Abstract
Habitat degradation, mostly caused by human impact, is one of the key drivers of biodiversity loss. This is a global problem, causing a decline in the number of pollinators, such as hoverflies. In the process of digitalizing ecological studies in Serbia, remote-sensing-based land cover classification has become a key component for both current and future research. Object-based land cover classification, using machine learning algorithms of very high resolution (VHR) imagery acquired by an unmanned aerial vehicle (UAV) was carried out in three different study sites on Mt. Stara Planina, Eastern Serbia. UAV land cover classified maps with seven land cover classes (trees, shrubs, meadows, road, water, agricultural land, and forest patches) were studied. Moreover, three different classification algorithms—support vector machine (SVM), random forest (RF), and k-NN (k-nearest neighbors)—were compared. This study shows that the random forest classifier performs better with respect to the other classifiers in all three study sites, with overall accuracy values ranging from 0.87 to 0.96. The overall results are robust to changes in labeling ground truth subsets. The obtained UAV land cover classified maps were compared with the Map of the Natural Vegetation of Europe (EPNV) and used to quantify habitat degradation and assess hoverfly species richness. It was concluded that the percentage of habitat degradation is primarily caused by anthropogenic pressure, thus affecting the richness of hoverfly species in the study sites. In order to enable research reproducibility, the datasets used in this study are made available in a public repository.
Highlights
Most image classification methods rely on pixel-based techniques that have limitations when it comes to high-resolution satellite data and unmanned aerial vehicle (UAV) imagery [16,43,44]
As small-scale changes cannot be detected at low resolutions, it is important to address the possibility of obtaining high spatial and temporal resolution UAV data that are suitable for generating detailed land cover maps [89]
It is possible to obtain very high resolution (VHR) UAV land cover classified maps in more heterogeneous study sites, but pinpointing limiting factors of data acquired in a complex area with a high slope of the terrain needs to be addressed; Proposed UAV-based land cover classification, along with both the potential vegetation types obtained from the EPNV map and biodiversity expert knowledge, can be applied in order to quantify habitat degradation in selected study areas; The initial results of linking the quantified habitat degradation with the biodiversity loss indicate the utility of the proposed framework; Comprehensive supplementary materials, including image processing steps for producing the land cover classified map in the form of a video recording guidance, along with raw data, ensure research reproducibility
Summary
Containing high intra-class spectral variability, features in VHR images can be separated based on spatial, textural, and contextual information [43]. Taking this into consideration, pixel-based image classification is replaced with object-based image analysis (OBIA), a new approach for managing spectral variability. In order to reduce the intra-class spectral variability, OBIA works with groups of homogenous and contiguous pixels ( called geographical objects; segments) with similar information to base units in order to conduct the classification. In order to perform OBIA, two steps are required: image segmentation and image classification [47] Since this method implies obtaining geographic information from remote sensing imagery analysis, the new term
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
