SISKA MIATI JATI NINGSIH, . (2024) PERHITUNGAN MUATAN ELEKTRON PADA EKSPERIMEN TETES MINYAK MILIKAN MENGGUNAKAN ALGORITMA YOLOv5. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.
Text
Cover, Abstrak, Lembar Pengesahan, Lembar Originalitas,.pdf Download (2MB) |
|
Text
Bab 1 Pendahuluan.pdf Download (212kB) |
|
Text
Bab 2 Landasan Teori.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
Text
Bab 3 Metodologi.pdf Restricted to Registered users only Download (329kB) | Request a copy |
|
Text
Bab 4 Hasil&Pembahasan.pdf Restricted to Registered users only Download (646kB) | Request a copy |
|
Text
Bab 5 Kesimpulan dan Saran.pdf Restricted to Registered users only Download (123kB) | Request a copy |
|
Text
Daftar Pustaka.pdf Download (214kB) |
|
Text
Lampiran dan Daftar Riwayat Hidup.pdf Restricted to Registered users only Download (2MB) | Request a copy |
Abstract
Eksperimen tetes minyak Milikan adalah salah satu metode fundamental dalam fisika untuk menentukan muatan elektron. Namun, pada eksperimen ini sering kali mengalami kesulitan dalam mengamati dan menganalisis pergerakan tetesan minyak secara akurat. Penelitian ini bertujuan untuk memanfaatkan teknologi deteksi objek berbasis YOLOv5 guna meningkatkan akurasi perhitungan muatan elektron pada eksperimen tetes minyak Milikan. Eksperimen dilakukan lima kali dengan variasi tegangan: 91,7V; 108,4V; 163,5V; 254,2V; dan 323,7V. Data citra dan video resolusi tinggi (2160 x 3840 piksel, 60 FPS) diambil menggunakan kamera ponsel, kemudian diekstraksi dan diaugmentasi menghasilkan 4072 frame gambar untuk pelatihan, pengujian, dan validasi model YOLOv5 dibagi dengan rasio 82:12:6. Proses anotasi dilakukan secara manual menggunakan platform Roboflow untuk memberi label pada setiap gambar. Pelatihan model dilakukan di platform Kaggle dengan parameter learning rate (0,001), optimizer (SDG), batch size (32), image size (840), dan epoch (300). Berdasarkan parameter yang telah digunakan, model dapat memperoleh nilai mAP sebesar 71%. Dengan model tersebut, didapatkan hasil terbaik pada tegangan 163,5V perhitungan muatan elektron terhadap data referensi (1,60218×10^(-19)) sebesar (1,65486×10^(-19)±5,90972×10^(-21) )C dengan kesalahan relatif 3,29%. Selain itu, hasil perhitungan menggunakan data pengamatan langsung menghasilkan 1,22248×10^(-19) C dengan kesalahan relatif sebesar 23,70% terhadap data referensi. Hasil penelitian ini menunjukkan bahwa algoritma YOLOv5 dapat diaplikasikan dalam eksperimen tetes minyak Milikan dengan akurasi dan efisiensi yang baik, meskipun masih diperlukan peningkatan untuk deteksi objek berukuran kecil. Kata kunci: Eksperimen tetes minyak Milikan, Deteksi objek, Muatan elektron, Pengolahan citra, YOLOv5 ***** The Millikan oil drop experiment is one of the fundamental methods in physics for determining electron charge. However, this experiment often faces difficulties in accurately observing and analyzing the movement of oil droplets. This research aims to utilize YOLOv5-based object detection technology to improve the accuracy of electron charge calculations in the Millikan oil drop experiment.The experiment was conducted five times with varying voltages: 91.7V, 108.4V, 163.5V, 254.2V, and 323.7V. High-resolution image and video data (2160 x 3840 pixels, 60 FPS) were captured using a smartphone camera, then extracted and augmented to produce 4072 frames for training, testing, and validating the YOLOv5 model, divided in a ratio of 82:12:6. The annotation process was done manually using the Roboflow platform to label each image. Model training was conducted on the Kaggle platform with parameters: learning rate (0.001), optimizer (SDG), batch size (32), image size (840), and epochs (300). Based on these parameters, the model achieved a mAP score of 71%. The best result with the model was obtained at a voltage of 163.5V, with an electron charge calculation relative to the reference data (1,60218×10^(-19)) of 1,65486×10^(-19) ±5,90972×10^(-21) C and a relative error of 3.29%. In addition, calculations using direct observation data resulted in 1,22248×10^(-19) C with a relative error of 23.70% compared to the reference data. The results of this study indicate that the YOLOv5 algorithm can be effectively applied in the Millikan oil drop experiment with good accuracy and efficiency, although improvements are still needed for detecting small-sized objects. Keywords: Millikan’s oil drop experiment, Object detection, Electron charge, Image processing, YOLOv5
Item Type: | Thesis (Sarjana) |
---|---|
Additional Information: | 1). Dr. Hadi Nasbey, S.Pd., M.Si. ; 2). Haris Suhendar, S.Si., M.Sc. |
Subjects: | Sains > Fisika |
Divisions: | FMIPA > S1 Fisika |
Depositing User: | Users 25044 not found. |
Date Deposited: | 12 Aug 2024 04:17 |
Last Modified: | 12 Aug 2024 04:17 |
URI: | http://repository.unj.ac.id/id/eprint/49536 |
Actions (login required)
View Item |