Research
Comparative Analysis of Machine Learning Algorithms for Soil Erosion Modelling Based on Remotely Sensed Data
2023
This study performs a comparative analysis of three commonly used classification algorithms—Support Vector Machine (SVM), Random Forest (RF), and Multilayer Perceptron (MLP)—in combination with ground truth samples from regions across Iceland. The process is automated to predict soil erosion risk for larger, less accessible areas using Sentinel-2 images. The analysis supports the effectiveness of these approaches for modeling soil erosion, highlighting their differences.
Remote Mapping of Soil Erosion Risk in Iceland
2022
This study explores the application of remote sensing techniques for assessing soil erosion risk in Iceland. Due to Arctic challenges like short growing seasons, strong winds, and cloud cover, ground surveys are difficult. The paper presents a Support Vector Machine (SVM) classification model, trained with ground truth data from Iceland’s Soil Conservation Services. The methodology allows automated analysis of large, inaccessible areas using Sentinel-2 satellite images.
DOI: https://doi.org/10.5194/isprs-archives-XLVIII-4-W1-2022-135-2022
Guided Ultrasonic Wave Beam Skew in Silicon Wafers
2018
In the photovoltaic industry, monocrystalline silicon wafers are widely used for high-efficiency solar cells. Micro-cracks formed during the cutting process can lead to brittle wafer fractures, requiring non-destructive testing for quality control. This study explores the application of guided ultrasonic waves for detecting micro-crack density. The research investigates the effects of material anisotropy on guided wave characteristics, using contact piezoelectric transducers and non-contact laser interferometry. Experimental and theoretical results confirm significant wave beam skew, particularly for the S0 mode, with implications for wafer inspection techniques.
Lamb Wave Propagation in Monocrystalline Silicon Wafers
2018
Monocrystalline silicon wafers are essential in the photovoltaic industry for producing high-efficiency solar panels. This study investigates the behavior of guided ultrasonic waves, specifically Lamb waves, for detecting micro-cracks in thin wafers. The research focuses on the effects of material anisotropy on wave propagation characteristics, including phase slowness and wave beam skewing. Experimental measurements using contact wedge transducers and laser interferometry showed strong agreement with theoretical models, with significant wave skewing observed, particularly for the S0 mode. Finite element simulations confirmed these effects, highlighting considerations for non-destructive testing of wafers.
High Frequency Guided Wave Propagation in Monocrystalline Silicon Wafers
2017
Monocrystalline silicon wafers are widely used in the photovoltaic industry for high-efficiency solar panels. The cutting process introduces micro-cracks that can affect wafer integrity. This study investigates the propagation of high-frequency guided ultrasonic waves for wafer structural monitoring. The impact of anisotropy on wave characteristics was analyzed through three-dimensional Finite Element simulations and experimental measurements. The results confirm strong directional dependency, with selective wave mode excitation achieved using contact piezoelectric transducers and laser interferometry. Good agreement was found between simulations, experimental data, and theoretical models, demonstrating the effectiveness of guided waves for non-destructive wafer inspection.
