Search

The recent spread of low-cost and high-quality RGB-D and infrared sensors has supported the development of Natural User Interfaces (NUIs) in which the interaction is carried without the use of physical devices such as keyboards and mouse. In this paper, we propose a NUI based on dynamic hand gestures, acquired with RGB, depth and infrared sensors. The system is developed for the challenging automotive context, aiming at reducing the driver&rsquo, s distraction during the driving activity. Specifically, the proposed framework is based on a multimodal combination of Convolutional Neural Networks whose input is represented by depth and infrared images, achieving a good level of light invariance, a key element in vision-based in-car systems. We test our system on a recent multimodal dataset collected in a realistic automotive setting, placing the sensors in an innovative point of view, i.e., in the tunnel console looking upwards. The dataset consists of a great amount of labelled frames containing 12 dynamic gestures performed by multiple subjects, making it suitable for deep learning-based approaches. In addition, we test the system on a different well-known public dataset, created for the interaction between the driver and the car. Experimental results on both datasets reveal the efficacy and the real-time performance of the proposed method.

Debris flow events, generated by surface runoff, occur with great frequency in the Dolomites (Northeastern Italian Alps) during the summer season. Summer thunderstorms, which are common in the region, can quickly generate runoff at the base of rocky cliffs, which then entrains and propagates downstream the underlying unconsolidated material. In the past, the main atmospheric feature considered in evaluating the initiation of debris flow events was rainfall. Observations led to the development of rainfall intensity–duration thresholds for sediment mobilization, which compared incoming severe rainfalls with the potential for triggering debris flows. This study works toward the examination of another characteristic of the atmosphere, the atmospheric electric field. In particular, the behavior of the electric field prior to convective rainfall is investigated as an indicator of rainfall intensities capable of triggering debris flows, in a basin near Cortina d’Ampezzo (Italy). Results suggest that prior to bursts of intense rainfall, the electric field derivative frequency distribution exhibits a recurrent pattern roughly half the time. When it occurs, the amplitude of derivative frequency distribution intersects the zero axis twice before rainfall reaches maximum intensity. A regression model is designed which considers the amplitude maximum and the difference in time between the crossings of the zero axis. The validation of this model suggests a mild relationship between electric field and rainfall intensity in an alpine environment.

Rainfall is the most significant factor for debris flows triggering. Water is needed to saturate the soil, initiate the sediment motion (regardless of the mobilization mechanism) and transform the solid debris into a fluid mass that can move rapidly downslope. This water is commonly provided by rainfall or rainfall and snowmelt. Consequently, most warning systems rely on the use of rainfall thresholds to predict debris flow occurrence. Debris flows thresholds are usually empirically-derived from the rainfall records that caused past debris flows in a certain area, using a combination of selected precipitation measurements (such as event rainfall P, duration D, or average intensity I) that describe critical rainfall conditions. Recent years have also seen a growing interest in the use of coupled hydrological and slope stability models to derive physically-based thresholds for shallow landslide initiation.In both cases, rainfall thresholds are affected by significant uncertainty. Sources of uncertainty include: measurement errors; spatial variability of the rainfall field; incomplete or uncertain debris flow inventory; subjective definition of the “rainfall event”; use of subjective criteria to define the critical conditions; uncertainty in model parameters (for physically-based approaches). Rainfall measurement is widely recognized as a main source of uncertainty due to the extreme time-space variability that characterize intense rainfall events in mountain areas. However, significant errors can also arise by inaccurate information reported in landslide inventories on the timing of debris flows, or by the criterion used to define triggering intensities.This study analyzes the common sources of uncertainty associated to rainfall thresholds for debris flow occurrence and discusses different methods to quantify them. First, we give an overview of the various approaches used in the literature to measure the uncertainty caused by random errors or procedural defects. These approaches are then applied to debris flows using real data collected in the Dolomites (Northen Alps, Itay), in order to estimate the variabilty of each single factor (precipitation, triggering timing, triggering intensity..). Individual uncertainties are then combined to obtain the overall uncertain of the rainfall threshold, which can be calculated using the classical method of “summation in quadrature” or a more effective approach based on Monte Carlo simulations. The uncertainty budget allows to identify the biggest contributors to the final variability and it is also useful to understand if this variability can be reduced to make our thresholds more precise.

Runoff-generated debris flows are a significant hazard in steep mountain ranges across the world. During intense rainfall storms, runoff can rapidly form in small steep basins and mobilise large volumes of sediment triggering debris flows which can damage infrastructure and endanger lives. A common method for forecasting debris flows is deriving empirical rainfall intensity–duration (ID) thresholds from previously recorded debris flow events in a given area. However, the storms which trigger debris flows usually are short and intense with high spatial variation making an accurate recording of the conditions responsible for initiation difficult.In this study, we investigate the impact of the spatial variability of rainfall on debris flow initiation in small, steep, and debris flow prone catchments in the eastern Italian Alps (Dolomites) using the SWEHR (Shallow Water Equation Hairsine-Rose) numerical model. The modelled catchments are monitored by multiple rain gages which we use to quantify the uncertainty of the rainfall ID thresholds due to the spatial variation of rainfall by comparing empirical and numerically modelled thresholds. We also compare simulated triggering discharges for debris flows with available field observations in the study area. This study will help to improve the quality of hazard forecasting of debris flows in mountainous regions

Transformer-based neural networks represent a successful self-attention mechanism that achieves state-of-the-art results in language understanding and sequence modeling. However, their application to visual data and, in particular, to the dynamic hand gesture recognition task has not yet been deeply investigated. In this paper, we propose a transformer-based architecture for the dynamic hand gesture recognition task. We show that the employment of a single active depth sensor, specifically the usage of depth maps and the surface normals estimated from them, achieves state-of-the-art results, overcoming all the methods available in the literature on two automotive datasets, namely NVidia Dynamic Hand Gesture and Briareo. Moreover, we test the method with other data types available with common RGB-D devices, such as infrared and color data. We also assess the performance in terms of inference time and number of parameters, showing that the proposed framework is suitable for an online in-car infotainment system.

Landslides are widespread geological features in Italy’s Northern Apennines, with slow-moving earthflows among the most common types. They develop in fine-grained rocks and are subject to periodic rainfall-induced reactivations alternating to phases of dormancy. In this paper, we use radar interferometry (InSAR) and information about landslide activity to investigate deformation signals on an areal basis and to assess the dynamics of recently reactivated earthflows. We use traditional two-pass interferometry by taking advantage of the short revisit time of the Sentinel 1 satellite to characterize 4 years of slope deformations over the 60 km2 study area, where 186 landslides are mapped. Our results show that most intense and sustained deformation signals are associated with phenomena on the verge of reactivation, indicating that radar interferometry may have a potential for early warning purposes. By focusing on three specific earthflow reactivations, we analyze their dynamics through the years that preceded their failure. Despite inherent uncertainties, it was possible to retrieve the deformation signal’s temporal evolution, which displayed seasonally recurring accelerations, peaking during the major precipitation episodes in the area.

In this work we propose a deep learning pipeline to predict the visual future appearance of an urban scene. Despite recent advances, generating the entire scene in an end-to-end fashion is still far from being achieved. Instead, here we follow a two stages approach, where interpretable information is included in the loop and each actor is modelled independently. We leverage a per-object novel view synthesis paradigm; i.e. generating a synthetic representation of an object undergoing a geometrical roto-translation in the 3D space. Our model can be easily conditioned with constraints (e.g. input trajectories) provided by state-of-the-art tracking methods or by the user itself. This allows us to generate a set of diverse realistic futures starting from the same input in a multi-modal fashion. We visually and quantitatively show the superiority of this approach over traditional end-to-end scene-generation methods on CityFlow, a challenging real world dataset., Accepted at ICPR2020

The implications of the severe lockdown regime introduced in Italy in the context of the Covid-19 emergency can be correctly understood only through a broader look at how the text of the provisions adopted by the government is transformed by media reporting and law enforcement practice. From such a perspective, it appears clearly that we are witnessing nothing more than the most recent segment of a populist approach to the use of legal tools, the history of which starts well before the pandemic.

Landslides are widespread landscape features in the Northern Apennine mountain chain and their activity frequently cause damages to settlements and infrastructures. In such context, slow-moving landslides are very common and typically affect fine-grained weathered rocks. Long periods of sustained slow-movements (cms/year) can be interrupted by rapid acceleration and catastrophic failures (ms/day) that are caused by intense rainfall events. Space-borne synthetic aperture radar interferometry (InSAR) proved effective to detect actively deforming phenomena and monitor their evolution in the periods before and after failures. We present InSAR results derived from the Sentinel 1 satellite constellation for landslide cases that underwent reactivation during 2019. In all cases, the catastrophic failures were unexpected and no ground-based monitoring data are available. We processed pre- and post-failure interferograms of SAR images acquired by Sentinel 1 A/B with time spans ranging from 6 to 24 days, removing those having low coherence by manual inspection. The conventional 2-pass technique allowed us to obtain measurements of surface displacement despite the fact that sparse to none infrastructures nor bare rock outcrops are present on the landslide bodies. Our interferograms show that surface displacements are visible well in advance of the actual failure. They display nearly continuous downslope motion with seasonal velocity changes. Time series between 2015 and 2019 shows that surface displacements can be appreciated throughout most part of the year with snow cover and summer peak of vegetation being the most notable exceptions. Distinct accelerations can be detected in space and time during the weeks and months preceding the reactivation.We compare time-dependent deformations to precipitation patterns to explore their relationship and to document the transition from stable to unstable deformation. Our work suggests that InSAR interferometry can be successfully used to measure pre-failure displacements and detect slow-moving landslides that are more prone to reactivation in case of rainfall events.

Slow moving landslides are widespread geomorphological features in the Northern Apennines of Italy where they represent one of the main landscape forming processes. The lithology of the Northern Apennines fold and thrust belt is characterized by alternations of sandstone, siltstone and clayshales, also known as flysch, and clay shales with a chaotic block in matrix fabric, which are often interpreted as tectonic or sedimentary melanges. While flysch rocks with a high pelitic fraction host earthslides that occasionally evolve into flow like movements, earthflows are the dominant landslide type in chaotic clay shales. In the present work, we document the kinematic response to rainfall of landslides in these different lithologies using radar interferometry. The study area includes three river catchments in the Northern Apennines. Here, the Mediterranean climate is characterized by two wet seasons during autumn and spring respectively, separated by dry summers and winters with moderate precipitation. We use SAR imagery from the X-band satellite COSMO SkyMed and from the C-band satellite Sentinel 1 to retrieve spatial displacement measurements between 2009 and 2016 for 25 landslides in our area of interest. We also document detailed temporal and spatial deformation signals for eight representative landslides, although the InSAR derived deformation signal is only well constrained by our dataset during the years 2013 and 2015. In spring 2013, long enduring rainfalls struck the study area and numerous landslide reactivations were documented by the regional authorities. During 2013, we measured higher displacement rates on the landslides in pelitic flysch formations compared to the earthflows in the clay shales. Slower mean velocities were measured on most landslides during 2015. We analyse the temporal deformation signal of our eight representative landslides and compare the temporal response to precipitation. We show that earthslides in pelitic flysch formations accelerate faster than earthflows in chaotic clay shales and reach higher velocities, while the kinematic behaviour of the earthflows can be described as rather steady with only minor accelerations. Although we have no detailed pore pressure measurements for the period of interest, the observed behaviour can be explained in our view by the morphological and hydrological characteristics of the different landslide types. On the one hand landslide material and bedrock in the pelitic flysch rocks are more resistant, which is why slope angles are higher in this lithology. On the other hand, landslides in the pelitic flysch formations have often deeper slip surfaces and landslide material is more permeable. This is why long persistent rainfall is necessary to saturate the landslide material and induce pore pressures that are high enough to trigger displacement.

Landslide is one of the repeated geological hazards during rainy season, which causes fatalities, damage to property and economic losses in Korea. Landslides are responsible for at least 17% of all fatalities from natural hazards worldwide, and nearly 25% of annual casualties caused by natural hazards in Korea. Due to global climate change, the frequency of landslide occurrence has been increased and subsequently, the losses and damages associated with landslides also have been increased. Therefore, accurate prediction of landslide occurrence, and monitoring and early warning for ground movements are very important tasks to reduce the damages and losses caused by landslides. Various studies on landslide prediction and reduction in landslide damage have been performed and consequently, much of the recent progress has been in these areas. In particular, the application of information and geospatial technologies such as remote sensing and geographic information systems (GIS) has greatly contributed to landslide hazard assessment studies over recent years. In this paper, the recent advances and the state-of-the-art in the essential components of the landslide hazard assessment, such as landslide susceptibility analysis, runout modeling, landslide monitoring and early warning, were reviewed. Especially, this paper focused on the evaluation of the landslide susceptibility using probabilistic approach and physically based method, runout evaluation using volume based model and dynamic model, in situ ground based monitoring techniques, remote sensing techniques for landslide monitoring, and landslide early warning using rainfall and physical thresholds.

Rainfall thresholds for the occurrence of debris flows are commonly defined by Intensity-Duration curves (ID thresholds). Interestingly, many empirical ID thresholds show up as straight lines in a log-log plot and therefore can be expressed by power-law functions of the form I=α⋅Dβ, where α is the scaling coefficient and β is the exponent of the power function. The different values of α and β reflect the variability of geological and hydrological conditions in the different areas. In most cases, however, field conditions are so complex that a quantitative interpretation of the empirical rainfall threshold is impossible and the physical meaning of α and β remains obscure. In this work, we provide a physical interpretation of the rainfall thresholds that characterize an active debris flow catchment in the Eastern Italian Alps (the Dimai Basin, Belluno Province). The catchment is affected by frequent debris flows generated by surface-water runoff and has been monitored since 2010 to investigate the initiation process. Monitoring data allowed for the detection of two rainfall thresholds: a lower one that identifies the arrival of water in the initiation area (Catchment Outflow Threshold) and an upper one that identifies the initiation of debris flows by channel runoff (Debris Flow Threshold). We demonstrate that these two thresholds can be satisfactorily reproduced by a simple physically based model in which the excess rainfall is routed over the catchment using a kinematic-wave scheme. This simple analysis provides a sound explanation of the observed thresholds and can be used to develop physically based thresholds for runoff-generated debris flows.

In the Dolomitic region, abundant coarse hillslope sediment is commonly found at the toe of rocky cliffs. Ephemeral channels originate where lower permeability bedrock surfaces concentrate surface runoff. Debris flows initiate along such channels following intense rainfall and determine the progressive erosion and deepening of the channels. Sediment recharge mechanisms include rock fall, dry ravel processes and channel-bank failures. Here we document debris flow activity that took place in an active debris flow basin during the year 2015. The Cancia basin is located on the southwestern slope of Mount Antelao (3264 m a.s.l.) in the dolomitic region of the eastern Italian Alps. The 2.5 km2 basin is incised in dolomitic limestone rocks. The data consist of repeated topographic surveys, distributed rainfall measurements, time-lapse (2 s) videos of two events and pore pressure measurements in the channel bed. During July and August 2015, two debris flow events occurred, following similarly intense rainstorms. We compared rainfall data to existing rainfall triggering thresholds and simulated the hydrological response of the headwater catchment with a distributed model in order to estimate the total and peak water discharge. Our data clearly illustrate how debris entrainment along the channel is the main contributor to the overall mobilized volume and that erosion is dominant when the channel slope exceeds 16°. Further downstream, sediment accumulation and depletion occurred alternately for the two successive events, indicating that sediment availability along the channel also influences the flow behaviour along the prevailing-transport reach. The comparison between monitoring data, topographical analysis and hydrological simulation allows the estimation of the average solid concentration of the two events and suggests that debris availability has a significant influence on the debris flow volume. © 2020 John Wiley & Sons, Ltd.

Landslides are common landscape features in the Northern Apennine mountain chain and cause frequent damages to human structures and infrastructure. Most landslides in the area can be classified as earthflows, where the clay-shales form the substrate, whereas complex landslides with flow and sliding components are common on the slopes where fine-grained turbidites form the substrate. Most of these landslides move periodically with contained velocities and, only after particular rainfall events, some of them accelerate abruptly. Space-borne synthetic aperture radar interferometry (InSAR) provides a particularly convenient way for studying the periods before and after failures. In this paper, we present InSAR-results derived from the Sentinel 1 satellite constellation for two landslide cases in the Northern Apennines. The first case is a complex landslide that is hosted on a pelitic flysch formation, whereas the second case is an earthflow located in chaotic clay shales. Both cases failed catastrophically and threatened or damaged important infrastructures. In the case of the complex landslide, we report spatially variations of the deformation field between repeated periods of acceleration. The data illustrate that the deformation initiated in the upper part of the slope and expanded over the whole landslide body afterward. In the case of the earthflow, we describe spatial and temporal kinematics during the period before a catastrophic failure in March 2018. We discuss the temporal deformation signal together with rainfall and snowmelt data from a nearby meteorological station. Deformation and precipitation data highlight that high total precipitation can be considered the trigger of the failure.

Slow moving earth-flows are a common geomorphological feature in the Northern Apennines and one of the main landscape forming agents. Often houses and small villages were built on these phenomena which are as consequence affected by damages that are caused by small displacements. In this short note, we present the InSAR results obtained from the Stanford Method of Persistent Scatterers (StaMPS) for Cosmo SkyMed and Envisat imagery together with a geologicalgeomorphological survey. We focus on the village of Gaggio Montano located in the Apennines south of Bologna. Here the tectonically sheared Flysch of the Monghidoro Formation structurally overlies the mélange of the Palombini shales and different complex landslides were mapped on the slope. We show that several landslide bodies are moving with steady deformation rates at least since 2003 and that the results obtained from Envisat and Cosmo SkyMed compare well to each other. Moreover the displacements measured by InSAR are in the same order of magnitude as displacements that have been measured in the past with inclinometers. Because the higher deformation rates are closely related to young houses and industrial complexes, it is possible that anthropogenic factors, like leakage from the water supply/discharge system, influence the displacements. A steady infiltration of water from the Flysch slabs into the clay shales might however represent a possible explanation.

In dolomitic headwater catchments, intense rainstorms of short duration produce runoff discharges that often trigger debris flows on the scree slopes at the base of rock cliffs. In order to measure these discharges, we placed a measuring facility at the outlet (elevation 1770 m a.s.l.) of a small, rocky headwater catchment (area ∼ 0.032 km2, average slope ∼ 320%) located in the Venetian Dolomites (North Eastern Italian Alps). The facility consists of an approximately rectangular basin, ending with a sharp-crested weir. Six runoff events were recorded in the period 2011-2014, providing a unique opportunity for characterizing the hydrological response of the catchment. The measured hydrographs display impulsive shapes, with an abrupt raise up to the peak, followed by a rapidly decreasing tail, until a nearly constant plateau is eventually reached. This behavior can be simulated by means of a distributed hydrological model if the excess rainfall is determined accurately. We show that using the Soil Conservation Service Curve-Number (SCS-CN) method and assuming a constant routing velocity invariably results in an underestimated peak flow and a delayed peak time. A satisfactory prediction of the impulsive hydrograph shape including peak value and timing is obtained only by combining the SCS-CN procedure with a simplified version of the Horton equation, and simulating runoff routing along the channel network through a matched diffusivity kinematic wave model. The robustness of the proposed methodology is tested through a comparison between simulated and observed timings of runoff or debris flow occurrence in two neighboring alpine basins. This article is protected by copyright. All rights reserved.

The Dolomite Alps of northeastern Italy experience debris flows with great frequency during the summer months. An ample supply of unconsolidated material on steep slopes and a summer season climate regime characterized by recurrent thunderstorms combine to produce an abundance of these destructive hydrogeologic events. In the past debris flow events have been studied primarily in the context of their geologic and geomorphic characteristics. The atmospheric contribution to these mass wasting events has been limited to recording rainfall and developing intensity thresholds for debris mobilization. This study aims to expand the examination of atmospheric processes that preceded both locally intense convective rainfall (LICR) and debris flows in the Dolomite region. 500 hPa pressure level plots of geopotential heights were constructed for a period of three days prior to debris flow events to gain insight into the synoptic scale processes which provide an environment conducive to LICR in the Dolomites. Cloud-to-ground (CG) lightning flash data recorded at the meso-scale were incorporated to assess the convective environment proximal to debris flow source regions. Twelve events were analyzed and from this analysis three common synoptic scale circulation patterns were identified. Evaluation of CG flashes at smaller spatial and temporal scales illustrated that convective processes vary in their production of CG flashes (total number) and the spatial distribution of flashes can also be quite different between events over longer periods. During the 60 min interval immediately preceding debris flow a majority of cases exhibited spatial and temporal collocation of LICR and CG flashes. Also a number of CG flash parameters were found to be significantly correlated to rainfall intensity prior to debris flow initiation.

The Piees de ra Mognes fan at the base of the Punta Nera cliffs, in the Venetian Dolomites (Italy), has been subject to debris flow activity for decades. Until recently, these debris flows never reached the National Road 51 on the valley bottom. Debris flows usually initiated at the base of an incised rocky channel in the Punta Nera cliffs where runoff is delivered to loose scree deposits of the fan. The main debris flow channel is strongly incised at the apex of the fan and splits into several minor channels at lower elevations. During the autumn 2014 and May 2016, two cliff collapses produced large debris deposits. Since then, the frequency of debris flows increased considerably because of the availability of debris deposits at very steep slope that lowered the runoff discharge needed for the debris flow initiation. In a few cases, debris flows that initiated in the rocky channel reached and interrupted the National Road 51, about 2 km downstream the well-known touristic village of Cortina d���Ampezzo. On July 2016, a monitoring station was placed at the beginning of the debris flow channel just downstream the base of the rocky channel. In the period between July and -September, the monitoring station recorded six debris flow events. Analysis of these data is used to describe the characteristics of debris flow initial routing. Moreover, we use video image analysis to investigate the velocity and depth of the surge from the 5 August 2016 event.