Coral reefs are the most biodiverse marine ecosystems, and host a wide range of taxonomic diversity in a complex spatial habitat structure. Existing coral reef survey methods struggle to accurately capture the taxonomic detail within the complex spatial structure of benthic communities. We propose a workflow to leverage underwater hyperspectral transects and two machine learning algorithms to produce dense habitat maps of 1150 m 2 of reefs across the Curaçao coastline. Our multi-method workflow labelled all 500+ million pixels with one of 43 classes at taxonomic family, genus or species level for corals, algae, sponges, or to substrate labels such as sediment, turf algae and cyanobacterial mats. With low annotation effort (2% pixels) and no external data, our workflow enables accurate (Fbeta 87%) survey-scale mapping, with unprecedented thematic and spatial detail. Our assessments of the composition and configuration of the benthic communities of 23 transect showed high consistency. Digitizing the reef habitat structure enables validation and novel analysis of pattern and scale in coral reef ecology. Our dense habitat maps reveal the inadequacies of point sampling methods to accurately describe reef benthic communities.

Mangrove forests provide valuable ecosystem services to coastal communities across tropical and subtropical regions. Current anthropogenic stressors threaten these ecosystems and urge researchers to create improved monitoring methods for better environmental management. Recent efforts that have focused on automatically quantifying the above-ground biomass using image analysis have found some success on high resolution imagery of mangrove forests that have sparse vegetation. In this study, we focus on stands of mangrove forests with dense vegetation consisting of the endemic Pelliciera rhizophorae and the more widespread Rhizophora mangle mangrove species located in the remote Utría National Park in the Colombian Pacific coast. Our developed workflow used consumer-grade Unoccupied Aerial System (UAS) imagery of the mangrove forests, from which large orthophoto mosaics and digital surface models are built. We apply convolutional neural networks (CNNs) for instance segmentation to accurately delineate (33% instance average precision) individual tree canopies for the Pelliciera rhizophorae species. We also apply CNNs for semantic segmentation to accurately identify (97% precision and 87% recall) the area coverage of the Rhizophora mangle mangrove tree species as well as the area coverage of surrounding mud and water land-cover classes. We provide a novel algorithm for merging predicted instance segmentation tiles of trees to recover tree shapes and sizes in overlapping border regions of tiles. Using the automatically segmented ground areas we interpolate their height from the digital surface model to generate a digital elevation model, significantly reducing the effort for ground pixel selection. Finally, we calculate a canopy height model from the digital surface and elevation models and combine it with the inventory of Pelliciera rhizophorae trees to derive the height of each individual mangrove tree. The resulting inventory of a mangrove forest, with individual P. rhizophorae tree height information, as well as crown shape and size descriptions, enables the use of allometric equations to calculate important monitoring metrics, such as above-ground biomass and carbon stocks.

Metropolitan areas around the world are experiencing a surge in air pollution levels due to different anthropogenic causes, making accurate air quality prediction a critical task for public health. Although many prediction systems have been researched and modelled, many of them have neglected the different effects that air pollution has on each individual citizen. Hence, we present a novel context prediction model that includes context-aware computing concepts to merge an accurate air pollution prediction algorithm (using Long Short-Term Memory Deep Neural Network) with information from both surrounding pollution sources (e.g., bushfire incidents, traffic volumes) and user’s health profile. This model is then integrated into a tool called My Air Quality Index (MyAQI), which is further implemented and evaluated in a real-life use case set up in Melbourne Urban Area (Victoria, Australia). Results obtained with MyAQI show both that (i) high precision levels are reached (90–96%) when forecasting air quality situations in four air quality monitoring stations, and (ii) the proposed model is highly adaptable to users’ individual health condition effects under the same airborne pollutant levels.

Predicting the time and place where concentrations of pollutants will be the highest is critical for air quality monitoring- and early-warning systems in urban areas. Much of the research effort in this area is focused only on improving air pollution prediction algorithms, disregarding valuable environmental- and user-based context. In this paper we apply context-aware computing concepts in the MyAQI system, to develop an integral air quality monitoring and prediction application, that shifts the focus towards the individual needs of each end-user, without neglecting the benefits of the latest air pollution forecasting algorithms. We design and describe a novel context and situation reasoning model, that considers external environmental context, along with user based attributes, to feed into the prediction model. We demonstrate the adaptability and customizability of the design and the accuracy of the prediction technique in the implementation of the responsive MyAQI web application. We test the implementation with different user profiles and show the results of the system's adaptation. We demonstrate the prediction model's accuracy, when using extended context for 4 air quality monitoring stations in the Melbourne Region in Victoria, Australia.

Air pollution is a growing global concern that affects the health and livelihood of millions of people worldwide. The advent of the Internet of Things (IoT) has made available a plethora of data sources that provide near real-time information on air pollution. Many studies and systems have taken advantage of data stemming from the IoT and have been dedicated to enhancing the monitoring and prediction of air quality, from a fairly analytical angle, often disregarding the user's perspective in processing and presenting this data. In this paper, we research and present a novel context-aware air quality monitoring and prediction system called My Air Quality Index (MyAQI). MyAQI takes into consideration user's context (e.g. health conditions, individual sensitivities and preferences) to tailor the visualisation and notifications. We propose a context model that is used to combine user's context with air pollution data to provide context-aware recommendations to the specific user. MyAQI also incorporates a prediction algorithm based on Long Short-Term Memory Neural Network (LSTM) to predict future air quality. MyAQI is implemented as a web-based application and has the capability to consume data from a wide range of data sources including IoT devices and open data sources (via Application Programming Interfaces (API)). We demonstrate the context-aware visualisation techniques implemented in MyAQI, which adapt to changing user's context, and validate the performance of the air quality prediction algorithm.