There are many studies that point out that the climate change and the global warming are real threats to the humankind. To understand the effects of those phenomena’s and take contra measures we need sometimes to down to the earth and analyse ecosystems dynamics. In this study, you will have the chance to work with competitive remote sensing technologies and contribute to this important large-scale work that will bring benefits for all of us.
Hemispherical Photography (HP) is a technique that makes use of fisheye (wide- angle) lenses to measure canopy Plant Area Index (PAI). This index is traditionally used by foresters and ecologists to characterize ecosystems. The images acquired are manually processed through a computer vision technique denoted as thresholding to separate the sky from the canopy. In this way, the canopy volume can be computed. A major challenge in this technique it’s to have a procedure that computes automatically the canopy for different HP images and workspace conditions.
The goal of this work is to use computer visions and machine learning approaches for designing an automatic thresholding system for a fisheye camera and canopy workspaces. This thesis will be carry out in the follow steps: a) Review of the state-of-the art of canopy calculation with HP; b) Design an automatic threshold approach; c) Field experiments
- Pueschel, P., Buddenbaum, H., & Hill, J. (2012). An efficient approach to standardizing the processing of hemispherical images for the estimation of forest structural attributes. Agricultural and Forest Meteorology, 160, 1–13.
- Jonckheere I, Nackaerts K, Muys B, Coppin P. 2005. Assessment of automatic gap fraction estimation of forests from digital hemispherical photography. Agric. For. Meteorol. 132(1–2): 96-114
- Nobis M, Hunziker U. 2005. Automatic thresholding for hemispherical canopy-photographs based on edge detection. Agric. For. Meteorol. 128(3–4): 243-250
- Willingness to change the world
- Artificial intelligence enthusiast
- Excited to work with computer vision
Theme(s): Sensing & measuring, Modelling & visualisation