Double-Branch Multi-Scale Contextual Network: A Model for Multi-Scale Street Tree Segmentation in High-Resolution Remote Sensing Images

Sensors, Journal Year: 2024, Volume and Issue: 24(4), P. 1110 - 1110

Published: Feb. 8, 2024

Street trees are of great importance to urban green spaces. Quick and accurate segmentation street from high-resolution remote sensing images is significance in space management. However, traditional methods can easily miss some targets because the different sizes trees. To solve this problem, we propose Double-Branch Multi-Scale Contextual Network (DB-MSC Net), which has two branches a (MSC) block encoder. The MSC combines parallel dilated convolutional layers transformer blocks enhance network’s multi-scale feature extraction ability. A channel attention mechanism (CAM) added decoder assign weights features RGB normalized difference vegetation index (NDVI). We proposed benchmark dataset test improvement our network. Experimental research showed that DB-MSC Net demonstrated good performance compared with typical like Unet, HRnet, SETR recent methods. overall accuracy (OA) was improved by at least 0.16% mean intersection over union 1.13%. model’s meets requirements

Language: Английский

---

Evaluation of tree stump measurement methods for estimating diameter at breast height and tree height

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 129, P. 103828 - 103828

Published: April 27, 2024

Language: Английский

---

Integration of Hyperspectral Imaging and Deep Learning for Sustainable Mangrove Management and Sustainable Development Goals Assessment

Wetlands, Journal Year: 2025, Volume and Issue: 45(1)

Published: Jan. 1, 2025

Language: Английский

---

Advances in Instance Segmentation: Technologies, Metrics and Applications in Computer Vision

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129584 - 129584

Published: Jan. 1, 2025

Language: Английский

---

Earth Observation Data for Mangrove Monitoring and Management at the Red Sea Coastline, Egypt

Springer remote sensing/photogrammetry, Journal Year: 2025, Volume and Issue: unknown, P. 145 - 175

Published: Jan. 1, 2025

Language: Английский

---

ResLMFFNet: a real-time semantic segmentation network for precision agriculture

Journal of Real-Time Image Processing, Journal Year: 2024, Volume and Issue: 21(4)

Published: May 28, 2024

Abstract Lightweight multiscale-feature-fusion network (LMFFNet), a proficient real-time CNN architecture, adeptly achieves balance between inference time and accuracy. Capturing the intricate details of precision agriculture target objects in remote sensing images requires deep SEM-B blocks LMFFNet model design. However, employing numerous units leads to instability during backward gradient flow. This work proposes novel residual-LMFFNet (ResLMFFNet) for ensuring smooth flow within blocks. By incorporating residual connections, ResLMFFNet improved accuracy without affecting speed number trainable parameters. The results experiments demonstrate that this architecture has achieved superior performance compared other architectures across diverse applications involving UAV satellite images. Compared LMFFNet, enhances Jaccard Index values by 2.1% tree detection, 1.4% crop 11.2% wheat-yellow rust detection. Achieving these remarkable levels involves maintaining almost identical computational complexity as model. source code is available on GitHub: https://github.com/iremulku/Semantic-Segmentation-in-Precision-Agriculture .

Language: Английский

---

EIAGA-S: Rapid Mapping of Mangroves Using Geospatial Data without Ground Truth Samples

Forests, Journal Year: 2024, Volume and Issue: 15(9), P. 1512 - 1512

Published: Aug. 29, 2024

Mangrove forests are essential for coastal protection and carbon sequestration, yet accurately mapping their distribution remains challenging due to spectral similarities with other vegetation. This study introduces a novel unsupervised learning method, the Elite Individual Adaptive Genetic Algorithm-Semantic Inference (EIAGA-S), designed high-precision semantic segmentation of mangrove using remote sensing images without need ground truth samples. EIAGA-S integrates an adaptive Algorithm elite individual’s evolution strategy, optimizing process. A new Enhanced Vegetation Index (MEVI) was developed better distinguish mangroves from vegetation types within feature space. constructs rules through iterative rule stacking enhances boundary information connected component analysis. The method evaluated multi-source dataset covering Hainan Dongzhai Port Nature Reserve in China. experimental results demonstrate that achieves superior overall mIoU (mean intersection over union) 0.92 F1 score 0.923, outperforming traditional models such as K-means SVM (Support Vector Machine). detailed analysis confirms EIAGA-S’s ability extract fine-grained patches. includes five categories: canopy, terrestrial vegetation, buildings streets, bare land, water bodies. proposed model offers precise data-efficient solution while eliminating dependency on extensive field sampling labeled data. Additionally, MEVI index facilitates large-scale monitoring. In future work, can be integrated long-term data analyze forest dynamics under climate change conditions. innovative approach has potential applications rapid detection, environmental protection, beyond.

Language: Английский

---

Progress in Remote Sensing Monitoring of Mangrove Carbon Storage

National Remote Sensing Bulletin, Journal Year: 2024, Volume and Issue: 0(0), P. 1 - 21

Published: Jan. 1, 2024

2024年5月11日,广西大学海洋学院的邓淞文团队与中国科学院地理科学与资源研究所的杨飞团队在《遥感学报》发表文章,深入剖析红树林碳库遥感研究。两团队系统梳理了红树林碳库遥感的发展历程,并探讨了光学遥感和雷达遥感技术在红树林碳储量估算中的应用。同时,他们关注生物碳库与土壤碳库的碳储量研究,提出红树林在碳中和目标中的关键角色。该研究为红树林碳库遥感研究提供了新视角,并展望了无人机遥感技术和人工智能在此领域的应用前景。

---