Building Extraction and Floor Area Estimation at the Village Level in Rural China Via a Comprehensive Method Integrating UAV Photogrammetry and the Novel EDSANet

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(20), P. 5175 - 5175

Published: Oct. 16, 2022

Dynamic monitoring of building environments is essential for observing rural land changes and socio-economic development, especially in agricultural countries, such as China. Rapid accurate extraction floor area estimation at the village level are vital overall planning development intensive use “beautiful countryside” construction policy Traditional situ field surveys an effective way to collect information but time-consuming labor-intensive. Moreover, buildings usually covered by vegetation trees, leading incomplete boundaries. This paper proposes a comprehensive method perform village-level homestead combining unmanned aerial vehicle (UAV) photogrammetry deep learning technology. First, tackle problem complex surface feature scenes remote sensing images, we proposed novel Efficient Deep-wise Spatial Attention Network (EDSANet), which uses dual attention refinement aggregate multi-level semantics enhance accuracy extraction, high-spatial-resolution imagery. Qualitative quantitative experiments were conducted with newly built dataset (named Weinan dataset) different networks examine performance EDSANet model task extraction. Then, number floors each was estimated using normalized digital (nDSM) generated from UAV oblique photogrammetry. The entire rapidly calculated multiplying floors. case study Helan village, Shannxi province, results show that images 0.939 precision reached 0.949. primarily have two stories, their total 3.1 × 105 m2. survey verified nDSM 0.94; RMSE 0.243. workflow experimental highlight potential rapid efficient China, well worldwide.

Language: Английский

---

Identification of Streamside Landslides with the Use of Unmanned Aerial Vehicles (UAVs) in Greece, Romania, and Turkey

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(4), P. 1006 - 1006

Published: Feb. 11, 2023

The alleviation of landslide impacts is a priority since they have the potential to cause significant economic damage as well loss human life. Mitigation can be achieved effectively by using warning systems and preventive measures. development improved methodologies for analysis understanding landslides at forefront this scientific field. Identifying effective monitoring techniques (accurate, fast, low cost) pursued objective. Geographic Information Systems (GISs) remote sensing are utilized in order achieve goal. In study, four methodological approaches (manual delineation, segmentation process, two mapping models, specifically object-based image pixel-based (OBIA PBIA)) were proposed tested with use Unmanned Aerial Vehicles (UAVs) data methods showcase state evolution landslides. digital surface model (DSM)-based classification approach was also used support aforementioned approaches. This study focused on streamside research sites three different countries: Greece, Romania, Turkey. results highlight that areas OBIA-based classifications most similar (98%) our control (manual) all sites. landslides’ perimeters Lefkothea Chirlesti showed (93%), opposed Sirtoba site, where from not corroborated manual classification. Deposition extend beyond trees revealed DSM-based encouraging because methodology monitor accuracy high performance regions. Specifically, terrains difficult access surveyed UAVs their ability take aerial images. obtained provide framework unitary modern tools.

Language: Английский

---

Comparison of Individual Tree Height Estimated from LiDAR and Digital Aerial Photogrammetry in Young Forests

Sustainability, Journal Year: 2022, Volume and Issue: 14(7), P. 3720 - 3720

Published: March 22, 2022

Biomass stored in young forests has enormous potential for the reduction of fossil fuel consumption. However, to ensure long-term sustainability, measurement accuracy tree height is crucial forest biomass and carbon stock monitoring, particularly forests. Precise using traditional field measurements challenging time consuming. Remote sensing (RS) methods can, however, replace field-based inventory. In our study, we compare individual estimation from Light Detection Ranging (LiDAR) Digital Aerial Photogrammetry (DAP) with measurements. It should be noted, that there was a one-year temporal difference between LiDAR/DAP scanning. A total 130 trees (32 Scots Pine, 29 Norway Spruce, 67 Silver Birch, 2 Eurasian Aspen) were selected private south-east Finland. Statistical correlation based on paired t-tests analysis variance (ANOVA, one way) used measured different methods. Comparative results remote showed LiDAR had stronger higher pine (R2 = 0.86, bias 0.70, RMSE 1.44) birch 0.81, 1.56) than DAP, which values 0.71, 0.82, 2.13) 0.69, 1.19, 2.08) birch. The two very similar spruce: 0.83, 0.30, 1.17) DAP 0.44, 1.26). Moreover, highly significant, minimum error mean 0.79–0.98, MD 0.12–0.33, RMSD 0.45–1.67) all species. t-test suggested significant (p < 0.05) observation test output are not significantly spruce. Presumably, campaign reason these results. Additionally, ANOVA indicated overall means estimated We concluded utilization estimating possible acceptable comparable measurement. Hence, inventory can carried out or at level as an alternative approaches.

Language: Английский

---

MaiZaic: a robust end-to-end pipeline for mosaicking freely flown aerial video of agricultural fields

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

ABSTRACT Unmanned aerial vehicles (UAVs) are increasingly used for high throughput phenotyping. In principle, freely flown would permit real-time flexibility in identifying and scouting regions of interest. Mosaicking multiple images provides a resolution global image consumer-grade UAVs offer low cost, ease flying, excellent RGB cameras. The vehicles’ inaccurate telemetry complicates estimating the homographies between pairs frames, standard mosaicking approach. Moreover, errors accumulate during computation, distorting later portions mosaic. Finally, crop fields particularly challenging to mosaic because their planting is so regular plants similar, eliminating distinctive features that could guide mosaicking. We propose MaiZaic , an end-to-end pipeline dynamically samples video frames using optical flow, automates camera gimbal calibration, estimates with unsupervised convolutional neural network, detects shots among generates mini-mosaics. Together, these techniques significantly reduce output mosaics. Our deep learning model trained on comprehensive dataset comprising different flight trajectories, maize lines, growth stages, augmented illumination data. more accurate faster than ASIFT robust our earlier CorNet CorNetv2 . demonstrate ’s effectiveness generating mosaics imagery captured by freely-flown explore its generalizability. Core ideas agricultural UAVs. introduces novel algorithms efficiently choose calibrate, imagery. homography estimator, CorNetv3 14 times 8.59% accurare ASIFT. generalizes well mosaicks at objects, cameras, pilots. mini-mosaicking algorithm improves accuracy interrupting error accumulation.

Language: Английский

---

Application of UAV Photogrammetry and Multispectral Image Analysis for Identifying Land Use and Vegetation Cover Succession in Former Mining Areas

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(3), P. 405 - 405

Published: Jan. 24, 2025

Variations in vegetation indices derived from multispectral images and digital terrain models satellite imagery have been successfully used for reclamation hazard management former mining areas. However, low spatial resolution the lack of sufficiently detailed information on surface morphology restricted such studies to large sites. This study investigates application small, unmanned aerial vehicles (UAVs) equipped with sensors land cover classification monitoring. The UAVs bridges gap between large-scale remote sensing techniques terrestrial surveys. Photogrammetric orthoimages (RGB multispectral) obtained repeated mapping flights November 2023 May 2024 were combined an ALS-based reference model object-based image classification. collected data enabled differentiation natural forests areas affected by activities, as well identification variations density growth rates results confirm that small provide a versatile efficient platform classifying monitoring forested landslides.

Language: Английский

---

UAV-SfM Photogrammetry for Canopy Characterization Toward Unmanned Aerial Spraying Systems Precision Pesticide Application in an Orchard

Drones, Journal Year: 2025, Volume and Issue: 9(2), P. 151 - 151

Published: Feb. 18, 2025

The development of unmanned aerial spraying systems (UASSs) has significantly transformed pest and disease control methods crop plants. Precisely adjusting pesticide application rates based on the target conditions is an effective method to improve use efficiency. In orchard spraying, structural characteristics canopy are crucial for guiding system adjust parameters. This study selected mango trees as research sample evaluated differences between UAV photography with a Structure from Motion (SfM) algorithm airborne LiDAR in results extracting maximum height, projection area, volume parameters were extracted height model SfM (CHMSfM) (CHMLiDAR) by grids same width planting rows (5.0 m) 14 different heights (0.2 m, 0.3 0.4 0.5 0.6 0.8 1.0 2.0 3.0 4.0 5.0 6.0 8.0 10.0 m), respectively. Linear regression equations used fit obtained sensors. correlation was using R2 rRMSE, t-test (α = 0.05) employed assess significance differences. show that grid increases, values CHMSfM CHMLiDAR increase, while rRMSE decrease. When two models 92.85%, 0.0563. For 97.83%, 0.01, volume, 98.35%, 0.0337. exceeds three all greater than 0.05, accepting hypothesis there no significant difference Additionally, coordinates x0 intersection linear equation y=x reference, tends overestimate lower underestimate higher area compared CHMLiDAR. some extent reflects surface smoother. demonstrates effectiveness guide UASS variable-rate oblique combined algorithm.

Language: Английский

---

Accuracy Assessment and Optimization of the Photogrammetric Process Variables for 3D Mapping Using Unmanned Aerial Vehicle (Drones)

Journal of the Indian Society of Remote Sensing, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 22, 2025

Language: Английский

---

Influence of Ground Control Point Placement and Surrounding Environment on Unmanned Aerial Vehicle-Based Structure-from-Motion Forest Resource Estimation

Drones, Journal Year: 2025, Volume and Issue: 9(4), P. 258 - 258

Published: March 28, 2025

Ground control points (GCPs) are used in forest surveys employing unmanned aerial vehicle (UAV)-based structure from motion (SfM). In that context, the influence of surrounding environment on GCP placement requires further analysis. This study investigated effects and estimation information by UAV-SfM. Forest resource was performed using UAV (Inspire2) images SfM analysis (via Pix4Dmapper) under varying environmental conditions around GCPs within same stand. The results indicated had no significant effect processing, tree top extraction (the number extracted target trees 151 or 150), crown area (RMSEs ranged approximately 5 to 6.5 m2). However, when were placed open areas, height accuracy improved, without differences between estimated measured values (patterns A, B, D E, RMSEs 1.60 3.09 m; patterns C 5.69 7.92 m). These findings suggest UAV-SfM-based surveys, particularly for estimation, both GCPs, as well environment, crucial enhancing accuracy.

Language: Английский

---

The Synergistic Effects of GCPs and Camera Calibration Models on UAV-SfM Photogrammetry

Drones, Journal Year: 2025, Volume and Issue: 9(5), P. 343 - 343

Published: May 1, 2025

Previous studies have shown that the use of appropriate ground control points (GCPs) and camera calibration models can optimize photogrammetry. However, synergistic effects GCPs on UAV-SfM photogrammetry are still unknown. This study used with varying complexities under different GCP conditions (in terms number quality) for The correlation matrix root mean squared error (RMSE) were to analyze models. results show (1) without GCPs, complex reduce distortion parameter improve terrain modeling accuracy by about 70%, Model C (with F, Cx, Cy, K1–K4, P1–P4) being most widely applicable. (2) Increasing enhances more effectively than increasing model complexity, reducing RMSE 45–70%, while complexity does not affect required number. (3) A strong interaction exists between quality models: High-quality enhance performance, requirement quality. provides both theoretical insights practical guidance efficient low-cost in scenarios.

Language: Английский

---

Effects of Flight and Smoothing Parameters on the Detection of Taxus and Olive Trees with UAV-Borne Imagery

Drones, Journal Year: 2022, Volume and Issue: 6(8), P. 197 - 197

Published: Aug. 8, 2022

Recent technical and jurisdictional advances, together with the availability of low-cost platforms, have facilitated implementation unmanned aerial vehicles (UAVs) in individual tree detection (ITD) applications. UAV-based photogrammetry or structure from motion is an example such a technique, but requires detailed pre-flight planning order to generate desired 3D-products needed for ITD. In this study, we aimed find most optimal flight parameters (flight altitude image overlap) processing options (smoothing window size) taxus trees Belgium. Next, tested transferability developed marker-controlled segmentation algorithm by applying it delineation olive orchard Greece. We found that had larger effect on accuracy precision ITD than parameters. particular, smoothing 3 × pixels performed best (F-scores 0.99) compared no between 0.88 0.90) size 5 0.90 0.94). Furthermore, results show model can still be bottleneck as does not capture management induced characteristics typical crown shape 0.55 0.61).

Language: Английский

---