Deep learning algorithm to detect tree defects via circular scans
This research aims to advance the accuracy and efficiency of tree defect detection and analysis by integrating Ground-penetrating Radar (GPR) with Deep Learning techniques. It delves into discussions on GPR-based techniques, circular scanning methods for tree imaging, and the application of Convolut...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/177109 |
| _version_ | 1811680095830540288 |
|---|---|
| author | Nguyen, Thanh Tin |
| author2 | Abdulkadir C. Yucel |
| author_facet | Abdulkadir C. Yucel Nguyen, Thanh Tin |
| author_sort | Nguyen, Thanh Tin |
| collection | NTU |
| description | This research aims to advance the accuracy and efficiency of tree defect detection and analysis by integrating Ground-penetrating Radar (GPR) with Deep Learning techniques. It delves into discussions on GPR-based techniques, circular scanning methods for tree imaging, and the application of Convolutional Neural Network (CNN) for deep learning.
The report provides detailed insights into the processing of dataset generation, encompassing simulation and measurement techniques. Nonetheless, this chapter includes the procedure for preprocessing and multiplying the dataset for different scenarios by applying edge detection and data augmentation techniques. It further indicates and describes the formulation of the U-net model architecture, a key component of the methodology.
For performance evaluation, accuracy metrics such as Structural Similarity Index (SSIM) and error metrics including Mean Squared Error (MSE), Mean Absolute Error (MAE), and Mean Relative Error (MRE) are properly indicated for performance evaluation. Furthermore, the concepts of examining the model quality globally or locally by the Canny edge detection is also introduced and explained.
In the subsequent chapter, numerical and graphical performance analyses and findings in experiment and model refining processes are provided, demonstrating the efficacy of the proposed approach in accurately detecting and characterizing tree defects.
The study's conclusions summarize the advancements achieved in tree defect detection, emphasizing the significance of the integrated approach. Finally, recommendations for future improvements are proposed. |
| first_indexed | 2024-10-01T03:19:36Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/177109 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:19:36Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1771092024-05-31T15:43:09Z Deep learning algorithm to detect tree defects via circular scans Nguyen, Thanh Tin Abdulkadir C. Yucel Lee Yee Hui School of Electrical and Electronic Engineering NParks acyucel@ntu.edu.sg, EYHLee@ntu.edu.sg Physics Tree trunk Deep learning CNN This research aims to advance the accuracy and efficiency of tree defect detection and analysis by integrating Ground-penetrating Radar (GPR) with Deep Learning techniques. It delves into discussions on GPR-based techniques, circular scanning methods for tree imaging, and the application of Convolutional Neural Network (CNN) for deep learning. The report provides detailed insights into the processing of dataset generation, encompassing simulation and measurement techniques. Nonetheless, this chapter includes the procedure for preprocessing and multiplying the dataset for different scenarios by applying edge detection and data augmentation techniques. It further indicates and describes the formulation of the U-net model architecture, a key component of the methodology. For performance evaluation, accuracy metrics such as Structural Similarity Index (SSIM) and error metrics including Mean Squared Error (MSE), Mean Absolute Error (MAE), and Mean Relative Error (MRE) are properly indicated for performance evaluation. Furthermore, the concepts of examining the model quality globally or locally by the Canny edge detection is also introduced and explained. In the subsequent chapter, numerical and graphical performance analyses and findings in experiment and model refining processes are provided, demonstrating the efficacy of the proposed approach in accurately detecting and characterizing tree defects. The study's conclusions summarize the advancements achieved in tree defect detection, emphasizing the significance of the integrated approach. Finally, recommendations for future improvements are proposed. Bachelor's degree 2024-05-27T01:40:16Z 2024-05-27T01:40:16Z 2024 Final Year Project (FYP) Nguyen, T. T. (2024). Deep learning algorithm to detect tree defects via circular scans. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/177109 https://hdl.handle.net/10356/177109 en B3002-231 application/pdf Nanyang Technological University |
| spellingShingle | Physics Tree trunk Deep learning CNN Nguyen, Thanh Tin Deep learning algorithm to detect tree defects via circular scans |
| title | Deep learning algorithm to detect tree defects via circular scans |
| title_full | Deep learning algorithm to detect tree defects via circular scans |
| title_fullStr | Deep learning algorithm to detect tree defects via circular scans |
| title_full_unstemmed | Deep learning algorithm to detect tree defects via circular scans |
| title_short | Deep learning algorithm to detect tree defects via circular scans |
| title_sort | deep learning algorithm to detect tree defects via circular scans |
| topic | Physics Tree trunk Deep learning CNN |
| url | https://hdl.handle.net/10356/177109 |
| work_keys_str_mv | AT nguyenthanhtin deeplearningalgorithmtodetecttreedefectsviacircularscans |