This study contributes and focuses on identifying trash materials found in Pasig River, mainly Manila City from Santa Ana to Intramuros. The proponents will use Scaled-YOLOv4, which is a state-of-the-art object detection model that will be trained to specifically detect plastic and paper.
The main concern of this study is to test the performance of Scaled-YOLOv4-CSP configuration in detecting paper and plastic found on the surface of Pasig River, Philippines. Specifically, the study aims to answer the following questions:
1. How will the dilapidated trash dataset affect the performance of the Scaled-YOLOv4 model in detecting objects?
2. What will the Scaled-YOLOv4 model’s two classes yield in terms of precision?
3. How will the environment hinder the Scaled-YOLOv4 model in detecting objects?
In Pasig City, Philippines there is a river ferry service which is the only water-based transportation in Metro Manila that spans from Pinagbuhatan Pasig City, Mandaluyong City, Makati City, and Intramuros Manila City. It is owned by a private company called SCC Nautical Transport Services Incorporated. It is more similar to a water taxi than a ferry, and other water vessels also use the Pasig River as a route for transportation.
In previous years, Pasig river is known for being one of the most polluted rivers in the Philippines. Trashes in the water surface cause major damages to water vessels passing through bodies of water. The damage done to these water vessels cost a lot of money in repairing it. The river is now clean due to the recent efforts of the Pasig River Rehabilitation Commission. Having a clean river doesn’t mean that it will stay like that forever, a viable solution is having an automated trash detection on the surface of the water. These can be a way to minimize water debris and pollution. The detected trash can be documented on a weekly basis and notify the local government if there's an alarming increase in trash on the surface of the water.
Table of Contents
1. Introduction
1.1 Statement of the Problem
1.2 Objectives of the Study
1.3 Significance of the Study
1.4 Scope and Limitation
2. Related Works
2.1 Deep Learning
2.2 YOLO (You Only Look Once)
2.3 Real Time Object Detection
2.4 Damages to Boats caused by Water Pollution
2.5 Automated Waste Sorting
3. Methodology
3.1 Machine Specifications
3.2 Theoretical Framework
3.3 Conceptual Framework
3.4 Background of the COCO Dataset
3.5 Proposed Method
3.6 Trash Detection
3.7 Architecture of the Model
3.8 Data Gathering
3.9 Dataset
3.10 Data Augmentation
3.11 Training the Model
3.12 Validating the Model
3.13 Testing the Model
4. Results and Discussion
4.1 Objectives
4.2 Training Graph
4.3 Results Analysis
4.4 Test Images with Inferences
4.5 Discussion Summary
5. Conclusion and Future Work
5.1 Interpretation of Findings and Conclusions
5.2 Future Work
Research Objectives and Focus
This study aims to assess the performance of the Scaled-YOLOv4-CSP object detection model in identifying plastic and paper waste on the surface of the Pasig River, Philippines, using a self-curated, dilapidated trash dataset to address environmental pollution concerns.
- Testing the effectiveness of the Scaled-YOLOv4-CSP configuration for detecting specific waste categories.
- Developing a specialized dataset of 1,000 images of river debris for model training.
- Evaluating model precision, recall, and detection speed under real-world environmental conditions.
- Investigating the impact of environmental factors and trash deformation on detection accuracy.
- Providing data-driven insights for future automated water waste management systems.
Excerpt from the Book
1.1 Statement of the Problem
The main concern of this study is to test the performance of Scaled-YOLOv4-CSP configuration in detecting paper and plastic found on the surface of Pasig River, Philippines. Specifically, the study aims to answer the following questions:
1. How will the dilapidated trash dataset affect the performance of the Scaled-YOLOv4 model in detecting objects?
2. What will the Scaled-YOLOv4 model’s two classes yield in terms of precision?
3. How will the environment hinder the Scaled-YOLOv4 model in detecting objects?
Summary of Chapters
1. Introduction: Presents the background of water pollution in Pasig River and outlines the study's objective to implement a Scaled-YOLOv4 model for automated trash detection.
2. Related Works: Reviews existing literature on deep learning, YOLO architectures, real-time object detection, and the physical dangers posed by water pollution to marine vessels.
3. Methodology: Details the technical implementation, including GPU specifications, architectural configuration of Scaled-YOLOv4-CSP, data collection, augmentation techniques, and training procedures.
4. Results and Discussion: Analyzes the experimental outcomes, evaluating the model's precision, recall, and performance on test images while discussing the impact of environmental interference.
5. Conclusion and Future Work: Summarizes the effectiveness of the proposed system and suggests potential future applications such as integration with CCTV, drones, or autonomous waste-collecting robots.
Keywords
Scaled-YOLOv4, Object Detection, Trash Detection, Pasig River, Computer Vision, Deep Learning, Plastic Waste, Paper Waste, Machine Learning, Artificial Intelligence, Image Processing, Dilapidated Trash, Model Precision, Environmental Protection, Automated Surveillance
Frequently Asked Questions
What is the primary objective of this research?
The study aims to test and evaluate the performance of the Scaled-YOLOv4-CSP model in detecting plastic and paper waste floating on the surface of the Pasig River in the Philippines.
What are the core themes addressed in the paper?
The paper focuses on environmental protection through computer vision, the efficiency of deep learning object detectors in challenging real-world environments, and the challenge of detecting non-uniform, dilapidated trash.
What specific methodology does the research employ?
The authors use the Scaled-YOLOv4-CSP configuration, training it on a custom-built dataset of 1,000 images, which are further expanded using rotation-based data augmentation techniques.
What does the main body of the work cover?
It provides a comprehensive overview of existing object detection algorithms, the specific architecture of the proposed system, the data gathering process, training configurations, and a detailed analysis of detection results.
What are the key findings regarding model performance?
The model achieved an average precision of 63%, with the paper category performing at 67% precision and the plastic category at 59% precision, confirming its viability for use in real-time monitoring.
Which keywords best describe this study?
Key terms include Scaled-YOLOv4, Object Detection, Trash Detection, Pasig River, Computer Vision, and Deep Learning.
How does water deformation affect the detection process?
Deformation changes the object's shape and size over time, which makes it difficult for the model to match the extracted features against the training data, ultimately reducing precision.
What are the potential practical applications mentioned in the conclusion?
The authors propose that the system could be integrated into river-based CCTV systems, boat dashboard cameras, or autonomous drones to monitor and report pollution levels in real-time.
What role does the environment play in model performance?
Environmental factors such as sunlight reflection, occlusion by water lilies, and variable illumination significantly hinder the model's ability to consistently identify trash objects.
- Arbeit zitieren
- Timothy Kyle Chan (Autor:in), 2021, Trash Detection for Computer Vision. Using Scaled-YOLOv4 on Water Surface, München, GRIN Verlag, https://www.grin.com/document/1151588
