Neural networks better recognize urban areas in drone images
Recognition of Urbanized Areas in UAV-Derived Very-High-Resolution Visible-Light Imagery
Computer Vision and Pattern Recognition
Summary
Identifying urban areas from drone photos is not always easy because places with buildings and plants can look similar. The authors tested different methods, including simple color-based rules and more complex neural networks, to see which works best. They found that neural networks correctly spot urban areas about 96% of the time, better than the color-based rules that were around 87%. They also showed that things like the time of year and how images are divided affect the accuracy. This helps improve mapping cities using drone images.
What this means in practice
- •For urban planning teams: Create high-accuracy maps of city development from detailed drone images to support planning decisions.
- •For environmental monitoring teams: Use improved classification techniques to track changes in urban sprawl versus green spaces over time from aerial imagery.
Authors
Edyta Puniach, Wojciech Gruszczyński, Paweł Ćwiąkała, Katarzyna Strząbała, Elżbieta Pastucha
Abstract
This study compared classifiers that differentiate between urbanized and non-urbanized areas based on unmanned aerial vehicle (UAV)-acquired RGB imagery. The tested solutions in-cluded numerous vegetation indices (VIs) thresholding and neural networks (NNs). The analysis was conducted for two study areas for which surveys were carried out using different UAVs and cameras. The ground sampling distances for the study areas were 10 mm and 15 mm, respectively. Reference classification was performed manually, obtaining approximately 24 million classified pix-els for the first area and approximately 3.8 million for the second. This research study included an analysis of the impact of the season on the threshold values for the tested VIs and the impact of image patch size provided as inputs for the NNs on classification accuracy. The results of the con-ducted research study indicate a higher classification accuracy using NNs (about 96%) compared with the best of the tested VIs, i.e., Excess Blue (about 87%). Due to the highly imbalanced nature of the used datasets (non-urbanized areas constitute approximately 87% of the total datasets), the Mat-thews correlation coefficient was also used to assess the correctness of the classification. The analysis based on statistical measures was supplemented with a qualitative assessment of the classification results, which allowed the identification of the most important sources of differences in classification between VIs thresholding and NNs.