A comparative study of remote sensing classification methods for monitoring and assessing desert vegetation using a UAV-based multispectral sensor

Al-Ali²,

et al. 2021

Plants

Self Cite

The rapid assessment and monitoring of native desert plants are essential in restoration and revegetation projects to track the changes in vegetation patterns in terms of vegetation coverage and structure. This work investigated advanced vegetation monitoring methods utilizing UAVs and remote sensing techniques at the Al Abdali protected site in Kuwait. The study examined the effectiveness of using UAV techniques to assess the structure of desert plants. We specifically examined the use of very-high-resolution aerial imagery to estimate the vegetation structure of Rhanterium epapposum (perennial desert shrub), assess the vegetation cover density changes in desert plants after rainfall events, and investigate the relationship between the distribution of perennial shrub structure and vegetation cover density of annual plants. The images were classified using supervised classification techniques (the SVM method) to assess the changes in desert plants after extreme rainfall events. A digital terrain model (DTM) and a digital surface model (DSM) were also generated to estimate the maximum shrub heights. The classified imagery results show that a significant increase in vegetation coverage occurred in the annual plants after rainfall events. The results also show a reasonable correlation between the shrub heights estimated using UAVs and the ground-truth measurements (R² = 0.66, p < 0.01). The shrub heights were higher in the high-cover-density plots, with coverage >30% and an average height of 77 cm. However, in the medium-cover-density (MD) plots, the coverage was <30%, and the average height was 52 cm. Our study suggests that utilizing UAVs can provide several advantages to critically support future ecological studies and revegetation and restoration programs in desert ecosystems.

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

The Use of Very-High-Resolution Aerial Imagery to Estimate the Structure and Distribution of the Rhanterium epapposum Community for Long-Term Monitoring in Desert Ecosystems

Al-Ali²,

et al. 2021

Plants

Self Cite

“…Kuwait hosts three main desalination plants. Furthermore, the area is affected by severe dust storms during the summer season, which highly contribute to pollution in this area [19,20]. The K-EPA maintains 15 distributed air quality monitoring stations to achieve an adequate area coverage.…”

Section: Description Of the Study Areamentioning

confidence: 99%

Air Quality Assessment in the State of Kuwait during 2012 to 2017

et al. 2021

This study aimed to examine the trend of ambient air pollution (i.e., ozone (O3), nitrogen monoxide (NO), nitrogen dioxide (NO2), nitrogen oxides (NOx), sulfur dioxide (SO2), carbon monoxide (CO), benzene (C6H6) and particulate matter with an aerodynamic diameter smaller than 10 microns (PM10), and non-methane hydrocarbons (NMHCs) at 10 monitoring stations located in the main residential and industrial areas in the State of Kuwait over 6 years (2012–2017). We found that the SO2 level in industrial areas (0.065 ppm) exceeded the allowable range of SO2 in residential areas (0.030 ppm). Air pollution variables were defined by the Environmental Public Authority of Kuwait (K-EPA). In this study, integrated statistical analysis was performed to compare an established air pollution database to Kuwait Ambient Air Quality Guidelines and to determine the association between pollutants and meteorological factors. All pollutants were positively correlated, with the exception of most pollutants and PM10 and O3. Meteorological factors, i.e., the ambient temperature, wind speed and humidity, were also significantly associated with the above pollutants. Spatial distribution mapping indicated that the PM10 level remained high during the southwest monsoon (the hot and dry season), while the CO level was high during the northeast monsoon (the wet season). The NO2 and O3 levels were high during the first intermonsoon season.

“…This benefits from the detailed spectral or morphological information offered by the images which also enables the utilization of various classical and state-of-the-art machine learning algorithms. Random forest, maximum likelihood classifier, and support vector machine techniques are commonly used machine learning classifiers in vegetation classification [40][41][42]. Besides the classical machine learning algorithms, rich information included in the ultra-high resolution images can be well utilized by the CNN-based deep learning models, with which a series of features are automatically extracted using the convolutional layers [43].…”

Section: Introductionmentioning

confidence: 99%

Early Detection of Encroaching Woody Juniperus virginiana and Its Classification in Multi-Species Forest Using UAS Imagery and Semantic Segmentation Algorithms

et al. 2021

Woody plant encroachment into grasslands ecosystems causes significantly ecological destruction and economic losses. Effective and efficient management largely benefits from accurate and timely detection of encroaching species at an early development stage. Recent advances in unmanned aircraft systems (UAS) enabled easier access to ultra-high spatial resolution images at a centimeter level, together with the latest machine learning based image segmentation algorithms, making it possible to detect small-sized individuals of target species at early development stage and identify them when mixed with other species. However, few studies have investigated the optimal practical spatial resolution of early encroaching species detection. Hence, we investigated the performance of four popular semantic segmentation algorithms (decision tree, DT; random forest, RF; AlexNet; and ResNet) on a multi-species forest classification case with UAS-collected RGB images in original and down-sampled coarser spatial resolutions. The objective of this study was to explore the optimal segmentation algorithm and spatial resolution for eastern redcedar (Juniperus virginiana, ERC) early detection and its classification within a multi-species forest context. To be specific, firstly, we implemented and compared the performance of the four semantic segmentation algorithms with images in the original spatial resolution (0.694 cm). The highest overall accuracy was 0.918 achieved by ResNet with a mean interaction over union at 85.0%. Secondly, we evaluated the performance of ResNet algorithm with images in down-sampled spatial resolutions (1 cm to 5 cm with 0.5 cm interval). When applied on the down-sampled images, ERC segmentation performance decreased with decreasing spatial resolution, especially for those images coarser than 3 cm spatial resolution. The UAS together with the state-of-the-art semantic segmentation algorithms provides a promising tool for early-stage detection and localization of ERC and the development of effective management strategies for mixed-species forest management.