Mapping Wetland Plant Communities Using Unmanned Aerial Vehicle Hyperspectral Imagery by Comparing Object/Pixel-Based Classifications Combining Multiple Machine-Learning Algorithms

Du, Baojia; Mao, Dehua; Wang, Zongming; Qiu, Zhiqiang; Yan, Hengqi; Feng, Kaidong; Zhang, Zhongbin

doi:10.1109/jstars.2021.3100923

Cited by 32 publications

(15 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The short‐wave infrared region (SWIR) has been found to be important for predicting wetlands (Mcpartland et al., 2019; Meingast et al., 2014). Various topographical features such as topographical wetness index (TWI) have been found to be important for predicting wetlands, but only for specific spatial resolutions (Lidberg et al., 2020; Rasanen et al., 2014) Thus, both high spatial and spectral resolution is necessary to map wetland plant communities (Du et al., 2021). Along the ridge‐snowbed gradient, snowbed had the lowest accuracy.…”

Section: Discussionmentioning

confidence: 99%

Land cover classification of treeline ecotones along a 1100 km latitudinal transect using spectral‐ and three‐dimensional information from UAV‐based aerial imagery

Mienna

Klanderud

Ørka

et al. 2022

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

The alpine treeline ecotone is expected to move upwards in elevation with global warming. Thus, mapping treeline ecotones is crucial in monitoring potential changes. Previous remote sensing studies have focused on the usage of satellites and aircrafts for mapping the treeline ecotone. However, treeline ecotones can be highly heterogenous, and thus the use of imagery with higher spatial resolution should be investigated. We evaluate the potential of using unmanned aerial vehicles (UAVs) for the collection of ultra-high spatial resolution imagery for mapping treeline ecotone land covers. We acquired imagery and field reference data from 32 treeline ecotone sites along a 1100 km latitudinal gradient in Norway (60-69°N). Before classification, we performed a superpixel segmentation of the UAV-derived orthomosaics and assigned land cover classes to segments: rock, water, snow, shadow, wetland, tree-covered area and five classes within the ridge-snowbed gradient. We calculated features providing spectral, textural, three-dimensional vegetation structure, topographical and shape information for the classification. To evaluate the influence of acquisition time during the growing season and geographical variations, we performed four sets of classifications: global, seasonal-based, geographical regional-based and seasonalregional-based. We found no differences in overall accuracy (OA) between the different classifications, and the global model with observations irrespective of data acquisition timing and geographical region had an OA of 73%. When accounting for similarities between closely related classes along the ridgesnowbed gradient, the accuracy increased to 92.6%. We found spectral features related to visible, red-edge and near-infrared bands to be the most important to predict treeline ecotone land cover classes. Our results show that the use of UAVs is efficient in mapping treeline ecotones, and that data can be acquired irrespective of timing within a growing season and geographical region to get accurate land cover maps. This can overcome constraints of a short field-season or low-resolution remote sensing data.

show abstract

Section: Discussionmentioning

confidence: 99%

Land cover classification of treeline ecotones along a 1100 km latitudinal transect using spectral‐ and three‐dimensional information from UAV‐based aerial imagery

Mienna

Klanderud

Ørka

et al. 2022

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

show abstract

“…Based on Landsat 8 images and RF classifier, Sharma et al [98] Some studies using hyperspectral data covered only the visible spectrum (VIS) and the near-infrared range (NIR), excluding the short-wave infrared (SWIR) range, providing lower results. The studies achieved 69-73% OA (in a spectrum range of 450−950 nm) for six wetland plant communities [104] and 71% OA (in a spectrum range of 400-1000 nm) for 19 classes of herbaceous vegetation [97]. The higher results we obtained (OA > 83%) may have been because we used the whole range of the spectrum (416-2510 nm) instead of only a part of it.…”

Section: Discussionmentioning

confidence: 77%

Airborne HySpex Hyperspectral Versus Multitemporal Sentinel-2 Images for Mountain Plant Communities Mapping

2022

View full text Add to dashboard Cite

Climate change and anthropopression significantly impact plant communities by leading to the spread of expansive and alien invasive plants, thus reducing their biodiversity. Due to significant elevation gradients, high-mountain plant communities in a small area allow for the monitoring of the most important environmental changes. Additionally, being a tourist attraction, they are exposed to direct human influence (e.g., trampling). Airborne hyperspectral remote sensing is one of the best data sources for vegetation mapping, but flight campaign costs limit the repeatability of surveys. A possible alternative approach is to use satellite data from the Copernicus Earth observation program. In our study, we compared multitemporal Sentinel-2 data with HySpex airborne hyperspectral images to map the plant communities on Tatra Mountains based on open-source R programing implementation of Random Forest and Support Vector Machine classifiers. As high-mountain ecosystems are adapted to topographic conditions, the input of Digital Elevation Model (DEM) derivatives on the classification accuracy was analyzed and the effect of the number of training pixels was tested to procure practical information for field campaign planning. For 13 classes (from rock scree communities and alpine grasslands to montane conifer and deciduous forests), we achieved results in the range of 76–90% F1-score depending on the data set. Topographic features: digital terrain model (DTM), normalized digital surface model (nDSM), and aspect and slope maps improved the accuracy of HySpex spectral images, transforming their minimum noise fraction (MNF) bands and Sentinel-2 data sets by 5–15% of the F1-score. Maps obtained on the basis of HySpex imagery (2 m; 430 bands) had a high similarity to maps obtained on the basis of multitemporal Sentinel-2 data (10 m; 132 bands; 11 acquisition dates), which was less than one percentage point for classifications based on 500–1000 pixels; for sets consisting of 50–100 pixels, Random Forest (RF) offered better accuracy.

show abstract

“…Grey-level Co-occurrence Matrix (GLCM) was the most commonly used method to extract texture information from images [25], and it was proved useful in aquaculture area classification [10], [14]. In this study, we first used principal components analysis (PCA) [26] to generate the first few principal component layers which indicated over 99% of the information of the input Sentinel-2 images.…”

Section: Ndvi = (Nir -R) / (Nir + R)mentioning

confidence: 99%

Dynamic Mapping of Inland Freshwater Aquaculture Areas in Jianghan Plain, China

Han

Huang

Ling

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Freshwater aquaculture in Jianghan Plain plays a key role in the whole industry of China. It was expanding rapidly to meet the fast-growing demand of consumption in these decades. The spatial distribution change of aquaculture in Jianghan Plain has attracted many researchers in recent years and it is worth further investigating. However, the accuracy and the quality of inland aquaculture classification and mapping still have space to be improved. Our study attempts to use Sentinel-1 and Sentinel-2 data to classify aquaculture areas, non-aquaculture water, and non-water areas. We applied image segmentation and pixel-based feature computation to obtain candidate datasets. Recursive feature elimination (RFE) was used to find the optimal datasets which were then imported into a modified UNet for classification. This workflow was then applied to the annual datasets for better observation of the spatial change of aquaculture areas. In total 12 indices and features were found most influential and preserved after RFE. The overall classification accuracy reached 96.83%, while the precision and recall of aquaculture areas reached 84.47% and 90.48%, respectively. The combination of object-based and pixel-based image analysis showed its advantage in inland aquaculture classification compared to other studies. Integrated rice-crawfish farming was found a key factor motivating the rapid development of aquaculture. Policies about the fishing ban and ecosystem restoration were also found deeply affecting the spatial change of aquaculture in Jianghan Plain.

show abstract

Mapping Wetland Plant Communities Using Unmanned Aerial Vehicle Hyperspectral Imagery by Comparing Object/Pixel-Based Classifications Combining Multiple Machine-Learning Algorithms

Cited by 32 publications

References 65 publications

Land cover classification of treeline ecotones along a 1100 km latitudinal transect using spectral‐ and three‐dimensional information from UAV‐based aerial imagery

Land cover classification of treeline ecotones along a 1100 km latitudinal transect using spectral‐ and three‐dimensional information from UAV‐based aerial imagery

Airborne HySpex Hyperspectral Versus Multitemporal Sentinel-2 Images for Mountain Plant Communities Mapping

Dynamic Mapping of Inland Freshwater Aquaculture Areas in Jianghan Plain, China

Contact Info

Product

Resources

About