Barend van der Merwe

email: [email protected]
Tel: +27 (0) 12 420 3699
Office: Geography Building 2-11
Monday: 10:30 - 12:30
Thursday: 14:30-16:30
Book a Consultation

Research profile

My research interests focus primarily on computational methodologies and how they can be sued to collect, analyse, and draw new insights from data. I am particularly interested in investigating the application of artificial intelligence (AI), computer vision, and physical computing to geographical and environmental problems. My PhD thesis answered the question of whether convolutional neural networks (CNNs) can learn to distinguish between barchan dune morphologies. Currently, I'm cosupervising a PhD Student (Christiaan Schutte) who is investigating the use of Recurrent Neural Networks (RNNs) as a tool to predict catchment discharge. I'm also supervising a MA student (Christopher Ruiters) on the use of the MobileNet architecture to classify plastic waste. My work on computer vision has largely been in developing applications to assist in teaching. Aside from automatically grading bubble sheets (i.e. multiple choice questions), I developed an OpenCV application to assist in the grading of map work by superimposing a model answer over the student's submission. Recently, I started investigating the potential of physical computing (i.e. the combination of microcontrollers, sensors, and actuators) as a means to successfully replicate research in environmental science and geomorphology.


  • BSc (Environmental Science), University of Pretoria
  • BSc Honours (Geography), University of Pretoria
  • MSc (Geography), University of Pretoria
    Dissertation: Comparing statistical inference and spatial analysis in geomorphology: a case study of Rietvlei Nature Reserve, Gauteng
  • PhD (Geography), University of Pretoria
    Thesis: Classifying barchan outlines into morphological classes using convolutional neural networks: a proof of concept

Courses presented in 2023

  • ENV 201 - Environmental Sciences 201
  • GGY 168 - Introduction to Physical Geography
  • GGY 718 - Applied Geomorphology

Conference Presentations/Webinars

  • Classifying barchan shape and asymmetry using convolutional neural networks, Oral Presentation, Society of South African Geographers and Southern African Association of Geomorphologists 2021 Joint Biennial Conference (6-8 September).
  • The Performance of Different CNN Architectures on Barchan Asymmetry Classification, IAG Regional Webinar (Africa), International Geomorphology Week 2021.
  • Barchan shape as an aid to process explanation: An historical overview, Society of South African Geographers (SSAG) centennial conference 2016, 25-28 September, Stellenbosch, South Africa
  • Using Geometric Morphometrics to Quantify Barchan Shape from Satellite Imagery, Oral Presentation, South African Association of Geomorphologists (SAAG) 2015 Conference, 19-20 September, Sani Mountain Lodge, Lesotho

Research publications (since 2012)

ORCID ID | Google Scholar | LinkedIN | Twitter | GitHub

  • van der Merwe, B., Pillay, N. and Coetzee, S. (2022). An application of CNN to classify barchan dunes into asymmetry classes, Aeolian Research, 56, 100801, doi: 10.1016/j.aeolia.2022.100801.
  • Van der Merwe, B.J. (2021). The relationship between barchan size and barchan morphology: A case study from Northern Namibia, South African Geographical Journal, 103, 119-138, doi: 10.1080/03736245.2021.1876753.
  • Van der Hoven C, Ubomba-Jaswa E, van der Merwe B, Loubser M, Abia, ALK (2017) The impact of various land uses on the microbial and physiochemical quality of surface water bodies in developing countries: Prioritisation of water resources management areas, Environmental Nanotechnology, Monitoring and Management, 8, 280-289.
  • Hansen C, Meiklejohn, K, Nel W, Loubser MJ, van der Merwe BJ (2013). Aspect-controlled weathering observed on a blockfield in Dronning Maud Land, Antarctica, Geografiska Annaler Series A: Physical Geography, 95, 305-313.
- Author Barend vd Merwe

Copyright © University of Pretoria 2023. All rights reserved.

COVID-19 Corona Virus South African Resource Portal

To contact the University during the COVID-19 lockdown, please send an email to [email protected]

FAQ's Email Us Virtual Campus Share Cookie Preferences