As part of my own research and through collaboration with colleagues from the Faculty of Mining Surveying and Environmental Engineering, I was able to form a research team. Below is a brief description of the research topics that our team focuses on along with a summary of our ongoing and past research activities.
Members
PhD Paweł Ćwiąkała
Responsible for conducting research using unmanned aerial vehicles (UAVs), particularly the optimization and development of UAV missions to assess the accuracy of on-board navigation systems.
PhD Wojciech Gruszczyński
Responsible for developing methodology to determine the height of the terrain surface and vertical displacement using UAV-derived photogrammetric data.
PhD Edyta Puniach
Responsible for developing methodology to determine horizontal displacement using UAV-derived photogrammetric data.
PhD Wojciech Matwij
Responsible for developing methodology to determine the displacement field using terrestrial laser scanning.
Msc Paweł Wiącek
Responsible for testing and calibrating optical sensors and developing a comprehensive validation procedure for UAV-derived photogrammetric products.
Research topics
The research team focuses on the acquisition, development, evaluation, and processing of high-resolution data from UAV photogrammetry and terrestrial laser scanning with the goal of determining the displacement field of the terrain surface and the objects on it. The solutions developed by the team make use of a wide range of computational tools based on computer vision, including image registration algorithms built on autocorrelation and keypoint descriptors, algorithms for processing and registering point clouds, and artificial neural networks.
Current projects
The “Jaworzno” project (2020–2021) aims to extend the scope of the team’s research by determining the displacement field of the terrain surface and objects on it using both UAV photogrammetric data and terrestrial laser scanning data. Under this project, the team is also verifying the developed algorithms to determine the heights of the terrain surface points. Quarterly measurements are planned in areas subject to land deformation due to underground coal mining. This project is being conducted in collaboration with the team of Professor Zygmunt Niedojadło, which specializes in observing, forecasting, and modeling land deformation in areas subjected to underground mining operations.
Part of the results of work to date in this project is described in the articles:
Determination of underground mining-induced displacement field using multi-temporal TLS point cloud registration. DOI: https://doi.org/10.1016/j.measurement.2021.109482
Application of UAV-based orthomosaics for determination of horizontal displacement caused by underground mining. DOI: 10.1016/j.isprsjprs.2021.02.006
Correction of Low Vegetation Impact on UAV-Derived Point Cloud Heights With U-Net Networks. DOI: https://doi.org/10.1109/TGRS.2021.3057272
UAV Applications for Determination of Land Deformations Caused by Underground Mining. DOI: https://doi.org/10.3390/rs12111733.
Further articles based on the results obtained in this project are under development.
Finished projects
The “Laziska” pilot project (2015–2016) compared the accuracy of determining relief geometry using UAVs and terrestrial laser scanning. Based on the acquired data, a method for vegetation filtration was proposed, leading to a reduction in the error in the determined height of the terrain surface of up to 40% compared to the raw data. The proposed algorithm is based on the search for local minima in the point cloud.
The results of this project are described in the article entitled:
“Comparison of low-altitude UAV photogrammetry with terrestrial laser scanning as data-source methods for terrain covered in low vegetation”. DOI: https://doi.org/10.1016/j.isprsjprs.2017.02.015
The “Jerzmanowice” project (2018–2019) aimed at developing filtration methodology to minimize the effect of vegetation on the geometry of the land surface determined using UAV-collected photogrammetric data. The research team collected UAV data at different times of the year and for different vegetation phases. Reference measurements were also carried out. The influence of vegetation on the determined heights of the land surface was minimized using convolutional neural network-based algorithms derived from the data collected during the “Łaziska” and “Jerzmanowice” projects. These algorithms use “classic” convolutional networks along with so-called fully convolutional U-Net neural networks. The proposed data processing methodology allows one to estimate the accuracy with which the height of each point in the unified point cloud is determined. During this project, the team also estimated the significance of the season and the related vegetation cycle in the accuracy of determining the height of the terrain surface with and without the developed algorithms.
The results of this project are described in the article entitled:
“Application of convolutional neural networks for low vegetation filtering from data acquired by UAVs”. DOI: https://doi.org/10.1016/j.isprsjprs.2019.09.014
The team conducted a project to test the positioning accuracy of UAVs from 2015–2018. A field base for testing the positioning accuracy of UAVs was established, and a set of basic tests to determine accuracy parameters in specific mission scenarios was created. The work done under this project makes it possible to determine the boundary conditions for the applicability of UAVs in various missions and represents an important step in the full test procedure of on-board computers and sensors. The team also conducted an analysis of positioning accuracy for selected UAVs.
The results of this project are described in the articles entitled:
“Testing procedure of Unmanned Aerial Vehicles (UAVs) trajectory in automatic missions” (DOI: https://doi.org/10.3390/app9173488) and “Assessment of the accuracy of positioning unmanned aerial vehicles.”
From 2016–2017, the team used Tatra National Park as a study site to evaluate the suitability of using UAVs to inventory hiking trails in alpine areas. The research team developed a complete methodology for collecting data in the difficult and demanding terrain presented by the trails on the steep slopes of the Tatra Mountains. The accuracy of the methodology was also assessed, and the possibility of using the developed methodology to assess the extent and severity of erosion along tourist trails and examine the succession of vegetation or stand condition in the area directly adjacent to tourist trails was evaluated.
The results this project are described in the article entitled:
“Assessment of the possibility of using unmanned aerial vehicles (UAVs) for the documentation of hiking trails in alpine areas”. DOI: https://doi.org/10.3390/s18010081
The “Kłodne Landslide” project focused on the landslide that began in June 2010 in the town of Kłodne (Limanowa commune) on the southern slope of Chełm Mountain, Island Beskids. This landslide is one of the largest and most dangerous landslides that began in recent years in the Polish Carpathians. Due to its scale (it covers an area of approximately 50 ha) and the extent of damage to buildings and infrastructure, the Kłodne landslide is studied by many scientific and research institutions. The National Research Institute of the Polish Geological Institute has studied the landslide since 2010, and students and employees of AGH UST have carried out work at the landslide site since 2013. Our team’s work aimed to determine the external borders of the landslide and test its stability. An analysis of the results collected since 2013 indicated that the landslide currently exhibits negligible activity. The research team performed classic geodetic surveys and used UAVs to update the cadastral data. Analyses were conducted to estimate the accuracy of determining the coordinates of the points constituting the content of the EGiB database based on the UAV-collected data.
The results of this project are described in the article entitled:
“Use of unmanned aerial vehicles (UAVs) for updating farmland cadastral data in areas subject to landslides”. DOI: https://doi.org/10.3390/ijgi7080331