Farhan Nurzaman, - (2023) PENDETEKSIAN OBJEK PADA KONDISI PENCAHAYAAN MINIM MENGGUNAKAN YOLOV7 DAN LOW-LIGHT IMAGE ENHANCEMENT. S1 thesis, Universitas Pendidikan Indonesia.
![[img]](http://repository.upi.edu/style/images/fileicons/text.png) |
Text
S_KOM_1904908_Title.pdf Download (919kB) |
|
Text
S_KOM_1904908_Chapter1.pdf Download (58kB) |
|
Text
S_KOM_1904908_Chapter2.pdf Restricted to Staf Perpustakaan Download (1MB) |
|
Text
S_KOM_1904908_Chapter3.pdf Download (321kB) |
|
Text
S_KOM_1904908_Chapter4.pdf Restricted to Staf Perpustakaan Download (6MB) |
|
Text
S_KOM_1904908_Chapter5.pdf Download (76kB) |
Abstract
Object detection atau deteksi objek merupakan salah satu teknik yang penting dalam computer vision. Object detection merupakan suatu metode untuk mengidentifikasi objek, seperti manusia, binatang, atau kendaraan, dan letak objek-objek tersebut pada gambar digital. Salah satu tantangan pada object detection adalah pencahayaan yang minim pada gambar sehingga terjadi penurunan kualitas gambar. Untuk mengatasi permasalahan tersebut penelitian ini menggunakan dua metode low-light image enhancement untuk meningkatkan pencahayaan dari gambar, yaitu Zero-DCE dan LLFlow. Gambar yang diperoleh dari proses low-light image enhancement kemudian digunakan sebagai gambar input dari YOLOV7 untuk dilakukan proses pendeteksian objek. Dari hasil pengujian terhadap dataset ExDark, diperoleh nilai mAP@0,5 sebesar 0,785 untuk penggunaan YOLOV7 tanpa low-light image enhancement, 0,794 menggunakan Zero-DCE, dan 0,781 menggunakan LLFlow. Object detection is a vital technique in computer vision, as it involves classifying and locating various objects within digital images. One of the main challenges of object detection is lighting variation, such as low-light conditions. This paper aims to resolve challenges associated with low-light conditions by using two low-light image enhancement methods, namely Zero-DCE and LLFlow. These enhancements are used to enhance the lighting condition within the image from a low-light image to sufficient lighting condition. Images processed by low-light image enhancement are used as inputs for YOLOV7. By evaluating the trained model on the ExDark dataset, it produces mAP@0,5 values of 0,785 when YOLOV7 is used without any enhancement, 0,794 when combined with Zero-DCE, and 0,781 when combined with LLFlow.
| Item Type: | Thesis (S1) |
|---|---|
| Additional Information: | Google Scholar Pembimbing: https://scholar.google.co.id/citations?user=R20SJKYAAAAJ&hl=en https://scholar.google.co.id/citations?user=3YFkydUAAAAJ&hl=en ID SINTA Dosen Pembimbing: Yaya Wihardi - 5994413 Herbert Siregar - 5991008 |
| Uncontrolled Keywords: | object detection, low-light image enhancement, yolov7, zerodce, llflow, low-light object detection, normalizing flow |
| Subjects: | L Education > L Education (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
| Divisions: | Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Ilmu Komputer |
| Depositing User: | Farhan Nurzaman |
| Date Deposited: | 16 Sep 2023 13:42 |
| Last Modified: | 16 Sep 2023 13:42 |
| URI: | http://repository.upi.edu/id/eprint/101522 |
Actions (login required)
 |
View Item |