Mega, Salma Puspa (2019) Aplikasi Sonar untuk Mengidentifikasi Obstacle Menggunakan Metode Naive Bayes. Diploma thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
Salma Puspa Mega-10311600000075.pdf Download (4MB) | Preview |
Abstract
Gambar yang diperoleh menggunakan sensor Sonar di Kapal Tempur Bawah Air (KTBA) belum mampu mengidentifikasi obstacle, sehingga sulit bagi user untuk membedakan objek di bawah laut. Dalam penelitian ini, dikembangkan aplikasi untuk mengidentifikasi obstacle menggunakan metode Naive Bayes. Tahapan yang dilakukan dalam mengembangkan aplikasi untuk identifikasi obstacle terdiri dari 4 tahap. Pertama, object detection untuk memisahakan objek dari backgroud dengan menggunakan metode Selective Search. Kedua, feature extraction untuk mengekstraksi ciri dari objek menggukan metode Oriented FAST and Rotated BRIEF (ORB). Ketiga, klasifikasi objek menggunakan metode Naive Bayes. Keempat, desain Graphical User Interface (GUI) untuk menampilkan hasil klasifikasi menggunakan framework Qt. Hasil pengujian menggunakan metode histogram of occupancy memiliki akurasi yang lebih tinggi daripada metode histogram of distance normal dan histogram of distance softweight. Hal tersebut menunjukkan algoritma menggunakan histogram of occupancy mampu melakukan identifikasi obstacle dalam bentuk wajah manusia dengan akurasi 88,5%. Dengan hasil akurasi 88,5 % menunjukkan metode yang digunakan dapat digunakan untuk mengidentifikasi obstacle pada KTBA.
================================================================================================================================
Images obtained using the Sonar sensor on the Underwater Combat Ship (KTBA) have not been able to overcome obstacles, making it difficult for users to determine objects under the sea. In this study, an application was developed to overcome obstacles using the Naif Bayes method. The steps taken in developing applications to complete obstacles consist of 4 cups. First, object detection to separate objects from backgroud by using the Selective Search method. Second, feature extraction for extracting features from objects uses FAST-Oriented and Rotated BRIEF (ORB) methods. Third, object classification uses the Naive Bayes method. Fourth, the design of Graphical User Interface (GUI) to display the results of classification using the Qt framework. The test results using the residential histogram method have higher testing than the normal distance histogram and soft distance histogram methods. This shows that the algorithm using a residential histogram is able to perform a barrier in the form of a human face with an accuracy of 88.5%. The evaluation results showed that 88.5% of the methods could be used for constraints at the KTBA.
| Item Type: | Thesis (Diploma) |
|---|---|
| Additional Information: | RSEO 621.389 5 Meg a-1 2019 |
| Uncontrolled Keywords: | KTBA, Sonar, Naive Bayes, Klasifikasi Objek |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems V Naval Science > VK > VK388 Sonar |
| Divisions: | Faculty of Vocational > 36304-Automation Electronic Engineering |
| Depositing User: | Salma Puspa Mega |
| Date Deposited: | 08 Jun 2023 06:33 |
| Last Modified: | 08 Jun 2023 06:33 |
| URI: | http://repository.its.ac.id/id/eprint/70580 |
Actions (login required)
 |
View Item |