Hyperspectral imagery provides detailed spectral information that can be used for tree species discrimination. The aim of this study is to assess spectral–spatial complexity reduction techniques for tree species classification using an airborne prism experiment (APEX) hyperspectral image. The methodology comprised the following main steps: (1) preprocessing (removing noisy bands) and masking out non-forested areas; (2) applying dimensionality reduction techniques, namely, independent component analysis (ICA), principal component analysis (PCA), and minimum noise fraction transformation (MNF), and stacking the selected dimensionality-reduced (DR) components to create new data cubes; (3) super-pixel segmentation on the original image and on each of the dimensionality-reduced data cubes; (4) tree species classification using a random forest (RF) classifier; and (5) accuracy assessment. The results revealed that tree species classification using the APEX hyperspectral imagery and DR data cubes yielded good results (with an overall accuracy of 80% for the APEX imagery and an overall accuracy of more than 90% for the DR data cubes). Among the classification results of the DR data cubes, the ICA-transformed components performed best, followed by the MNF-transformed components and the PCA-transformed components. The best class performance (according to producer’s and user’s accuracy) belonged to Picea abies and Salix alba. The other classes (Populus x (hybrid), Alnus incana, Fraxinus excelsior, and Quercus robur) performed differently depending on the different DR data cubes used as the input to the RF classifier.
1 January 2019 in European Journal of Remote Sensing
Scale is a critical factor when studying patterns and the processes that cause them. A variety of approaches have been used to define the concept of scale but confusion and ambiguities remain regarding scale types and their definitions. The objectives of this study were therefore (1) to review existing types and definitions of scale, and (2) to systematically investigate the ambiguities in scale definitions and to determine the applicability of the various scale types and definitions. Through a comprehensive literature review, we identified seven types of scales and designed a survey for the seven definitions of scale and interviewed 150 scientists. The results show that the more cartography related types of scale are relatively well known while the more abstract dimensions are less known and are most ambiguous. Based on graphical examples, participants were asked which spatial scales are most relevant for their work. Surprisingly, composite objects such as a forest stand were most relevant followed by individual objects such as single trees and, lastly, more generalized categorizes or meta-objects such as “forested area”. We have drawn some conclusions that will help to clarify the different types of scale in regard to their practical use.
Landslide mapping and analysis are essential aspects of hazard and risk analysis. Landslides can block rivers and create landslide-dammed lakes, which pose a significant risk for downstream areas. In this research, we used an object-based image analysis approach to map geomorphological features and related changes and assess the applicability of Sentinel-1 data for the fast creation of post-event digital elevation models (DEMs) for landslide volume estimation. We investigated the Hítardalur landslide, which occurred on the 7 July 2018 in western Iceland, along with the geomorphological changes induced by this landslide, using optical and synthetic aperture radar data from Sentinel-2 and Sentinel-1. The results show that there were no considerable changes in the landslide area between 2018 and 2019. However, the landslide-dammed lake area shrunk between 2018 and 2019. Moreover, the Hítará river diverted its course as a result of the landslide. The DEMs, generated by ascending and descending flight directions and three orbits, and the subsequent volume estimation revealed that—without further post-processing—the results need to be interpreted with care since several factors influence the DEM generation from Sentinel-1 imagery.
26 September 2018 in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
There is a growing use of Earth observation (EO) data for support planning in humanitarian crisis response. Information about number and dynamics of displaced population in camps is essential to humanitarian organizations for decision-making and action planning. Dwelling extraction and categorisation is a challenging task, due to the problems in separating different dwellings under different conditions, with wide range of sizes, colour and complex spatial patterns. Nowadays, so-called deep learning techniques such as deep convolutional neural network (CNN) are used for understanding image content and object recognition. Although recent developments in the field of computer vision have introduced CNN networks as a practical tool also in the field of remote sensing, the training step of these techniques is rather time-consuming and samples for the training process are rarely transferable to other application fields. These techniques also have not been fully explored for mapping camps. Our study analyses the potential of a CNN network for dwelling extraction to be embedded as initial step in a comprehensive object-based image analysis (OBIA) workflow. The results were compared to a semi-automated, i.e. combined knowledge-/sample-based, OBIA classification. The Minawao refugee camp in Cameroon served as a case study due to its well-organised, clearly distinguishable dwelling structure. We use manually delineated objects as initial input for the training samples, while the CNN network is structured with two convolution layers and one max pooling.
22 January 2022 in Science of The Total Environment
Landslide-dammed lakes pose a risk for upriver and downriver communities and infrastructure. The 2016 Kaikōura earthquake affected the northeastern region of the South Island in New Zealand, triggering numerous landslides that dammed river courses leading to the formation of hundreds of dammed lakes. Detecting and monitoring landslide-dammed lakes is important for disaster management. Satellite remote sensing imagery is often complementary to field acquisitions to obtain an overview of large and remote areas and thus can be exploited to monitor landslide-dammed lakes. Yet, the strengths and limitations of freely available multi-temporal satellite imagery for landslide-dammed lake assessment remain largely unexplored. This study aimed at automatically mapping landslide-dammed lakes caused by the 2016 Kaikōura earthquake and monitoring their evolution using time series of Sentinel-2 imagery and the computing capabilities of the Google Earth Engine. Our approach combined dynamic thresholding, change detection, and connected component analysis. Landslide-dammed lakes larger than 300 m2 and located on relatively flat terrain were detected with reasonable accuracy, while lakes located in steeply incised valleys were detected less frequently. Despite the challenging topographical and environmental characteristics of the study area, we were able to detect landslide-dammed lake candidates at a regional scale. Temporal monitoring of the evolution of the landslide-dammed lake area revealed four distinct patterns: 1) constant, 2) increasing, 3) decreasing, and 4) variable. Our approach contributes to the understanding of the utility and limitations of temporal and spatial monitoring of landslide-dammed lakes, their potential cascading hazards and their interactions.
Large rainfall-induced landslides are among the most dangerous natural hazards in Taiwan, posing a risk for people and infrastructure. Thus, better knowledge about the evolution of landslides and their impact on the downstream area is of high importance for disaster mitigation. The aim of this study is twofold: (1) to semi-automatically map the evolution of the Butangbunasi landslide in south-central Taiwan using satellite remote sensing data, and (2) to investigate the potential correlation between changes in landslide area and heavy rainfall during typhoon events. Landslide area, as well as temporary landslide-dammed lakes, were semi-automatically identified using object-based image analysis (OBIA), based on 20 Landsat images from 1984 to 2018. Hourly rainfall data from the Taiwan Central Weather Bureau (CWB) was complemented with rainfall data from Climate Hazards Group Infrared Precipitation with Station data (CHIRPS) to examine the potential relationship between landslide area changes and rainfall as a triggering factor. The OBIA mapping results revealed that the most significant landslide extension happened after typhoon Morakot in 2009. We found a moderate positive relationship between the landslide area change and the duration of the heavy rainfall event, whereas daily precipitation, cumulative rainfall and mean intensity did not present strong significant correlations.