Designing a Computer Vision System for Underwater Robotic Interventions

This thesis deals with the challenging problem of designing systems able to perceive objects in underwater environments. In the last few decades research activities in robotics have advanced the state of art regarding intervention capabilities of autonomous systems. State of art in fields such as localization and navigation, real time perception and cognition, safe action and manipulation capabilities, applied to ground environments (both indoor and outdoor) has now reached such a readiness level that it allows high level autonomous operations. On the opposite side, the underwater environment remains a very difficult one for autonomous robots. Water influences the mechanical and electrical design of systems, interferes with sensors by limiting their capabilities, heavily impacts on data transmissions, and generally requires systems with low power consumption in order to enable reasonable mission duration. Interest in underwater applications is driven by needs of exploring and intervening in environments in which human capabilities are very limited. Nowadays, most underwater field operations are carried out by manned or remotely operated vehicles, deployed for explorations and limited intervention missions. Manned vehicles, directly on-board controlled, expose human operators to risks related to the stay in field of the mission, within a hostile environment. Remotely Operated Vehicles (ROV) currently represent the most advanced technology for underwater intervention services available on the market. These vehicles can be remotely operated for long time but they need support from an oceanographic vessel with multiple teams of highly specialized pilots. Vehicles equipped with multiple state-of-art sensors and capable to autonomously plan missions have been deployed in the last ten years and exploited as observers for underwater fauna, seabed, ship wrecks, and so on. On the other hand, underwater operations like object recovery and equipment maintenance are still challenging tasks to be conducted without human supervision since they require object perception and localization with much higher accuracy and robustness, to a degree seldom available in Autonomous Underwater Vehicles (AUV). This thesis reports the study, from design to deployment and evaluation, of a general purpose and configurable platform dedicated to stereo-vision perception in underwater environments. Several aspects related to the peculiar environment characteristics have been taken into account during all stages of system design and evaluation: depth of operation and light conditions, together with water turbidity and external weather, heavily impact on perception capabilities. The vision platform proposed in this work is a modular system comprising off-the-shelf components for both the imaging sensors and the computational unit, linked by a high performance ethernet network bus. The adopted design philosophy aims at achieving high flexibility in terms of feasible perception applications, that should not be as limited as in case of a special-purpose and dedicated hardware. Flexibility is required by the variability of underwater environments, with water conditions ranging from clear to turbid, light backscattering varying with daylight and depth, strong color distortion, and other environmental factors. Furthermore, the proposed modular design ensures an easier maintenance and update of the system over time. Performance of the proposed system, in terms of perception capabilities, has been evaluated in several underwater contexts taking advantage of the opportunity offered by the MARIS national project. Design issues like energy power consumption, heat dissipation and network capabilities have been evaluated in different scenarios. Finally, real-world experiments, conducted in multiple and variable underwater contexts, including open sea waters, have led to the collection of several datasets that have been publicly released to the scientific community. The vision system has been integrated in a state of the art AUV equipped with a robotic arm and gripper, and has been exploited in the robot control loop to successfully perform underwater grasping operations.

Valutazione del rischio ecologico e sviluppo di strategie innovative per il monitoraggio della qualita’ dei corpi idrici: applicazione di un nuovo approccio integrato in un bacino pilota.

Freshwater is extremely precious; but even more precious than freshwater is clean freshwater. From the time that 2/3 of our planet is covered in water, we have contaminated our globe with chemicals that have been used by industrial activities over the last century in a unprecedented way causing harm to humans and wildlife. We have to adopt a new scientific mindset in order to face this problem so to protect this important resource. The Water Framework Directive (European Parliament and the Council, 2000) is a milestone legislative document that transformed the way that water quality monitoring is undertaken across all Member States by introducing the Ecological and Chemical Status. A “good or higher” Ecological Status is expected to be achieved for all waterbodies in Europe by 2015. Yet, most of the European waterbodies, which are determined to be at risk, or of moderate to bad quality, further information will be required so that adequate remediation strategies can be implemented. To date, water quality evaluation is based on five biological components (phytoplankton, macrophytes and benthic algae, macroinvertebrates and fishes) and various hydromorphological and physicochemical elements. The evaluation of the chemical status is principally based on 33 priority substances and on 12 xenobiotics, considered as dangerous for the environment. This approach takes into account only a part of the numerous xenobiotics that can be present in surface waters and could not evidence all the possible causes of ecotoxicological stress that can act in a water section. The mixtures of toxic chemicals may constitute an ecological risk not predictable on the basis of the single component concentration. To improve water quality, sources of contamination and causes of ecological alterations need to be identified. On the other hand, the analysis of the community structure, which is the result of multiple processes, including hydrological constrains and physico-chemical stress, give back only a “photograph” of the actual status of a site without revealing causes and sources of the perturbation. A multidisciplinary approach, able to integrate the information obtained by different methods, such as community structure analysis and eco-genotoxicological studies, could help overcome some of the difficulties in properly identifying the different causes of stress in risk assessment. In synthesis, the river ecological status is the result of a combination of multiple pressures that, for management purposes and quality improvement, have to be disentangled from each other. To reduce actual uncertainty in risk assessment, methods that establish quantitative links between levels of contamination and community alterations are needed. The analysis of macrobenthic invertebrate community structure has been widely used to identify sites subjected to perturbation. Trait-based descriptors of community structure constitute a useful method in ecological risk assessment. The diagnostic capacity of freshwater biomonitoring could be improved by chronic sublethal toxicity testing of water and sediment samples. Requiring an exposure time that covers most of the species’ life cycle, chronic toxicity tests are able to reveal negative effects on life-history traits at contaminant concentrations well below the acute toxicity level. Furthermore, the responses of high-level endpoints (growth, fecundity, mortality) can be integrated in order to evaluate the impact on population’s dynamics, a highly relevant endpoint from the ecological point of view. To gain more accurate information about potential causes and consequences of environmental contamination, the evaluation of adverse effects at physiological, biochemical and genetic level is also needed. The use of different biomarkers and toxicity tests can give information about the sub-lethal and toxic load of environmental compartments. Biomarkers give essential information about the exposure to toxicants, such as endocrine disruptor compounds and genotoxic substances whose negative effects cannot be evidenced by using only high-level toxicological endpoints. The increasing presence of genotoxic pollutants in the environment has caused concern regarding the potential harmful effects of xenobiotics on human health, and interest on the development of new and more sensitive methods for the assessment of mutagenic and cancerogenic risk. Within the WFD, biomarkers and bioassays are regarded as important tools to gain lines of evidence for cause-effect relationship in ecological quality assessment. Despite the scientific community clearly addresses the advantages and necessity of an ecotoxicological approach within the ecological quality assessment, a recent review reports that, more than one decade after the publication of the WFD, only few studies have attempted to integrate ecological water status assessment and biological methods (namely biomarkers or bioassays). None of the fifteen reviewed studies included both biomarkers and bioassays. The integrated approach developed in this PhD Thesis comprises a set of laboratory bioassays (Daphnia magna acute and chronic toxicity tests, Comet Assay and FPG-Comet) newly-developed, modified tacking a cue from standardized existing protocols or applied for freshwater quality testing (ecotoxicological, genotoxicological and toxicogenomic assays), coupled with field investigations on macrobenthic community structures (SPEAR and EBI indexes). Together with the development of new bioassays with Daphnia magna, the feasibility of eco-genotoxicological testing of freshwater and sediment quality with Heterocypris incongruens was evaluated (Comet Assay and a protocol for chronic toxicity). However, the Comet Assay, although standardized, was not applied to freshwater samples due to the lack of sensitivity of this species observed after 24h of exposure to relatively high (and not environmentally relevant) concentrations of reference genotoxicants. Furthermore, this species demonstrated to be unsuitable also for chronic toxicity testing due to the difficult evaluation of fecundity as sub-lethal endpoint of exposure and complications due to its biology and behaviour. The study was applied to a pilot hydrographic sub-Basin, by selecting section subjected to different levels of anthropogenic pressure: this allowed us to establish the reference conditions, to select the most significant endpoints and to evaluate the coherence of the responses of the different lines of evidence (alteration of community structure, eco-genotoxicological responses, alteration of gene expression profiles) and, finally, the diagnostic capacity of the monitoring strategy. Significant correlations were found between the genotoxicological parameter Tail Intensity % (TI%) and macrobenthic community descriptors SPEAR (p<0.001) and EBI (p<0.05), between the genotoxicological parameter describing DNA oxidative stress (ΔTI%) and mean levels of nitrates (p<0.01) and between reproductive impairment (Failed Development % from D. magna chronic bioassays) and TI% (p<0.001) as well as EBI (p<0.001). While correlation among parameters demonstrates a general coherence in the response to increasing impacts, the concomitant ability of each single endpoint to be responsive to specific sources of stress is at the basis of the diagnostic capacity of the integrated approach as demonstrated by stations presenting a mismatch among the different lines of evidence. The chosen set of bioassays, as well as the selected endpoints, are not providing redundant indications on the water quality status but, on the contrary, are contributing with complementary pieces of information about the several stressors that insist simultaneously on a waterbody section providing this monitoring strategy with a solid diagnostic capacity. Our approach should provide opportunities for the integration of biological effects into monitoring programmes for surface water, especially in investigative monitoring. Moreover, it should provide a more realistic assessment of impact and exposure of aquatic organisms to contaminants. Finally this approach should provide an evaluation of drivers of change in biodiversity and its causalities on ecosystem function/services provision, that is the direct and indirect contributions to human well-being.