).

Fig. 5.4
figure 4

Thermal outcomes for representative rock masses of the studied lithologies at the 4 measurement stages

Full size image

Figure 5.4 shows that, with reference to t4 images, positive linear anomalies mark the main discontinuity system, emphasizing their geometrical relationship; for example, the “X” shapes at dolostones, defined by two intersecting persistent and open sets, is clearly visible in the thermogram. Warm sectors are also related to crushed and/or hollow portions, proving that IRT allows a good definition of the crushed rock, thus representing a useful help during a traditional geostructural survey, especially at slopes with closely spaced discontinuities.

Porphyroids t4 thermograms show how the highest temperatures retrace the main discontinuities along with an intensely jointed zone in the upper portion of the image. The dip-slope plane occurring in the center of the thermogram is a cold area due to its low fracturing and smooth surface. This is one of the main unstable planes, giving rise to potential planar sliding failures, and its identification is a key point during a stability analysis.

Moreover, it has been noticed that the aperture of fractures is another element affecting the thermal output, as maximum surface temperatures were found at open discontinuities (even mm). Such temperature values decrease at filled or closed fractures (Fig. 5.4). Similarly, the intensity of thermal anomalies can be related to the persistence of fractures, starting from the assumption that a persistent discontinuity concurs to the heat exchange between the rock mass and the external ambient.

By reiterating the daily campaign on the same outcrops during winter, an interesting difference in the quality of thermograms can be outlined with respect to the summer images. In particular, winter thermograms show a poor definition and sometimes are characterized by a blurred effect (Fig. 5.5). It happens because, during winter, rock masses are less heated and the ambient temperature is sensibly lower during the day, therefore winter images need a deeper processing before being compared to summer images.

Fig. 5.5
figure 5

Winter representative thermograms, showing the blurred effect due to the low warming of the rock masses (Modified after [10])

Full size image

3.3 The Cooling Rate Index

The digital processing of an IRT image permits the assessment of the maximum, minimum and average surface temperatures of the surveyed object(s). While maximum and minimum are temperature values belonging to the hottest and coldest pixels respectively, the average surface temperature is representative of the entire shooting area. Therefore, in this case, we can consider it as the “average” surface temperature of the surveyed portion of rock mass. Based on such values, recorded at different times, the cooling attitude of the rock mass can be monitored. In this light, according to Newton’s Law of Cooling, the cooling of an object in a colder ambient is described by a curve; the slope of the tangent to the curve at any point gives the rate of the temperature loss. On a two-variable diagram, the plot of average surface temperature estimated at each outcrop against time describes the cooling of the rock through a curve with a decreasing trend (Fig. 5.6).

Fig. 5.6
figure 6

Representative cooling curves presented by [10] for a the studied lithologies

Full size image

Based on this outcome, Pappalardo et al. [10] have calculated a specific index representing the variation of surface temperature per unit of time and named it Cooling Rate Index (CRI), which can be computed by Eq. 5.2.

$$ \mathrm{CRI}=\Delta \mathrm{T}/\Delta \mathrm{t} $$
(5.2)

Where ΔT is the variation of temperature and Δt is the considered time window. Therefore, the higher CRI, the faster the rock mass cooling.

With reference to surveyed rock masses, cooling curve of dolostones and limestones show decreasing trends, while in the porphyroids curve a slight heating phase between t1 and t2 is outlined. This is caused by the local setting of the outcrops, since porphyroid rock masses are partly shadowed by a protruding rock portion on the top of the outcrop at t1. In this case, the projected shadow leads to a slower warming of the rock, which lasts until t2, when the cooling phase begins (Fig. 5.6).

According to acquired data, the cooling phase of dolostones is characterized by CRI ranging between 0.65 and 0.84 °C per hour.

Limestone outcrop is affected by a CRI value around 0.41° per hour, while porphyroids have an average CRI of 0.58 °C per hour.

Achieved values are indicative of a faster cooling at dolostones, which are the most fractured rock masses showing also open discontinuities and crushed portions. This consideration leads to hypothesize that there is a link between CRI and the degree of fracturing of the rock.

Therefore, a further challenge of the research reviewed herein is to ascertain the validity of this relationship. In this light, statistical correlations between CRI and two important geomechanical indexes, i.e. the Rock Quality Designation (RQD) [23] and the Volumetric Joint count (Jv) [24], previously estimated for each geostructural station, were carried out.

Resulting CRI vs Jv scatterplot shows a positive linear trend indicating that the rock mass cools down faster as Jv increases (i.e. an increasing rock fracturing). This outcome is in accordance with the field condition of the outcrops, as dolostones (highest Jv) are characterized by the highest CRI and the best fitting is achieved by calculating the Cooling Rate only within the nighttime (t3-t4), when the temperature difference between rock outcrop and external environment is maximum (Fig. 5.7a). According to this result, the cooling of rock masses accelerate proportionally to their degree of fracturing, resulting in higher CRI at the most fractured outcrops.

Fig. 5.7
figure 7

Statistical correlations proposed by [10] between (a) CRIt3–t4 and Jv; (b) CRIt3–t4 and RQD

Full size image

Correspondingly, the correlation between the daily CRI and RQD shows a negative trend, with a satisfactory best fit for t3-t4 data (Fig. 5.7b). A low CRI corresponds to a high RQD (i.e. good geomechanical quality of the rock); this is in accordance with the principle that intact rock (i.e. rock with no visible fractures) is not a good conductor of heat.

The above reported correlations prove the reliability of CRI as a potential index for the remote assessment of the state of fracturing of a rock mass. Therefore, further survey at outcrops with different properties is currently being performed to enrich the statistics of such new survey methodology.

4 Application of IRT to Weathered Rock Slopes

This section reports on an Infrared Thermography shooting campaign aimed at individuating potentially unstable areas along a crystalline weathered rock slope characterized by a high predisposition to fail. Several landslides, classifiable as rockfalls and debris-falls according to [25, 26], repeatedly affected the studied slope, even after the execution of remedial works. Mineo et al. [11] proposed an integrated approach to study the instability mechanisms affecting this slope and highlighted that the poor geomechanical condition of the rock is not only related to the rock face, but persists inside the outcrop.

In this case, the analysis of thermograms was aimed at mapping the potentially unstable portions of the slope according to the difference of temperature related to the elements of the outcrop. This kind of survey proved a helpful tool during the stability study of rock slopes, especially where it is logistically hard to accomplish direct surveys.

4.1 IRT Shooting Campaign

The studied outcrop is a weathered paragneiss rock mass, characterized by scars of previous rockfalls, well visible among partly vegetated sectors (Fig. 5.8a). The surface temperature of the slope, along with its variations, can be employed as an indicator to recognize the different elements occurring at the slope face according to Wu et al. [5] and Mineo and Pappalardo [12].

Fig. 5.8
figure 8

Comparison between a digital photo and an IRT image of the rock slope taken from the same shooting point: (a) photo taken in daylight; (b) IRT image taken in nighttime [11]

Full size image

In this case, the shooting campaign was carried out during the dry season, in night time, when the possible influence of parasite radiations could be neglected.

The minimum and maximum surface temperatures were about 14 °C and 29 °C, respectively.

The lowest values are found along the vegetated upper part of the outcrop, while maximum values are found at the source areas of past rockfalls, that is the place where the rock is bared (Fig. 5.8b). In particular, by highlighting different ranges of surface temperatures, six different frames were considered pointing out the elements of the slope. Such approach is aimed at linking the different temperature values to the physical elements (rock, vegetation, and debris) along the slope (Fig. 5.9).

Fig. 5.9
figure 9

Slope portions with homogeneous temperature range [11]

Full size image

According to [11, 12], a first differentiation based on selected ranges of temperature can be carried out as follows:

  1. 1.

    18–19 °C: vegetated areas;

  2. 2.

    19–20 °C: weathered and fractured, partly vegetated rock;

  3. 3.

    20–21 °C: talus and debris;

  4. 4.

    21–22 °C: bare, fractured rock.

In order to enhance the thermal contrast between the mentioned elements, the considered temperature range was constricted between 19 and 24 °C (Fig. 5.10). In this way, the contribution of shrubs and trees on the top of the slope (lowest surface temperature) was avoided and the bare rock (warmest regions) was emphasized. Assuming that the bare rock crops out at the source areas of already occurred events, such sectors may be regarded as the currently stable portions where the unstable material has already failed. Therefore, particular attention is focused on all the other elements lying between these two extremes.

Fig. 5.10
figure 10

(a) Photo of the studied rock slope taken in daylight; (b) 19–24 °C thermogram of the slope; (c) Thermogram overlapped to the digital photo of the slope. Dashed ellipses indicate the most critical sectors. Numbers are the ID of each sector [12]

Full size image

In fact, lowest surface temperatures label partly vegetated and weathered sectors of the slope due to the effect of the local vegetation and of a natural moisture rate. Indeed, weathered rocks, especially when broken down to sand and silt, tend to absorb humidity from the atmosphere favoring the growth of plants and shrubs. By the radical apparatus, these concur in the disaggregation of the rock enhancing its instability.

According to these considerations, Mineo and Pappalardo [12] found 11 potentially unstable sectors along the surveyed slope face, with average surface temperatures ranging between 19.3 and 20.7 °C (Fig. 5.10c).

Identified sectors mostly border the source areas of latest events suggest that, there are still some rock materials which may fail, which was un-stabilized by the occurred landslides. In particular, a brief description of the 11 sectors is reported as follows.

  • 1 Located above the crown of a landslide occurred in 2008, this area holds an intensely weathered rock material populated by vegetation. The main criticism is related to the basal sapping process caused by the loss of the already failed material, especially in the steepest portions.

  • 2/3 These two similar sectors lie between two denudation surfaces. In particular, sector 3 seems to hold a great volume of material, whose failure could occur along the same sliding surface of the adjacent events.

  • 4 Partly vegetated rock material, not involved in previous landslide events.

  • 5 This portion of rock mass is identified by a cold region, caused by a morphological step at the top of the slope. Since this sector has not been involved in previous landslides, this step could also be considered as a potential detachment point for rock material.

  • 6/7/8 Similar in features to sector 1.

  • 9/10 Remnants of weathered and partly vegetated rock material in the middle of the slope. Due to their peculiar morphological features, these sectors could also hold a rate of accumulated talus, which is likely to be pushed down the slope by heavy rainfalls.

  • 11 This is a particular sector since the negative thermal anomaly highlights a sub-flat morphology. Apparently, it does not seem to have a relevant feature, but it lays on a protruding portion of the rock mass composed of weathered and fractured rock. Rainwater, flowing on the slope face, may stagnate on such flat morphology and infiltrate within the rock mass, thus favoring weathering and disaggregation of the already poor rock.

Mineo and Pappalardo [12] graphically summarized achieved outcomes in a map, obtained by overlapping the most critical areas to a digital image of the slope and showing the identified most potentially unstable sectors (Fig. 5.11).

Fig. 5.11
figure 11

Map of the unstable sectors of the slope showing the potential evolution of the instability [12]

Full size image

5 Application of IRT for the Study of Unstable Areas for Potential Reactivation

According to the results reported in Sect. 5.3, a further utility of the application of IRT to the analysis of rock masses was found in the study of a slope affected by sliding of wedges. The case was taken into account and was initially presented by Pappalardo et al. [27] and it is about a dolostone outcrop already affected by a rockfall, which left an evident scar on the slope face. The source area of the wedge is bordered by two intersecting discontinuities planes (S1-S2) and by a third plane (TC) acting as a tension crack since it strikes almost parallel to the face (Fig. 5.12a). It is likely that the initial rock volume slides along the intersection line and its movement was enhanced by TC.

Fig. 5.12
figure 12

(a) Digital photograph of the surveyed slope; (b) 18–22.4 °C thermogram overlapped to a digital photo of the slope; (c) 20.2–23.2 °C thermogram overlapped to a digital photo of the slope; (d) portions with surface temperature >21 °C highlighted on the slope

Full size image

IRT images were taken in nighttime, when the heat is released by the rock masses through its discontinued systems. In this view, positive anomalies are related to fractures, caves and hollow parts, as reported before. Here, negative anomalies indicate low fractured planes as well as the weathered rock and jutting portions.

5.1 Thermal Outcomes

Acquired thermograms show the alternation of cold and warm areas, apparently without any specific shape, where the two main discontinuities bordering the source area of the wedge area can be individuated (Fig. 5.12b). In fact, S1 and S2 are highlighted by warm colors, while TC represents the internal face of the source area, where the highest surface temperatures were recorded.

By processing the IRT image, so to highlight the highest temperature values related to the most fractured rock, S1 and S2 traces become clearly visible (Fig. 5.12c). In particular, S2 appears as a smooth surface externally bordered by intensely fractured rock.

Moreover, the traces of other discontinuities occurring next to the source area run parallel to the main discontinuity sets and are affected by close spacing. Such discontinuities border positive anomalies, highlighting a heavy fractured rock around the source area. This is relevant information obtainable by the analysis of a thermogram; in fact, these areas could represent potential enlargement sectors of the wedge, thus suggesting possible reactivation of the landslide movement.

A further element pointed out by Pappalardo et al. [27] is represented by the warm region occurring at TC plane, within the source area (Fig. 5.12d): such anomaly can be related to fractured and loosened rock, which is probably underwent decompression phenomena after the initial wedge sliding. This is a further hint provided by the IRT analysis, which suggests the presence of a potential source area for future failures, which would lead to retreat phenomenon of the slope.

6 Conclusions

This review paper reports an innovative approach for the remote survey of rock masses through InfraRed Thermography. Interesting results were achieved with reference to three different applications, whose outcomes are summarized by the following points, aimed at evaluating peculiar geomechanical features along rock slopes.

  1. 1.

    The first discussed case is related to a daily IRT shooting campaign commented to highlight the utility of such survey methodology applied to intensely jointed rock masses of different lithologies. The main discontinuity systems, as well as crushed portions of the rock are identified by positive thermal anomalies (highest surface temperature), while smooth and low fractured surfaces keep a lower temperature. Such an output allows mapping heavily fractured sector of the surveyed outcrop along with its open and persistent discontinuities, which are responsible for instability. Indeed, the maximum temperature was found at the most open and persistent fractures, while the lowest values were proper of closed or filled discontinuities.

    The use of IRT to highlight the different degree of fracturing along a rock slope proved a suitable tool, which could give a great contribution for the geostructural characterization of hardly-reachable rock masses, such as coastal cliffs or higher sectors.

    Furthermore, a new index describing the cooling attitude of rock masses was introduced. It was named as Cooling Rate Index (CRI) [10] and it is the expression of the cooling velocity of the rock, which is strongly conditioned by its degree of fracturing. CRI proved well correlated with two geomechanical parameters used for the classification of the rock quality, i.e., RQD and Jv. Their statistical correlation with CRI returned satisfactory positive and negative linear relationships for CRI-Jv and CRI-RQD respectively, demonstrating that the cooling speed of a rock mass is proportional to its degree of fracturing.

  2. 2.

    The second application discussed herein is focused on the IRT imaging analysis carried out to detect unstable portions along a weathered rock slope, which had already suffered numerous rockfalls and debris-falls. In this case, the utility of IRT resides in the possibility of distinguishing elements with a different thermal attitude along the slope by highlighting the contrast between vegetation, debris/weathered rock and bare rock. Shooting was carried out in nighttime, because the difference between the above mentioned elements is not always clear in daylight conditions, due to the effects of light and shadows caused by the irregularity of the rock face [11]. In this way, the potential source areas of future events, including also possible reactivation sectors of previous landslides, were mapped and 11 critical portions were identified. The potentially unstable sectors are placed both at the crown of already occurred failures and between two or more denudation areas [12].

    Achieved outcomes confirm the utility of IRT as a support methodology during the modeling of the slope stability, even in highly weathered rock masses.

  3. 3.

    According to the afore mentioned outcomes, the application of IRT for the survey of potential reactivation areas along a rock mass has been discussed, highlighting how thermal images proved useful in the detection of relevant criticalities related to such cases. In detail, IRT was aimed at analyzing the source area of a wedge failure some years ago [27]. Such methodology provided interesting information on the poor quality of the rock mass, especially around and within the source area itself. In fact, the analysis of thermograms highlighted the presence of potential enlargement sectors of the main wedge, suggesting potential reactivation of the failure, involving new volumes of rock.

Results achieved by the researches reviewed herein demonstrate the high potentiality owned by IRT when applied to the geomechanics. The detection of unstable areas along a slope is a precious hint, especially when complex rock masses are surveyed.

It is clear that such application is at its pioneering stage and that further tests in different settings are needed to refine the outcomes. Nevertheless, this paper can be considered a starting point for possible classification systems and survey methods based on the thermal behavior of rock masses. Further studies will follow approaching to the analysis of thermograms not only from a qualitative point of view, but also considering the thermal behavior of the rock over time, at any scale [28].