A Novel Fusion Perception Algorithm of Tree Branch/Trunk and Apple for Harvesting Robot Based on Improved YOLOv8s

Agronomy, Год журнала: 2024, Номер 14(9), С. 1895 - 1895

Опубликована: Авг. 24, 2024

Aiming to accurately identify apple targets and achieve segmentation the extraction of branch trunk areas trees, providing visual guidance for a picking robot actively adjust its posture avoid trunks obstacle avoidance fruit picking, spindle-shaped which are widely planted in standard modern orchards, were focused on, an algorithm tree detection robots was proposed based on improved YOLOv8s model design. Firstly, image data trees orchards collected, annotations object pixel-level conducted data. Training set then augmented improve generalization performance algorithm. Secondly, original network architecture’s design by embedding SE module attention mechanism after C2f Backbone architecture. Finally, dynamic snake convolution embedded into Neck structure architecture better extract feature information different branches. The experimental results showed that can effectively recognize images segment branches trunks. For recognition, precision 99.6%, recall 96.8%, mAP value 98.3%. 81.6%. compared with YOLOv8s, YOLOv8n, YOLOv5s algorithms recognition test images. other three algorithms, increased 1.5%, 2.3%, 6%, respectively. 3.7%, 15.4%, 24.4%, fruits, branches, is great significance ensuring success rate harvesting, provide technical support development intelligent harvesting robot.

Язык: Английский

---

Optimization of Sassafras tzumu leaves color quantification with UAV RGB imaging and Sassafras-net

Information Processing in Agriculture, Год журнала: 2025, Номер unknown

Опубликована: Фев. 1, 2025

Язык: Английский

---

Forestry Segmentation Using Depth Information: A Method for Cost Saving, Preservation, and Accuracy

Forests, Год журнала: 2025, Номер 16(3), С. 431 - 431

Опубликована: Фев. 27, 2025

Forests are critical ecosystems, supporting biodiversity, economic resources, and climate regulation. The traditional techniques applied in forestry segmentation based on RGB photos struggle challenging circumstances, such as fluctuating lighting, occlusions, densely overlapping structures, which results imprecise tree detection categorization. Despite their effectiveness, semantic models have trouble recognizing trees apart from background objects cluttered surroundings. In order to overcome these restrictions, this study advances management by integrating depth information into the YOLOv8 model using FinnForest dataset. Results show significant improvements accuracy, particularly for spruce trees, where mAP50 increased 0.778 0.848 mAP50-95 0.472 0.523. These findings demonstrate potential of depth-enhanced limitations RGB-based segmentation, complex forest environments with structures. Depth-enhanced enables precise mapping species, health, spatial arrangements, habitat analysis, wildfire risk assessment, sustainable resource management. By addressing challenges size, distance, lighting variations, approach supports accurate monitoring, improved conservation, automated decision-making forestry. This research highlights transformative integration models, laying a foundation broader applications environmental conservation. Future studies could expand dataset diversity, explore alternative technologies like LiDAR, benchmark against other architectures enhance performance adaptability further.

Язык: Английский

---

Detección Automática De Palmas Ceroxylon Mediante Aprendizaje Profundo En Un Área Protegida Del Amazonas (No Perú)

Опубликована: Янв. 1, 2025

---

Refined Classification of Mountainous Vegetation Based on Multi-Source and Multi-Temporal High-Resolution Images

Forests, Год журнала: 2025, Номер 16(4), С. 707 - 707

Опубликована: Апрель 21, 2025

Distinguishing vegetation types from satellite images has long been a goal of remote sensing, and the combination multi-source multi-temporal sensing for classification is currently hot topic in field. In species-rich mountainous environments, this study selected four different seasons (two aerial images, one WorldView-2 image, UAV image) proposed method integrating hierarchical extraction object-oriented approaches 11 types. This innovatively combines Random Forest algorithm with decision tree model, constructing strategy based on feature combinations to progressively address challenge distinguishing similar spectral characteristics. Compared traditional single-temporal methods, our approach significantly enhances accuracy through fusion comparative experimental validation, offering novel technical framework fine-grained under complex land cover conditions. To validate effectiveness features, we additionally performed classifications individual images. The results indicate that (1) classification, best performance was achieved autumn reaching an overall 72.36%, while spring had worst performance, only 58.79%; (2) features reached 89.10%, which improvement 16.74% compared (autumn). Notably, producer species such as Quercus acutissima Carr., Tea plantations, Camellia sinensis (L.) Kuntze, Pinus taeda L., Phyllostachys spectabilis C.D.Chu et C.S.Chao, thunbergii Parl., Castanea mollissima Blume all exceeded 90%, indicating relatively ideal outcome.

Язык: Английский

---

The Role of RPAS in Vegetation Height Estimation: Challenges and Future Perspectives in the Forestry Context

Current Forestry Reports, Год журнала: 2025, Номер 11(1)

Опубликована: Апрель 22, 2025

Язык: Английский

---

Precise identification of individual tree species in urban areas with high canopy density by multi-sensor UAV data in two seasons

International Journal of Digital Earth, Год журнала: 2025, Номер 18(1)

Опубликована: Апрель 25, 2025

Язык: Английский

---

Geospatial Data and Google Street View Images for Monitoring Kudzu Vines in Small and Dispersed Areas

Earth, Год журнала: 2025, Номер 6(2), С. 40 - 40

Опубликована: Май 13, 2025

Comprehensive reviews of continuously vegetated areas to determine dispersed locations invasive species require intensive use computational resources. Furthermore, effective mechanisms aiding identification specific approaches relying on geospatial indicators and ancillary images. This study develops a two-stage data workflow for the Kudzu vine (Pueraria montana) often found in small along roadsides. The INHABIT database from United States Geological Survey (USGS) provided vines Google Street View (GSV) set Stage one built up images be implemented an object detection technique, You Only Look Once (YOLO v8s), training, validating, testing. two defined dataset confirmed which was followed retrieve GSV analyzed with YOLO v8s. effectiveness v8s model assessed identified georeferenced demonstrated that field observations can virtually conducted by integrating images; however, its potential is confined updated periodicity or similar services.

Язык: Английский

---

Tree Species Detection and Enhancing Semantic Segmentation Using Machine Learning Models with Integrated Multispectral Channels from PlanetScope and Digital Aerial Photogrammetry in Young Boreal Forest

Remote Sensing, Год журнала: 2025, Номер 17(11), С. 1811 - 1811

Опубликована: Май 22, 2025

The precise identification and classification of tree species in young forests during their early development stages are vital for forest management silvicultural efforts that support growth renewal. However, achieving accurate geolocation through field-based surveys is often a labor-intensive complicated task. Remote sensing technologies combined with machine learning techniques present an encouraging solution, offering more efficient alternative to conventional methods. This study aimed detect classify using remote imagery techniques. mainly involved two different objectives: first, detection the latest version You Only Look Once (YOLOv12), second, semantic segmentation (classification) random forest, Categorical Boosting (CatBoost), Convolutional Neural Network (CNN). To best our knowledge, this marks first exploration utilizing YOLOv12 identification, along integrates digital aerial photogrammetry Planet achieve forests. used datasets: RGB from unmanned vehicle (UAV) ortho photography RGB-NIR PlanetScope. For YOLOv12-based detection, only was used, while performed three sets data: (1) Ortho (3 bands), (2) + canopy height model (CHM) (8 (3) CHM 12 vegetation indices (20 bands). With models applied these datasets, nine were trained tested 57 images (1024 × 1024 pixels) corresponding mask tiles. achieved 79% overall accuracy, Scots pine performing (precision: 97%, recall: 92%, mAP50: mAP75: 80%) Norway spruce showing slightly lower accuracy 94%, 82%, 90%, 71%). segmentation, CatBoost 20 bands outperformed other models, 85% 80% Kappa, 81% MCC, CHM, EVI, NIRPlanet, GreenPlanet, NDGI, GNDVI, NDVI being most influential variables. These results indicate simple boosting like can outperform complex CNNs

Язык: Английский

---

ICESat-2 Performance for Terrain and Canopy Height Retrieval in Complex Mountainous Environments

Remote Sensing, Год журнала: 2025, Номер 17(11), С. 1897 - 1897

Опубликована: Май 30, 2025

Accurate estimation of forest canopy height and understory terrain in mountainous regions is crucial for carbon stock assessment under the Paris Agreement but remains challenging. This study aimed to evaluate ICESat-2’s performance these complex environments. To achieve this, ICESat-2 ATL03 Version 6 photon data were processed using a novel adaptive DBSCAN algorithm (BDT-ADBSCAN) Pu’er City, China, biodiversity hotspot, results validated against airborne LiDAR. achieved high retrieval accuracy (R2 = 1.00, RMSE 0.91 m), primarily affected by slope, while was less accurate 0.53, 6.45 m) with systematic underestimation, mainly influenced itself. Nighttime strong-beam acquisitions substantially improved accuracies both products. research demonstrates viability high-resolution digital modeling provides quality control thresholds structure challenging regions, addressing validation gaps Asian hotspots supporting monitoring UN Sustainable Development Goals.

Язык: Английский

---