Faishal , Hariz Makaarim Gandadipoera (2024) IMPLEMENTASI CONVOLUTIONAL NEURAL NETWORK (CNN) UNTUK DETEKSI KERUSAKAN POHON BERBASIS MOBILE. FAKULTAS MATEMATIKA DAN ILMU PENGETAHUAN ALAM, UNIVERSITAS LAMPUNG.
|
File PDF
ABSTRAK - Faishal Hariz.pdf Download (3922Kb) | Preview |
|
![[img]](http://digilib.unila.ac.id/style/images/fileicons/text.png) |
File PDF
SKRIPSI FULL - Faishal Hariz.pdf Restricted to Hanya staf Download (4Mb) | Minta salinan |
|
|
File PDF
SKRIPSI TANPA BAB PEMBAHASAN - Faishal Hariz.pdf Download (4Mb) | Preview |
Abstrak (Berisi Bastraknya saja, Judul dan Nama Tidak Boleh di Masukan)
Kerusakan pohon merupakan salah satu parameter dari indikator vitalitas dalam metode Forest Health Monitoring (FHM). Saat ini, pengukuran kerusakan pohon dalam FHM masih dinilai secara manual. Disisi lain, seiring kemajuan teknologi terdapat suatu sistem yang dapat membantu pengidentifikasian suatu objek dengan metode Convolutional Neural Network (CNN). Oleh karena itu, penelitian ini bertujuan mengimplementasikan Convolutional Neural Network (CNN) untuk deteksi kerusakan pohon berbasis mobile. Tahapan penelitian meliputi Dataset, pre-processing, Candidate CNN model, Training and Validation CNN model, Converting model to mobile, dan Model deploy. Hasil penelitian menunjukkan bahwa model MobileNet mencapai akurasi 99,37% dan setelah dikonversi ke TensorFlow Lite Model memiliki akurasi yang sama. Model SSD MobileNet V2 mencapai 94,90% mAP, namun dengan augmentasi terjadi penurunan menjadi 92,44%. Penelitian ini juga mencakup pengujian aplikasi yang dikembangkan pada framework Flutter dengan metode black-box testing dan User Acceptance Testing, serta aplikasi Dgt Diagnosis yang dapat diunduh di Google Play Store. Penelitian ini berkesimpulan bahwa Convolutional Neural Network (CNN) dapat diimplementasikan pada aplikasi mobile untuk melakukan deteksi dan klasifikasi kerusakan pohon. Kata kunci: Forest Health Monitoring, Convolutional Neural Network, Mobile Device, Framework Flutter, MobileNet. Tensorflow Lite Model Tree damage is one of the parameters of the vitality indicator in the Forest Health Monitoring (FHM) method. Currently, the measurement of tree damage in FHM is still assessed manually. On the other hand, along with technological advances there is a system that can help identify an object with the Convolutional Neural Network (CNN) method. Therefore, this research aims to implement Convolutional Neural Network (CNN) for mobile-based tree damage detection. The research stages include Dataset, pre-processing, Candidate CNN model, Training and Validation CNN model, Converting model to mobile, and Model deploy. The results showed that the MobileNet model achieved 99.37% accuracy and after being converted to TensorFlow Lite Model had the same accuracy. The MobileNet V2 SSD model achieved 94.90% mAP, but with augmentation it decreased to 92.44%. This research also includes testing applications developed on the Flutter framework with black-box testing and User Acceptance Testing methods, as well as the Dgt Diagnosis application which can be downloaded on the Google Play Store. This research concludes that Convolutional Neural Network (CNN) can be implemented in mobile applications to detect and classify tree damage. Keywords: Forest Health Monitoring, Convolutional Neural Network, Mobile Device, Flutter Framework, MobileNet. Tensorflow Lite Model
| Jenis Karya Akhir: | Skripsi |
|---|---|
| Subyek: | 000 Ilmu komputer, informasi dan pekerjaan umum |
| Program Studi: | FAKULTAS MIPA > Prodi Ilmu Komputer |
| Pengguna Deposit: | UPT . Siswanti |
| Date Deposited: | 16 Apr 2025 04:13 |
| Terakhir diubah: | 16 Apr 2025 04:13 |
| URI: | http://digilib.unila.ac.id/id/eprint/86136 |
Actions (login required)
 |
Lihat Karya Akhir |
