Design of tiny object detection algorithm in aerial imagery using MMdetection
The effectiveness of object detection techniques has been greatly enhanced through the development of deep learning algorithms. Nevertheless, conventional horizontal bounding box object detection algorithms fall short in specialized scenarios such as aerial images with dense, tiny, and oriented obj...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/172237 |
| _version_ | 1826110190172241920 |
|---|---|
| author | Liu, Haoran |
| author2 | Wen Bihan |
| author_facet | Wen Bihan Liu, Haoran |
| author_sort | Liu, Haoran |
| collection | NTU |
| description | The effectiveness of object detection techniques has been greatly enhanced through
the development of deep learning algorithms. Nevertheless, conventional horizontal bounding box object detection algorithms fall short in specialized scenarios such as aerial images with dense, tiny, and oriented objects. This dissertation aims to investigate enhancing the performance concerning detecting densely
clustered objects of small size in aerial imagery, utilizing oriented object detection algorithms. As an extension of horizontal object detection, the oriented
information of detection network excels in handling detection tasks involving
many densely arranged and arbitrarily oriented targets.
The research can be divided into three main contents. Initially, it envelops the
significance and widespread application demand for oriented object detection, especially in the realm of aerial imagery. Following this, a comparative analysis
of some of the existing benchmarks of oriented object detection algorithms is
conducted on notable optical datasets like DOTA v2.0 and SAR datasets including HRSID and SSDD. This experiment provides a substantial reference
for selecting foundational algorithms for subsequent optimizations. Ultimately,
the research pivots towards the optimization based on R3Det, focusing on enhancing the feature fusion network, adjusting the stride of detection head, and
embedding attention mechanisms within the backbone to elevate the perception
capacity towards tiny objects in aerial vistas.
Keywords: Tiny object detection; oriented object detection; aerial images. |
| first_indexed | 2024-10-01T02:30:21Z |
| format | Thesis-Master by Coursework |
| id | ntu-10356/172237 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T02:30:21Z |
| publishDate | 2023 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1722372023-12-08T15:43:12Z Design of tiny object detection algorithm in aerial imagery using MMdetection Liu, Haoran Wen Bihan School of Electrical and Electronic Engineering bihan.wen@ntu.edu.sg Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision The effectiveness of object detection techniques has been greatly enhanced through the development of deep learning algorithms. Nevertheless, conventional horizontal bounding box object detection algorithms fall short in specialized scenarios such as aerial images with dense, tiny, and oriented objects. This dissertation aims to investigate enhancing the performance concerning detecting densely clustered objects of small size in aerial imagery, utilizing oriented object detection algorithms. As an extension of horizontal object detection, the oriented information of detection network excels in handling detection tasks involving many densely arranged and arbitrarily oriented targets. The research can be divided into three main contents. Initially, it envelops the significance and widespread application demand for oriented object detection, especially in the realm of aerial imagery. Following this, a comparative analysis of some of the existing benchmarks of oriented object detection algorithms is conducted on notable optical datasets like DOTA v2.0 and SAR datasets including HRSID and SSDD. This experiment provides a substantial reference for selecting foundational algorithms for subsequent optimizations. Ultimately, the research pivots towards the optimization based on R3Det, focusing on enhancing the feature fusion network, adjusting the stride of detection head, and embedding attention mechanisms within the backbone to elevate the perception capacity towards tiny objects in aerial vistas. Keywords: Tiny object detection; oriented object detection; aerial images. Master of Science (Computer Control and Automation) 2023-12-04T05:32:26Z 2023-12-04T05:32:26Z 2023 Thesis-Master by Coursework Liu, H. (2023). Design of tiny object detection algorithm in aerial imagery using MMdetection. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/172237 https://hdl.handle.net/10356/172237 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Liu, Haoran Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title | Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title_full | Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title_fullStr | Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title_full_unstemmed | Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title_short | Design of tiny object detection algorithm in aerial imagery using MMdetection |
| title_sort | design of tiny object detection algorithm in aerial imagery using mmdetection |
| topic | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision |
| url | https://hdl.handle.net/10356/172237 |
| work_keys_str_mv | AT liuhaoran designoftinyobjectdetectionalgorithminaerialimageryusingmmdetection |