Short-lived climate pollutants (SLCPs) and municipal solid wastes (MSWs) have been found to be viable sources of clean energy. This study integrates the Intergovernmental Panel on Climate Change (IPCC) guidelines for methane flow rate estimation in the software Long-Range Energy Alternatives Planning-Integrated Benefit Calculator (LEAP-IBC) system to estimate and project the methane emissions coming from the waste generated by Metro Manila, disposed in sanitary landfills. It aims to analyze the environmental impacts of the emissions coming from the non-energy sector using the IBC feature of LEAP and by developing two scenarios with 2010 and 2050 as the base and end years: the baseline and methane recovery scenario, where the latter represents the solid waste management undertaken to counter the emissions. Under the baseline, 97.30 million metric tonnes of methane emissions are expected to be produced and are predicted to continuously increase. In the same scenario, the cities of Quezon, Manila, and Caloocan account for the biggest methane emissions. On the other hand, in the methane recovery scenario, the methane emissions are expected to have a decline of 36% from 127.036 to 81.303 million metric tonnes by 2025, 52% from 135.358 to 64.972 million metric tonnes by 2030, and 54% from 150.554 to 69.254 million metric tonnes by 2040. For the 40-year projection of the study under the 100-year global warming potential analysis, a total of 10,249 million metric tonnes of CO₂ equivalent is avoided in the methane recovery compared to the BAU, and a maximum of 0.019 °C temperature increase can also be avoided. Moreover, electricity costs without LFG technology increase from 2.21 trillion to 8.75 trillion, while costs with LFG technology also rise but remain consistently lower, ranging from 2.20 trillion to 8.74 trillion. This consistent reduction in electricity costs underscores the long-term value and importance of adopting LFG technology, even as its relative savings impact diminishes over time. Finally, the fixed effects and random effects panel data regression analysis reinforces and asserts that the solid waste management is really improved by means of the methane recovery technology, leading the methane emissions to decrease.

The maritime Search and Rescue (SAR) operation requires effective and accurate object detection systems capable of identifying various targets in dynamic sea environments and low-light situations. The paper presents a comparative study of the YOLOv10, YOLOv11, and YOLOv12 networks in multi-class marine detection using UAV images. The SeaDronesSee Odv2 dataset has been preprocessed using physics-based augmentation that mimics environmental changes, such as fog, noon, sunset, dawn, and cloudy scenarios. A multi-resolution tiling procedure was implemented to preserve the image consistency of small objects. Results show that YOLOv11s is the model that has the least accuracy-efficiency trade-off, with an mAP@0.5 of 0.888 and an F1-Score of 0.872 at a reasonable inference time. Precision-recall analysis has shown that large maritime objects were detected with high precision, while small objects were detected with average recall. The results show that multi-resolution preprocessing, as well as physicsbased augmentation, enhance the robustness and generalization of the network. Altogether, YOLOv11s is the most stable version to use in real-time maritime SAR missions with UAVs due to its ability to handle a variety of visual conditions.

Unmanned Aerial Vehicles (UAVs) are finding applications in marine Search and Rescue (SAR) operations due to their ability to cover a large geographical area safely and efficiently. Human detection in maritime scenarios is challenging due to scale variation, poor lighting and visibility conditions, and the limited computing resources available on UAVs. In this paper, we introduce YOLOv11 Star, a lightweight variant of YOLOv11 Nano optimized for detecting humans in UAV images. This approach features a multi-resolution tiling processing pipeline for handling high-resolution input images. It adopts architectural modifications that replace traditional C3K2 and C2PSA modules with the Starblock component, which reduces the model size and complexity. In evaluations on the SeaDronesSee Odv2 dataset, YOLOv11n-Star achieves a 53.7% (1.19M) parameter reduction and a 52% (2.53 MB) model size reduction over YOLOv11n baseline, with a competitive mAP@0.50 of 0.714 (compared to a baseline of 0.763). Performance can also be compared to YOLOv12n, YOLOv10n, and YOLOv9n. The current findings reveal that YOLOv11 Star has a strong balance between detection and computational efficiency, thus making it a potential solution to the problem of efficient real-time deployment of maritime Search and Rescue (SAR).

Accurate real-time monitoring of cattle behavior is essential for enabling data-driven decision-making in precision livestock farming. However, existing monitoring solutions often rely on cloud-based processing or high-power hardware, which are impractical for deployment in remote or low-infrastructure agricultural environments. There is a critical need for low-cost, energy-efficient, and autonomous sensing systems capable of operating independently at the edge. This paper presents a compact, sensor-integrated system for real-time cattle behavior monitoring using an embedded vision sensor and a TinyML-based inference pipeline. The system is designed for low-power deployment in field conditions and integrates the OV2640 image sensor with the Sipeed Maixduino platform, which features the Kendryte K210 RISC-V processor and an on-chip neural network accelerator (KPU). The platform supports fully on-device classification of cattle postures using a quantized convolutional neural network trained on the publicly available cattle behavior dataset, covering standing and lying behavioral states. Sensor data is captured via the onboard camera and preprocessed in real time to meet model input specifications. The trained model is quantized and converted into a K210-compatible. kmodel using the NNCase toolchain, and deployed using MaixPy firmware. System performance was evaluated based on inference latency, classification accuracy, memory usage, and energy efficiency. Results demonstrate that the proposed TinyML-enabled system can accurately classify cattle behaviors in real time while operating within the constraints of a low-power, embedded platform, making it a viable solution for smart livestock monitoring in remote or under-resourced environments.

Mangrove forests play a critical role in protecting tropical coastlines, storing carbon, and supporting marine life, yet mapping their fine-scale canopy structure remains difficult. This study introduces a structured ground-truthing strategy for automatic tree crown delineation of mangrove forests using high-resolution UAV imagery and deep learning. UAV-derived RGB orthomosaics from the Batangas coastline, Philippines, were manually annotated into three canopy classes: Individual Crowns (IC), Crown Clusters (CC), and Canopy Gaps (CG), which form the reference foundation for model training. A multi-head DeepLabV3–ResNet50 network was developed to jointly predict canopy masks and distance maps that capture crown geometry and spatial relationships. From these distance surfaces, local maxima were extracted as potential crown centers, and a Voronoi-style watershed segmentation was applied to delineate discrete crown units. An adaptive merging rule, based on peak prominence and spatial proximity, refined the segmentation by merging coalescent crowns while preserving distinct individuals. The results show interpretable delineation of crowns across varying canopy densities and environmental contexts, from isolated individuals in coastal fringes to coalescent crowns in mature stands. High-prominence peaks corresponded to dominant crowns, while moderate peaks represented subordinate or merging individuals. The framework integrates manual annotation with automation through deep learning models, establishing a reproducible foundation for crown-level mapping and long-term monitoring of mangrove canopy dynamics.

The use of Vegetation Indices (VIs) is important in remote sensing for inferring biophysical parameters and monitoring environmental processes. Since VIs are derived from reflectance measurements, the scale and platform of acquisition-satellite, commercial multispectral drone, or simulation model—can potentially influence their derived values and inter-relationships. This study addresses this influence by presenting a multiscale comparison of four common indices: MPRI, NDVI, NDWI, and OSAVI. Data were sourced from three distinct platforms: Sentinel-2 satellite imagery, high-resolution data acquired via a multispectral drone, and a simulated spectral library generated using the PROSPECT-4 via ARTMO software. The core objective was to assess the consistency of the correlational structure among these VIs across the three scale measurements. Results indicate a statistically significant similarity in the overall correlational pattern of the indices, suggesting that the intrinsic mathematical relationships between these VIs are largely scale-invariant. However, the study identifies a critical need to standardize the data acquisition and processing protocol for drone measurements to match the known consistency of the Sentinel-2 mission and the rigorous parameters of the PROSPECT-4 simulation. This standardization is essential for future multi-sensor integration and calibration efforts.