Tree Canopy Volume Extraction Fusing ALS and TLS Based on Improved PointNeXt

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(14), P. 2641 - 2641

Published: July 19, 2024

Canopy volume is a crucial biological parameter for assessing tree growth, accurately estimating forest Above-Ground Biomass (AGB), and evaluating ecosystem stability. Airborne Laser Scanning (ALS) Terrestrial (TLS) are advanced precision mapping technologies that capture highly accurate point clouds digitization studies. Despite advances in calculating canopy volume, challenges remain extracting the removing gaps. This study proposes extraction method based on an improved PointNeXt model, fusing ALS TLS cloud data. In this work, first utilized to extract canopy, enhancing accuracy mitigating under-segmentation over-segmentation issues. To effectively calculate divided into multiple levels, each projected xOy plane. Then, Mean Shift algorithm, combined with KdTree, employed remove gaps obtain parts of real canopy. Subsequently, convex hull algorithm area part, sum areas all multiplied by their heights yields volume. The proposed method’s performance tested dataset comprising poplar, willow, cherry trees. As result, model achieves mean intersection over union (mIoU) 98.19% test set, outperforming original 1%. Regarding algorithm’s Root Square Error (RMSE) 0.18 m3, high correlation observed between predicted volumes, R-Square (R2) value 0.92. Therefore, efficiently acquires providing stable technical reference biomass statistics.

Language: Английский

---

Predicting Fine Dead Fuel Load of Forest Floors Based on Image Euler Numbers

Forests, Journal Year: 2024, Volume and Issue: 15(4), P. 726 - 726

Published: April 21, 2024

The fine dead fuel load on forest floors is the most critical classification feature in description systems, affecting several parameters manifestation of wildland fires. An accurate determination this contributes substantially to effective fire prevention, monitoring, and suppression. This study investigated viability incorporating image Euler numbers into predictive models via taking photos method. Pinus massoniana needles Quercus fabri broad leaves—typical types karst areas—served as research subjects. Accurate field data were collected Tianhe Mountain forests, China, while artificial fuelbeds differing loads constructed laboratory. Images captured uniformly digitized according various conversion thresholds. Thereafter, extracted, their relationship with was analyzed, applied generate three load-prediction based stepwise regression, nonlinear fitting, random algorithms. number had a significant both P. Q. loads. At low thresholds, negatively correlated load, whereas positive correlation recorded when threshold exceeded certain value. model showed best prediction performance, mean relative errors 9.35% 14.54% for fabri, respectively. fitting displayed next regression exhibited largest error, which significantly different from that model. first propose use features predict surface. results are more objective, accurate, time-saving than current estimates, benefiting scientific management

Language: Английский

---

Unoccupied aerial system (UAS) Structure-from-Motion canopy fuel parameters: Multisite area-based modelling across forests in California, USA

Remote Sensing of Environment, Journal Year: 2024, Volume and Issue: 312, P. 114310 - 114310

Published: July 28, 2024

There is a pressing need for well-informed management to reduce wildfire hazard and restore fire's beneficial ecological role in the Mediterranean- temperate-climate forests of California, USA. These efforts rely upon accessibility high spatial temporal resolution data on biomass canopy fuel parameters such as base height (CBH), mean height, bulk density (CBD), cover, leaf area index (LAI). Remote sensing using unoccupied aerial system Structure-from-Motion (UAS-SfM) presents promising technology this application due its accessibility, relatively low cost, possibility cadence. However, date, method has not been studied complex mosaic forest types found across California. In study we examined capacity structural multispectral information obtained from UAS-SfM, conjunction with machine learning methods, model aboveground an area-based approach multiple sites representing diversity Based correlations field measurements, separated into vertical (biomass, CBH, height) horizontal (LAI, CBD, cover) groups. UAS-SfM random models performed well modelling structure fuels (R2 0.69–0.75). exhibited strong performance comparison ALS, when transferred novel site. Vertical predictors were prominent these models, did improve addition spectral predictors. mainly used raster-based indices (primarily NDVI) had 0.49–0.59). addition, underperformed ALS poor applied When region widespread coverage, both groups successfully produced contiguous maps that could be fire behavior or decision making monitoring. findings indicate without sensors, suited mapping vertical-structure diverse landscapes supporting wide range types. contrast, identification variables suggests potential multi- hyperspectral sensors high-resolution satellite imagery meeting needs.

Language: Английский

---

High-Resolution Mapping of Litter and Duff Fuel Loads Using Multispectral Data and Random Forest Modeling

Fire, Journal Year: 2024, Volume and Issue: 7(11), P. 408 - 408

Published: Nov. 7, 2024

Forest fuels are the core element of fire management; each fuel component plays an important role in behavior. Therefore, accurate determination their characteristics and spatial distribution is crucial. This paper introduces a novel method for mapping litter duff loads using data collected by unmanned aerial vehicles. The approach leverages very high-resolution multispectral analysis within machine learning framework to achieve precise detailed results. A set vegetation indices texture metrics derived from data, optimized “Variable Selection Using Random Forests” (VSURF) algorithm, were used train random forest (RF) models, enabling modeling maps loads. field campaign measure was conducted mixed natural protected area “Sierra de Quila”, Jalisco, Mexico, obtain reference calibration validation purposes. results revealed moderate coefficients between observed predicted with R2 = 0.32, RMSE 0.53 Mg/ha 0.38, 13.14 loads, both significant p-values 0.018 0.015 respectively. Moreover, relative root mean squared errors 33.75% 27.71% bias less than 5% 20% coherent structure vegetation, despite high complexity study area. Our allows us estimate continuous aligned ecological context, which dictates dynamics variability. achieved acceptable accuracy monitoring providing researchers managers timely expedite decision-making management.

Language: Английский

---

Fuel Load Models for Different Tree Vegetation Types in Sichuan Province Based on Machine Learning

Forests, Journal Year: 2024, Volume and Issue: 16(1), P. 42 - 42

Published: Dec. 29, 2024

(1) Objective: To improve forest fire prevention, this study provides a reference for risk assessment in Sichuan Province. (2) Methods: This research focuses on various vegetation types Given data from 6848 sample plots, five machine learning models—random forest, extreme gradient boosting (XGBoost), k-nearest neighbors, support vector machine, and stacking ensemble (Stacking)—were employed. Bayesian optimization was utilized hyperparameter tuning, resulting models predicting fuel loads (FLs) across different types. (3) Results: The FL model incorporates not only characteristics but also site conditions climate data. Feature importance analysis indicated that structural factors (e.g., canopy closure, diameter at breast height, tree height) dominated cold broadleaf, subtropical mixed forests, while mean annual temperature seasonality) were more influential coniferous forests. Machine learning-based outperform the multiple stepwise regression both fitting ability prediction accuracy. XGBoost performed best coniferous, with coefficient of determination (R2) values 0.79, 0.85, 0.81, 0.83, respectively. Stacking excelled achieving an R2 value 0.82. (4) Conclusions: establishes theoretical foundation capacity It is recommended be applied to predict broadleaf suggested FLs Furthermore, offers management, assessment, prevention control

Language: Английский

---