Accelerated forest modeling from tree canopy point clouds via deep learning

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 132, P. 104074 - 104074

Published: Aug. 1, 2024

Rapid generation of tree models from point clouds canopies holds wide-ranging applications in the field earth sciences, including forest ecology research, environmental monitoring, and management. Traditional modeling methods rely on procedural to simulate growth, which are time-consuming due their extensive manual parameterization. Furthermore, existing deep learning struggle generate visually realistic because complex branch structures specific natural patterns trees. To address these challenges, this paper proposes a novel learning-based method for rapidly generating that align with shape canopy. Different traditional methods, we use neural networks build graphs models. Our consists two main steps: i) 3D coordinates each node generated canopy cloud by designed coordinate network; ii) graph network is proposed predict attributes adjacency relationship between nodes. form structure, discrete nodes connected using minimum spanning algorithm combined relationship. The include width, whether it leaf node, size, used subsequent construction tree's mesh. validate effectiveness our method, large-scale dataset containing 10 forests 3216 constructed open sourced study canopies. Experimental results demonstrate method's efficiency quickly (reducing average canopy-to-tree reconstruction time 7 min less than 0.5 s) while preserving visual authenticity accurately matching shapes, making suitable wide range reconstructions.

Language: Английский

---

Fruit Water Stress Index of Apple Measured by Means of Temperature-Annotated 3D Point Cloud

Plant Phenomics, Journal Year: 2024, Volume and Issue: 6

Published: Jan. 1, 2024

In applied ecophysiological studies related to global warming and water scarcity, the status of fruit is increasing importance in context fresh food production. present work, a stress index (

Language: Английский

---

Comparative Analysis of Machine Learning Techniques and Data Sources for Dead Tree Detection: What Is the Best Way to Go?

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(16), P. 3086 - 3086

Published: Aug. 21, 2024

In Central Europe, the extent of bark beetle infestation in spruce stands due to prolonged high temperatures and drought has created large areas dead trees, which are difficult monitor by ground surveys. Remote sensing is only possibility for assessment tree areas. Several options exist mapping individual including different sources processing techniques. Satellite images, aerial images from UAVs can be used as sources. Machine deep learning techniques included techniques, although models often presented without proper realistic validation.This paper compares methods monitoring using three data sources: multispectral imagery, PlanetScope satellite Sentinel-2 well two methods. The classification Random Forest (RF) neural network (NN) modalities: pixel- object-based. total, 12 combinations presented. results were evaluated types reference data: accuracy model on validation vector-format semi-automatic polygons a human evaluator, referred real Ground Truth. imagery was found have highest accuracy, with CNN achieving up 98% object classification. A higher achieved combining pixel RF (87% Sentinel-2). For Imagery, best result 89%, combination object-based classifications. comparison Truth showed decrease 89% around 70%. conclusion, most effective tool calamity terms precision but advantage fast availability shorter time, together larger coverage

Language: Английский

---

Cucumber seedlings segmentation network based on multi-view geometric graph encoder from 3D point clouds

Plant Phenomics, Journal Year: 2024, Volume and Issue: 6

Published: Jan. 1, 2024

Plant phenotyping plays a pivotal role in observing and comprehending the growth development of plants. In phenotyping, plant organ segmentation based on 3D point clouds has garnered increasing attention recent years. However, using only geometric relationship features Euclidean space still cannot accurately segment measure To this end, we mine more propose network multiview graph encoder, called SN-MGGE. First, construct cloud acquisition platform to obtain cucumber seedling dataset, employ CloudCompare software annotate data. The GGE module is then designed generate features, including relationships shape structure, via encoder over hyperbolic spaces. Finally, semantic results are obtained downsampling operation multilayer perceptron. Extensive experiments dataset clearly show that our proposed SN-MGGE outperforms several mainstream networks (e.g., PointNet++, AGConv, PointMLP), achieving mIoU OA values 94.90% 97.43%, respectively. On basis results, 4 phenotypic parameters (i.e., height, leaf length, width, area) extracted through K-means clustering method; these very close ground truth,

Language: Английский

---

A novel BH3DNet method for identifying pine wilt disease in Masson pine fusing UAS hyperspectral imagery and LiDAR data

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 134, P. 104177 - 104177

Published: Sept. 27, 2024

Language: Английский

---