TDTony Dolphin

Tony Dolphin

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Tony Dolphin has a total of 32 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Crowdsourcing Experiment and Fully Convolutional Neural Networks for Coastal Remote Sensing of Seagrass and Macroalgae

    by Brandon Hobley, Michal Mackiewicz, Julie Bremner, Tony Dolphin, Riccardo Arosio

    Recently, convolutional neural networks (CNNs) and fully convolutional neural networks (FCNs) have been successfully used for monitoring coastal marine ecosystems, in particular vegetation. However, even with recent advances in computational modelling and data acquisition, deep learning models require substantial amounts of good quality reference data to effectively self-learn internal representations of input imagery. The classical approach for coastal mapping requires experts to transcribe insitu records and delineate polygons from high-resolution imagery such that FCNs can self-learn. However, labelling by a single individual limits the training data, whereas crowdsourcing labels can increase the volume of training data, but may compromise label quality and consistency. In this paper we assessed the reliability of crowdsourced labels on a complex multi-class problem domain for estuarine vegetation and unvegetated sediment. An inter-observer variability experiment was conducted in order to assess the statistical differences in crowdsourced annotations for plant species and sediment. The participants were grouped based on their discipline and level of expertise, and the statistical differences were evaluated using the Cochran's Q-test and the annotation accuracy of each group to determine for observation biases. Given the crowdsourced labels, FCNs were trained with majority-vote annotations from each group to check whether observation biases were propagated to FCN performance. Two scenarios were examined: first, a direct comparison of FCNs trained with transcribed in-situ labels and crowdsourced labels from each group was established. Then, transcribed in-situ labels were supplemented with crowdsourced labels to investigate the feasibility of training FCNs with crowdsourced labels in coastal mapping applications. We show that annotations sourced from discipline experts (ecologists and geomorphologists) familiar with the study site were more accurate than experts with no prior knowledge of the site and non-experts, with our results confirming that biases in participant annotation were propagated in FCN performance. Furthermore, FCNs trained with a combined dataset of in-situ and crowdsourced labels performed better than FCNs trained on the same imagery with in-situ labels.


  • Open Access

    Semi-Supervised Segmentation for Coastal Monitoring Seagrass Using RPA Imagery

    by Brandon Hobley, Riccardo Arosio, Geoffrey French, Julie Bremner, Tony Dolphin, Michal Mackiewicz

    Intertidal seagrass plays a vital role in estimating the overall health and dynamics of coastal environments due to its interaction with tidal changes. However, most seagrass habitats around the globe have been in steady decline due to human impacts, disturbing the already delicate balance in the environmental conditions that sustain seagrass. Miniaturization of multi-spectral sensors has facilitated very high resolution mapping of seagrass meadows, which significantly improves the potential for ecologists to monitor changes. In this study, two analytical approaches used for classifying intertidal seagrass habitats are compared—Object-based Image Analysis (OBIA) and Fully Convolutional Neural Networks (FCNNs). Both methods produce pixel-wise classifications in order to create segmented maps. FCNNs are an emerging set of algorithms within Deep Learning. Conversely, OBIA has been a prominent solution within this field, with many studies leveraging in-situ data and multiresolution segmentation to create habitat maps. This work demonstrates the utility of FCNNs in a semi-supervised setting to map seagrass and other coastal features from an optical drone survey conducted at Budle Bay, Northumberland, England. Semi-supervision is also an emerging field within Deep Learning that has practical benefits of achieving state of the art results using only subsets of labelled data. This is especially beneficial for remote sensing applications where in-situ data is an expensive commodity. For our results, we show that FCNNs have comparable performance with the standard OBIA method used by ecologists.


  • Open Access

    Semi-supervised segmentation for coastal monitoring seagrass using RPA imagery

    by Brandon Lewis Hobley, Riccardo Arosio, Geoffrey French, Julie Bremner, Tony Dolphin, Michal Mackiewicz

    Intertidal seagrass plays a vital role in estimating the overall health and dynamics of coastal environments due to its interaction with tidal changes. However, most seagrass habitats around the globe have been in steady decline due to human impacts, disturbing the already delicate balance in environmental conditions that sustain seagrass. Miniaturization of multi-spectral sensors has facilitated very high resolution mapping of seagrass meadows, which significantly improve the potential for ecologists to monitor changes. In this study, two analytical approaches used for classifying intertidal seagrass habitats are compared: Object-based Image Analysis (OBIA) and Fully Convolutional Neural Networks (FCNNs). Both methods produce pixel-wise classifications in order to create segmented maps, however FCNNs are an emerging set of algorithms within Deep Learning with sparse application towards seagrass mapping. Conversely, OBIA has been a prominent solution within this field, with many studies leveraging in-situ data and multiscale segmentation to create habitat maps. This work demonstrates the utility of FCNNs in a semi-supervised setting to map seagrass and other coastal features from an optical drone survey conducted at Budle Bay, Northumberland, England. Semi-supervision is also an emerging field within Deep Learning that has practical benefits of achieving state of the art results using only subsets of labelled data. This is especially beneficial for remote sensing applications where in-situ data is an expensive commodity. For our results, we show that FCNNs have comparable performance with standard OBIA method used by ecologists, while also noting an increase in performance for mapping ecological features that are sparsely labelled across the study site.


  • Mixed Sand/Gravel Beach Response to the Beast from the East Storms

    by Tony Dolphin, Helene Burningham, Dafni Sifnioti, William Manning, Steven Wallbridge, Adrian Farcas

    Dolphin, T.; Burningham, H.; Sifnioti, D.; Manning, W.; Wallbridge, S., and Farcas, A., 2020. Mixed sand/gravel beach response to the Beast from the East storms. In: Malvárez, G. and Navas, F. (eds.), Global Coastal Issues of 2020. Journal of Coastal Research, Special Issue No. 95, pp. 463–467. Coconut Creek (Florida), ISSN 0749-0208.Two large, very long-duration, shore-normal (easterly) storms worked on the Minsmere – Sizewell mixed sand and gravel (MSG) beach frontage (Suffolk, UK) in March 2018. These storms were unusual in their sustained high waves (Hs > 2.1 m, 97.5th percentile) for 140 hours from the east (16% occurrence), compared to the regional median duration of 8.5 hours per storm. Patterns in contour advance/retreat were compared with modelled nearshore wave conditions every 100 m along the 3.3 km frontage. Although there was a positive correlation between the upper supra-tidal beach and wave period, patterns in the beach response were inconsistent and zones of common behaviour could not be explained by the wave data. Postulated explanations include spatial variation in the antecedent morphology and particle-size arrangement within the MSG sediment matrix. Despite the high magnitude of work done by these storms, the lack of barrier erosion suggests that barrier stability is likely for the present sea level under these perceived more-erosive, though less common, easterly storms.

  • Velocity of RFID Tagged Gravel in a Non-uniform Longshore Transport System

    by Tony Dolphin, Janette Lee, Roger Phillips, Colin J.L. Taylor, Keith R. Dyer

    Dolphin, T.; Lee, J.; Phillips, R.; Taylor, C.J.L., and Dyer, K., 2016. Velocity of RFID tagged gravel in a nonuniform longshore transport system. In: Vila-Concejo, A.; Bruce, E.; Kennedy, D.M., and McCarroll, R.J. (eds.), Proceedings of the 14th International Coastal Symposium (Sydney, Australia). Journal of Coastal Research, Special Issue, No. 75, pp. 363–367. Coconut Creek (Florida), ISSN 0749-0208. Radio-frequency Identification (RFID) tags were embedded into 940 native beach gravel particles and deployed at source on Benhole Beach (Somerset, UK) to develop an understanding of gravel transport along a frontage that is likely to be subjected to construction activity associated with the proposed development of the Hinkley Point C power station. The study lasted for three years, the longest of its kind to be published, and maintained high detection rates (78% on completion) typical of RFID experiments. Tracer distribution through time highlighted non-uniformity in the longshore transport system ...


  • Use of GPS and X-Band Radar to track tidal currents around the shore parallel breakwaters at Sea Palling, Norfolk, UK.

    by Roger Phillips, Chris Vincent, Paul Bell, Tony Dolphin, John Bacon

    Phillips, R., Vincent, C.E., Bell, P.S., Dolphin, T., and Bacon, J., 2013. Use of GPS and X-Band Radar to track tidal currents around the shore parallel breakwaters at Sea Palling, Norfolk, UK Sea Palling on the East coast of the UK has a series of 9 shore parallel rock breakwaters. The numerical model TELEMAC2D has been used to simulate tidal currents around the system to assess the tidal contribution to sediment budgets and transport pathways. Numerical simulations are particularly useful for identifying the important processes involved in a complex system such as these breakwaters and understanding how the breakwaters might effect the local coastal region. Models are classically evaluated using a number of measurement stations to calibrate and then check it is giving realistic results. It is important to assess both the spatial and temporal performance of a model, but this is difficult with a model having tens of thousands of nodes, based on only a few discrete locations. This paper presents a...


  • Open Access

    Far-field vessel wakes in Tallinn Bay; 273–302

    by Kevin Parnell, Nicole Delpeche, Ira Didenkulova, Tony Dolphin, Ants Erm, Andres Kask, Loreta Kelpšaite, Dmitry Kurennoy, Ewald Quak, Andrus Räämet, et al.

    The properties of wave fields induced by high-speed ferries and recently introduced conventional ferries with increased cruise speeds are analysed for a site in Tallinn Bay, the Gulf of Finland, the Baltic Sea, located about 3 km from the sailing line and up to 8 km from the wave production area. The analysis is based on high-resolution profiling of the water surface for about 650 wakes from fast ferries, measured during 4 weeks in June–July 2008. The new large conventional ferries with cruise speeds of 25–30 knots (~ 45–55 km/h) sail at near-critical speeds along extensive sections of eastern Tallinn Bay, and excite wakes equivalent to those of high-speed ferries. The peak periods of these wakes are between 10 and 13 s. The typical daily highest ship wave is approximately 1.2 m, measured prior to wake breaking. The largest recorded ship wave in calm conditions had a height of 1.5 m and in the presence of some wind wave background 1.7 m. The cumulative impact of ship wakes results in a gradual increase in the suspended matter concentration in near-bottom water over the course of a day. The largest and longest ship waves produce considerable wave runup at the coast and prevent several coastal sections from achieving an equilibrium state. The largest ship waves have an asymmetric shape both in terms of the water surface elevation above and below the mean level and in terms of the shape of the wave front and back. The overall intensity of anthropogenic waves has remained at the same level as it was in the year 2002, although the ships that produced the highest waves in the past are no longer in service.