you're reading...


What do remote sensing, machine learning, and statistics have in common? Enhancing the accuracy of seagrass monitoring, for one.

Citation: Ha NT, Manley-Harris M, Pham TD, Hawes I. A Comparative Assessment of Ensemble-Based Machine Learning and Maximum Likelihood Methods for Mapping Seagrass Using Sentinel-2 Imagery in Tauranga Harbor, New Zealand. Remote Sensing [Internet]. MDPI AG; 2020 Jan 21;12(3):355. Available from: http://dx.doi.org/10.3390/rs12030355


What do remote sensing, machine learning, and statistics have in common?  Enhancing the accuracy of seagrass monitoring, for one.

Seagrass meadows are highly productive blue carbon ecosystems found in shallow, salty and brackish waters across the world, from the tropics to the Arctic Circle. Seagrasses belong to a group of plants known as monocotyledons, which include lilies, grasses, and palms; like their relatives, seagrasses have roots, leaves, and veins, and produce seeds and flowers. Now if you weren’t already amazed, here is the even more amazing part: they evolved around 100 million years ago and yet remain the most understudied of ecosystems, with their total geographic extent largely unknown due to challenges in mapping, globally (Hastings et al. 2020). Seagrass can form dense underwater meadows, many of which are large enough to be seen from space. Despite being one of the most productive ecosystems in the world, seagrasses do not receive enough research, funding, or attention.


Seagrasses have been acknowledged as one of the most productive ‘blue carbon’ (the potential for carbon to be stored in coastal and marine sediments) ecosystems, and are in significant decline across most of the globe. One of the first steps towards conservation is to map and monitor existing seagrass meadows. A number of different methods are used for this task, including statistical modeling and geospatial analyses. But what about using the power of machine learning to convert satellite images into seagrass meadow maps? This technology has been used in other conservation efforts, but is still uncommon in seagrass work. In a recent study, authors explore the possibility of bringing machine learning into the world of seagrass conservation.

Therefore, Dr. Ha and team aimed to develop a novel approach for seagrass monitoring, using state-of-the-art machine learning, along with data from Sentinel-2 imagery. Sentinel-2 is a constellation of two identical satellites in the same orbit, imaging land and coastal areas at high spatial resolutions. Tauranga Harbor, New Zealand served as a validation site for extensive ground-truthing data to compare methods. The team developed ensemble machine learning methods including random forest (RF), rotation forests (RoF), and canonical correlation forests (CCF) with the more traditional maximum likelihood classifier (MLC) technique. Using a group of validation metrics, their results indicated that their machine learning techniques outperformed the MLC, and RoF was the best performer. 


Now, you have probably lost me at this point: so, let’s break these terms apart to understand what Dr. Ha really did.


Canonical Corre-what?

In recent years, machine learning (ML) has emerged as a novel approach for seagrass mapping and monitoring. Machine learning has the benefits of rapid learning, working with non-linear data, and an increasing number of new, open source algorithms. In the field of seagrass mapping and monitoring, machine learning is still in its infancy. All of the statistical and machine learning applications used until now have included mixed results, but they support the exploration of new machine learning approaches, particularly for improving low coverage seagrass mapping.

Among the various machine learning algorithms, rotation forest (RoF) and canonical correlation forest (CCF) algorithms are now emerging as reliable techniques for land cover mapping, landslide mapping using multi-spectral or hyper-spectral imagery, and rapid building mapping using multi-source data. These machine learning algorithms are well-known for detecting boundaries. These techniques potentially offer benefits in the classification of low coverage through enhanced recognition of edge boundaries. Therefore, their goal in this study was to compare the use of three ML algorithms to the more traditional machine learning approaches for mapping the aboveground distribution of seagrass communities at low and high coverage using satellite imagery data.

The team’s target was Tauranga Harbor, New Zealand, as it offers a mosaic of dense, sparse, and zero seagrass coverage. The team discussed the difference in the performance of the selected models for seagrass detection at two densities in this paper. Their hope was that their results would contribute alternative solutions for the mapping and monitoring of seagrass at various regions in the world, and assist in the conservation of this important blue carbon ecosystem.

So what does it all mean?

Well, first, in this study the machine learning methods outperformed the statistical methods for all evaluation metrics.

Until now, no literature had looked at the comparative performance of the machine learning classifier methods for seagrass mapping, along with a full radiometric correction (calibrating the image pixel values and correcting for error) of the image. Additionally, of the machine learning ensemble approaches that Dr. Ha’s team used the RoF model to demonstrate superior performance than that of CCF and RF; this is a unique result because previous studies demonstrate superior performance of CCF. Of the methods tested here, only the RF technique has previously been applied to seagrass mapping using very high spatial resolution imagery. In that case, high precision (0.947) and recall (0.968) values were determined mapping Posidonia oceanica from digital airborne images, though no comparison to other methods was attempted. In another seagrass study, the overall accuracy only reached 82% using the RF algorithm applied to RapidEye imagery. Considering the size of the seagrass meadows and the mix of substrate in Tauranga Harbor, the measured scores in Dr. Ha’s team’s results were reliable for both dense and sparse seagrass mapping using medium spatial resolution of Sentinel-2 data (10 m pixel size), which is a really interesting result!

Dr. Ha’s results attest to the reliable application of the RoF model for the mapping and monitoring of seagrass in shallow water using satellite imagery. Despite a lower accuracy for sparse than dense seagrass meadow classification, the CCF model shows potential for the mapping of seagrass and merits further testing at various scales and in various case studies. Regarding MLC, this model is still an applicable candidate for dense seagrass meadows, however, it may not be applicable for the mapping of sparse to very sparse seagrass meadows.

With the development of computer vision and pattern recognition, deep learning approaches using a variety of algorithms such as convolutional neural networks (CNNs or recurrent neural networks (RNNs)) for semantic segmented imagery applied sub-pixel techniques should be encouraged for future studies.


No comments yet.

Post a Comment


  • by oceanbites 3 months ago
    Happy Earth Day! Take some time today to do something for the planet and appreciate the ocean, which covers 71% of the Earth’s surface.  #EarthDay   #OceanAppreciation   #Oceanbites   #CoastalVibes   #CoastalRI 
  • by oceanbites 4 months ago
    Not all outdoor science is fieldwork. Some of the best days in the lab can be setting up experiments, especially when you get to do it outdoors. It’s an exciting mix of problem solving, precision, preparation, and teamwork. Here is
  • by oceanbites 5 months ago
    Being on a research cruise is a unique experience with the open water, 12-hour working shifts, and close quarters, but there are some familiar practices too. Here Diana is filtering seawater to gather chlorophyll for analysis, the same process on
  • by oceanbites 6 months ago
    This week for  #WriterWednesday  on  #oceanbites  we are featuring Hannah Collins  @hannahh_irene  Hannah works with marine suspension feeding bivalves and microplastics, investigating whether ingesting microplastics causes changes to the gut microbial community or gut tissues. She hopes to keep working
  • by oceanbites 6 months ago
    Leveling up - did you know that crabs have a larval phase? These are both porcelain crabs, but the one on the right is the earlier stage. It’s massive spine makes it both difficult to eat and quite conspicuous in
  • by oceanbites 7 months ago
    This week for  #WriterWednesday  on  #Oceanbites  we are featuring Cierra Braga. Cierra works ultraviolet c (UVC) to discover how this light can be used to combat biofouling, or the growth of living things, on the hulls of ships. Here, you
  • by oceanbites 7 months ago
    This week for  #WriterWednesday  at  #Oceanbites  we are featuring Elena Gadoutsis  @haysailor  These photos feature her “favorite marine research so far: From surveying tropical coral reefs, photographing dolphins and whales, and growing my own algae to expose it to different
  • by oceanbites 7 months ago
    This week for  #WriterWednesday  on Oceanbites we are featuring Eliza Oldach. According to Ellie, “I study coastal communities, and try to understand the policies and decisions and interactions and adaptations that communities use to navigate an ever-changing world. Most of
  • by oceanbites 8 months ago
    This week for  #WriterWednesday  at  #Oceanbites  we are featuring Jiwoon Park with a little photographic help from Ryan Tabata at the University of Hawaii. When asked about her research, Jiwoon wrote “Just like we need vitamins and minerals to stay
  • by oceanbites 8 months ago
    This week for  #WriterWednesday  on  #Oceanbites  we are featuring  @riley_henning  According to Riley, ”I am interested in studying small things that make a big impact in the ocean. Right now for my master's research at the University of San Diego,
  • by oceanbites 8 months ago
    This week for  #WriterWednesday  at  #Oceanbites  we are featuring Gabby Stedman. Gabby is interested in interested in understanding how many species of small-bodied animals there are in the deep-sea and where they live so we can better protect them from
  • by oceanbites 8 months ago
    This week for  #WriterWednesday  at  #Oceanbites  we are featuring Shawn Wang! Shawn is “an oceanographer that studies ocean conditions of the past. I use everything from microfossils to complex computer models to understand how climate has changed in the past
  • by oceanbites 9 months ago
    Today we are highlighting some of our awesome new authors for  #WriterWednesday  Today we have Daniel Speer! He says, “I am driven to investigate the interface of biology, chemistry, and physics, asking questions about how organisms or biological systems respond
  • by oceanbites 9 months ago
    Here at Oceanbites we love long-term datasets. So much happens in the ocean that sometimes it can be hard to tell if a trend is a part of a natural cycle or actually an anomaly, but as we gather more
  • by oceanbites 10 months ago
    Have you ever seen a lobster molt? Because lobsters have exoskeletons, every time they grow they have to climb out of their old shell, leaving them soft and vulnerable for a few days until their new shell hardens. Young, small
  • by oceanbites 10 months ago
    A lot of zooplankton are translucent, making it much easier to hide from predators. This juvenile mantis shrimp was almost impossible to spot floating in the water, but under a dissecting scope it’s features really come into view. See the
  • by oceanbites 11 months ago
    This is a clump of Dead Man’s Fingers, scientific name Codium fragile. It’s native to the Pacific Ocean and is invasive where I found it on the east coast of the US. It’s a bit velvety, and the coolest thing
  • by oceanbites 11 months ago
    You’ve probably heard of jellyfish, but have you heard of salps? These gelatinous sea creatures band together to form long chains, but they can also fall apart and will wash up onshore like tiny gemstones that squish. Have you seen
  • by oceanbites 12 months ago
    Check out what’s happening on a cool summer research cruise! On the  #neslter  summer transect cruise, we deployed a tow sled called the In Situ Icthyoplankton Imaging System. This can take pictures of gelatinous zooplankton (like jellyfish) that would be
  • by oceanbites 1 year ago
    Did you know horseshoe crabs have more than just two eyes? In these juveniles you can see another set in the middle of the shell. Check out our website to learn about some awesome horseshoe crab research.  #oceanbites   #plankton   #horseshoecrabs 
WP2Social Auto Publish Powered By : XYZScripts.com