University Of The Philippines Open University Master's Research Proposal PDF

UNIVERSITY OF THE PHILIPPINES OPEN UNIVERSITY MASTER OF ENVIRONMENT AND NATURAL RESOURCES MANAGEMENT COASTAL RESOURCES MANAGEMENT DEXTER K. DAG-UMAN "Mapping the underwater forest: A deep learning approach to seagrass mapping distribution in Calatagan, Batangas, Philippines, using Sentinel-2 sa...

UNIVERSITY OF THE PHILIPPINES OPEN UNIVERSITY MASTER OF ENVIRONMENT AND NATURAL RESOURCES MANAGEMENT COASTAL RESOURCES MANAGEMENT DEXTER K. DAG-UMAN "Mapping the underwater forest: A deep learning approach to seagrass mapping distribution in Calatagan, Batangas, Philippines, using Sentinel-2 satellite imagery" Dr. Consuelo Habito Faculty of Management and Development Studies I. **INTRODUCTION** A. Background of the study (A summary of Review of Related Literature) Monitoring seagrass distribution and health is essential for effective conservation and management efforts. Traditional methods for seagrass mapping, such as field surveys and aerial photography, are time-consuming, labor-intensive, and often limited in spatial coverage. Remote sensing techniques, particularly satellite imagery, have emerged as a promising alternative for large-scale and cost-effective seagrass monitoring. Continue............. B. Rationale/Significance of the study - **Seagrass Importance:** Seagrass meadows are crucial coastal ecosystems that provide numerous benefits, including habitat for diverse marine life, shoreline protection, carbon sequestration, and water quality improvement. Continue.......... - **Monitoring Challenges:** Traditional methods for seagrass mapping, such as direct underwater surveys, are time-consuming, labor-intensive, and often limited in spatial coverage. Continue....... - **Deep Learning Potential:** Deep learning techniques, particularly convolutional neural networks, have shown promise in image analysis and object recognition, making them suitable for mapping seagrass distribution from satellite imagery. Continue........... - **Sentinel-2 Data:** Sentinel-2 satellites provide high-resolution multispectral imagery, which can capture subtle variations in water color and vegetation characteristics that are indicative of seagrass presence. - **Improved Mapping Accuracy:** Deep learning models can potentially achieve higher accuracy in seagrass mapping compared to traditional methods, providing more detailed and reliable information on seagrass distribution. Continue............ - **Cost-Effective Monitoring:** Using satellite imagery for seagrass mapping is more cost-effective and efficient than traditional methods, enabling large-scale monitoring efforts. Continue.................. - **Temporal Monitoring:** Sentinel-2 imagery provides frequent data acquisition, allowing for the monitoring of seagrass changes over time, providing insights into their health and resilience to environmental pressures. Continue.................. - **Conservation and Management:** The results of this study can contribute to informed conservation and management decisions for seagrass meadows in Calatagan, Batangas, and potentially other areas in the Philippines. C. Objectives 1. **Develop a Deep learning Model:** The primary objective is to develop a robust deep learning model, specifically a convolutional neural network (CNN), capable of accurately identifying and classifying seagrass areas within Sentinel-2 satellite imagery. This model will be trained on a dataset of labeled images, allowing it to learn the unique spectral and spatial characteristics of seagrass meadows. 2. **Map Seagrass Distribution:** The developed deep learning model will then be applied to a large dataset of Sentinel-2 imagery covering the study area in Calatagan, Batangas. This application will generate a detailed map of seagrass distribution, highlighting the extent and location of seagrass meadows within the study area. 3. **Assess Model Accuracy:** To ensure the reliability of the generated seagrass map, the study will rigorously assess the accuracy of the deep learning model. This will involve comparing the model's predictions with ground truth data, collected through traditional methods like underwater surveys, field observations and very high-resolution satellite imageries (Google Earth). 4. **Explore Temporal Monitoring:** The study aims to explore the potential of using Sentinel-2 imagery, with its frequent data acquisition, for temporal monitoring of seagrass meadows. This will involve analyzing changes in seagrass distribution over time, providing insights into factors influencing seagrass health and resilience. - **Improved Seagrass Monitoring:** The study's objectives directly address challenges associated with traditional seagrass mapping methods, aiming to provide a more accurate, efficient and cost-effective approach to monitoring these crucial ecosystems. - **Data-Driven Conservation:** By generating detailed and timely maps of seagrass distribution, the study provides valuable data for conservation and management efforts, enabling informed decision-making regarding habitat protection and resto ration. - **Understanding Seagrass Dynamics:** The exploration of temporal monitoring using Sentinel-2 imagery offers the potential to gain a deeper understanding of seagrass dynamics, including their response to environmental changes and human activities. II. **METHODOLOGY (with subtitles)** 1. - Sentinel-2 Satellite Imagery: - Data Source: Access Sentinel-2 data through the Copernicus Open Access Hub () or other data portals. - Selection: Identify and download Sentinel-2 satellite imagery covering the study area in Calatagan, Batangas, Philippines. Sentinel-2 offers high-resolution multispectral imagery (10-meter resolutions for visible and near-infrared bands) suitable for seagrass mapping. - Temporal Coverage: Choose imagery from different dates to capture seasonal variations in seagrass cover and assess potential changes over time. - Pre-process: Apply processing and image enhancement of Sentinel-2 data (such as image composite and clipping). - Ground Truth Data: - Existing Data: This research will use the existing seagrass maps and datasets from various government agencies like NAMRIA within the Calatagan area. 2. **Data Pre-processing:** - Spectral Band Selection: Choose relevant spectral bands from Sentinel-2 imagery that are sensitive to seagrass presence. This includes bands in the visible, near-infrared, and shortwave infrared regions. - Spectral Indices: Calculate vegetation indices (e.g., Normalized Difference Vegetation Index - NDVI, Normalized Difference Water Index -- NDWI) to enhance the contrast between seagrass and other features. - Image Pre-process: Apply image processing and enhancement of Sentinel-2 data (e.g., Composite and Clipping). III. **REFERENCES**