Ecological Risk Assessment and Prediction Based on Scale Optimization—A Case Study of Nanning, a Landscape Garden City in China

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(5), P. 1304 - 1304

Published: Feb. 26, 2023

Analysis and prediction of urban ecological risk are crucial means for resolving the dichotomy between preservation economic development, thereby enhancing regional security fostering sustainable development. This study uses Nanning, a Chinese landscape garden city, as an example. Based on spatial granularity extent perspectives, using 30 m land use data, optimal scale assessment (ERA) is confirmed. also explores patterns temporal changes in Nanning scale. At same time, Patch-generating Land Use Simulation model used to predict Nanning’s 2036 under two scenarios propose conservation recommendations light results. The results show that: 120 7 km best scales ERA Nanning. Although distribution levels obviously different, overall relatively low, scenario protection 2036, area high small. can provide theoretical support cities civilization construction.

Language: Английский

---

Segmentation of individual mangrove trees using UAV-based LiDAR data

Ecological Informatics, Journal Year: 2023, Volume and Issue: 77, P. 102200 - 102200

Published: July 4, 2023

Accurate assessment of structural parameters is essential to effectively monitor the mangrove resources. However, extraction results are closely related segmentation individual trees. Although tree influenced by many factors, specific factors affecting trees, such as data source, image resolution, algorithm, and stand density, have not yet been elucidated. Therefore, in this study, canopy height models (CHMs) with different spatial resolutions were derived from unmanned aerial vehicle (UAV)-based light detection ranging (LiDAR) data. Moreover, watershed algorithm (WA), regional growth (RG), improved K-nearest neighbour (KNN) bird's eye view (BEV) faster region-based convolutional neural network (R-CNN) algorithms used segment trees based on CHMs LiDAR at three sites varying densities. Finally, algorithms, resolutions, forest densities comparatively assessed determine their influence Segmentation accuracy KNN was highest among CHM-based WA, RG, an optimal F 0.893 minimum 0.628. R-CNN had value 0.931 0.612. Based results, overall ranking BEV Faster > RG WA. The for low-density (LD) medium-density (MD) high-density (HD). For LD MD sites, values (0.931 0.712, respectively). HD site, all performed poorly, except higher than 0.6. CHM result 0.1 m best, being better 0.25 0.5 m. Our demonstrated that affected deep learning those other site limited. further research necessary improve sites.

Language: Английский

---

Early detection of rubber tree powdery mildew using UAV-based hyperspectral imagery and deep learning

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 220, P. 108909 - 108909

Published: April 12, 2024

Language: Английский

---

Improved random forest classification model combined with C5.0 algorithm for vegetation feature analysis in non-agricultural environments

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: May 6, 2024

Abstract In response to the challenges posed by high computational complexity and suboptimal classification performance of traditional random forest algorithms when dealing with high-dimensional noisy non-agricultural vegetation satellite data, this paper proposes an enhanced algorithm based on C5.0 algorithm. The focuses Liaohe Plain, selecting two distinct landscape patterns in Shenbei New District Changtu County as research objects. High-resolution data from GF-2 serves experimental dataset. This introduces ensemble feature method bagging concept improve original model. enhances likelihood features beneficial classifying positive class samples, avoiding excessive removal useful negative samples. approach ensures importance model diversity. is then employed for selection, index (EVI) utilized coverage estimation. Results indicate that employing a multi-scale parameter selection tool, combined limited field-measured facilitates identification plant species landscapes. effectively selects features, minimizing information redundancy. established object-oriented achieves impressive accuracy 94.02% aerial imagery dataset, EVI-based estimation demonstrating accuracy. experiments same test set, proposed attains average 90.20%, outperforming common such bidirectional encoder representation transformer, FastText, convolutional neural network, which achieve accuracies ranging 84.41 88.33% identifying artificial habitat features. exhibits competitive edge compared other algorithms. These findings contribute scientific evidence protecting agricultural ecosystems restoring ecosystem biodiversity.

Language: Английский

---

Landscape Pattern and Ecological Risk Assessment in Guilin Based on Land Use Change

International Journal of Environmental Research and Public Health, Journal Year: 2023, Volume and Issue: 20(3), P. 2045 - 2045

Published: Jan. 22, 2023

The land use and ecological risk patterns in Guilin, which is the only innovation demonstration zone under National Sustainable Development Agenda China with a focus on sustainable of natural resources, have changed significantly as result combined impact climate change human activities, thus presenting challenges to development local area. This research employs an assessment model spatial analysis techniques order analyze correlation between risk, evaluate temporal evolution characteristics at overall county scales Guilin. results reveal following: (1) A total 1848.6 km2 types Guilin from 2000 2020, construction has gradually expanded central urban area suburbs increasing internal stability each year. (2) level showed decreasing trend city scale, but some regions still distribution scale. (3) value significant correlation, clustering effect, was consistent class areas. can provide reference for control landscape resource cities.

Language: Английский

---

Construction and Optimization of an Ecological Network in the Yellow River Source Region Based on MSPA and MCR Modelling

International Journal of Environmental Research and Public Health, Journal Year: 2023, Volume and Issue: 20(4), P. 3724 - 3724

Published: Feb. 20, 2023

The source region of the Yellow River (SRYR) is an important water conservation and farming area in China. Under dual influence natural environment external pressure, ecological patches are becoming increasingly fragmented, landscape connectivity continuously declining, which directly affect patch pattern SRYR sustainable development. In SRYR, morphological spatial analysis (MSPA) index methods were used to extract ecologically sources. Based on minimum cumulative resistance model (MCR), Linkage Mapper was generate a potential corridor, then stepped stone identified extracted by gravity betweenness centrality build optimal network. distribution core accounting for 80.53% total grassland area. 10 sources based 15 corridors MCR mainly distributed central eastern regions SRYR. Through centrality, added, 45 planned obtained optimize network enhance east west connectivity. Our research results can provide reference protection ecosystem, have guiding significance practical value construction fragmented areas.

Language: Английский

---

Assessing Derawan Island’s Coral Reefs over Two Decades: A Machine Learning Classification Perspective

Sensors, Journal Year: 2024, Volume and Issue: 24(2), P. 466 - 466

Published: Jan. 12, 2024

This study aims to understand the dynamic changes in coral reef habitats of Derawan Island over two decades (2003, 2011, and 2021) using advanced machine learning classification techniques. The motivation stems from urgent need for accurate, detailed environmental monitoring inform conservation strategies, particularly ecologically sensitive areas like reefs. We employed non-parametric algorithms, including Random Forest (RF), Support Vector Machine (SVM), Classification Regression Tree (CART), assess spatial temporal habitats. Our analysis utilized high-resolution data Landsat 9, 7, Sentinel-2, Multispectral Aerial Photos. RF algorithm proved be most achieving an accuracy 71.43% with 73.68% 78.28% findings indicate that is significantly influenced by geographic resolution quality field satellite/aerial image data. Over decades, there was a notable decrease area 2003 reduction 16 hectares, followed slight increase but more heterogeneous densities between 2011 2021. underscores nature efficacy monitoring. insights gained highlight importance analytical methods guiding efforts understanding ecological time.

Language: Английский

---

Explainable machine learning-based fractional vegetation cover inversion and performance optimization – A case study of an alpine grassland on the Qinghai-Tibet Plateau

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102768 - 102768

Published: Aug. 10, 2024

Fractional Vegetation Cover (FVC) serves as a crucial indicator in ecological sustainability and climate change monitoring. While machine learning is the primary method for FVC inversion, there are still certain shortcomings feature selection, hyperparameter tuning, underlying surface heterogeneity, explainability. Addressing these challenges, this study leveraged extensive field data from Qinghai-Tibet Plateau. Initially, selection algorithm combining genetic algorithms XGBoost was proposed. This integrated with Optuna tuning method, forming GA-OP combination to optimize learning. Furthermore, comparative analyses of various models inversion alpine grassland were conducted, followed by an investigation into impact heterogeneity on performance using NDVI Coefficient Variation (NDVI-CV). Lastly, SHAP (Shapley Additive exPlanations) employed both global local interpretations optimal model. The results indicated that: (1) exhibited favorable terms computational cost accuracy, demonstrating significant potential tuning. (2) Stacking model achieved among seven (R2 = 0.867, RMSE 0.12, RPD 2.552, BIAS −0.0005, VAR 0.014), ranking follows: > CatBoost LightGBM RFR KNN SVR. (3) NDVI-CV enhanced result reliability excluding highly heterogeneous regions that tended be either overestimated or underestimated. (4) revealed decision-making processes perspectives. allowed deeper exploration causality between features targets. developed high-precision scheme, successfully achieving accurate proposed approach provides valuable references other parameter inversions.

Language: Английский

---

Study on Individual Tree Segmentation of Different Tree Species Using Different Segmentation Algorithms Based on 3D UAV Data

Forests, Journal Year: 2023, Volume and Issue: 14(7), P. 1327 - 1327

Published: June 28, 2023

Individual structural parameters of trees, such as forest stand tree height and biomass, serve the foundation for monitoring dynamic changes in resources. are closely related to individual crown segmentation. Although three-dimensional (3D) data have been successfully used determine segmentation, this phenomenon is influenced by various factors, (i) source 3D data, (ii) segmentation algorithm, (iii) species. To further quantify effect factors on light detection ranging (LiDAR) image-derived points were obtained unmanned aerial vehicles (UAVs). Three different algorithms (PointNet++, Li2012, layer-stacking (LSS)) segment crowns four The results show that two accuracy LiDAR was generally better than using with a maximum difference 0.13 F values. For three algorithms, PointNet++ algorithm best, an value 0.91, whereas result LSS yields worst result, 0.86. Among tested species, Liriodendron chinense followed Magnolia grandiflora Osmanthus fragrans, Ficus microcarpa worst. Similar trees observed based data. fragrans superior according determined These demonstrate species all impact trees. greatest, source. Consequently, future research acquisition methods should be selected deep learning adopted improve

Language: Английский

---

Tradeoffs among multi-source remote sensing images, spatial resolution, and accuracy for the classification of wetland plant species and surface objects based on the MRS_DeepLabV3+ model

Ecological Informatics, Journal Year: 2024, Volume and Issue: 81, P. 102594 - 102594

Published: April 8, 2024

Classification of wetland plant species (PlatSpe) and surface objects (SurfObj) in remote sensing images faces significant challenges due to the high diversity PlatSpe fragmented nature SurfObj. Unmanned aerial vehicle (UAV) satellite are primary data sources for classification However, there is still insufficient research on effect various spatial resolutions results. This study essentially focuses Huixian Wetland Guilin, Guangxi, China through utilizing UAV with varying as sources. To this end, MRS_DeepLabV3+ model constructed based multi-resolution segmentation DeepLabV3+, SurfObj appropriately classified model. The obtained results reveal that: (1) optimal scale parameter (SP) capable achieving higher accuracy compared DeepLabV3+. SPs both gradually lessen decreasing resolution, require larger images. (2) In image models, OA kappa exhibit a trend reduction resolution. (3) overall accuracies models superior resolution intervals 2 16 m. investigation can be regarded valuable reference selecting classification.

Language: Английский

---