Pengenalan Kerusakan Jalan Menggunakan Histogram of Oriented Gradients dan Support Vector Machin (SVM) Bertingkat

Putri, Shinta Amalia Muchlis (2020) Pengenalan Kerusakan Jalan Menggunakan Histogram of Oriented Gradients dan Support Vector Machin (SVM) Bertingkat. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
06111640000019-Undergraduate_Thesis.pdf
Restricted to Repository staff only

Download (6MB) | Request a copy

Abstract

Kerusakan jalan menyebabkan aktivitas masyarakat menjadi terhambat hingga mengakibatkan kecelakaan lalu lintas. Identifikasi serta klasifikasi terhadap jenis kerusakan jalan diperlukan untuk meminimalkan dampak dari kerusakan jalan, sebelum dilakukan perbaikan. Dengan teknologi pengolahan citra digital, proses identifikasi dan klasifikasi jenis kerusakan jalan dapat dilakukan secara otomatis. Pada penelitian Tugas Akhir ini, pengenalan kerusakan jalan secara otomatis dengan menerapkan Histogram of Oriented Gradients pada proses ekstraksi fitur dan Support Vector Machine bertingkat untuk proses klasifikasi. Terdapat 3 kelas dalam proses klasifikasi, yaitu alligator, retak, dan lubang. Hasil klasifikasi digunakan untuk memberi label pada 23 citra kerusakan jalan agar setiap jenis kerusakannya mudah dikenali. Proses pengenalan kerusakan jalan menggunakan model klasifikasi SVM bertingkat dengan fungsi kernel polynomial dan strategi One Vs All, serta menggunakan ukuran sel 16×16 piksel saat proses ekstraksi fitur HOG dan menghasilkan nilai akurasi sebesar 85,71%, serta memerlukan waktu komputasi selama 4,375029 detik. =================================================================================================== Pavement distress causes inhibit humans activities until causes traffic accident. Identification and classification need to minimalize the impact of road damages before its reparation. Image processing can identify the variations of road damages instantly by using identification and classification. In this research of the final project, road distress identity can been done automatically with implementing Histogram of Oriented Gradients on the feature extraction and Multiclass Support Vector Machine for classification process. There are 3 classes in classification process. They are alligator, crack, and pothole. The result of classification is used for labelling on 23 images of road distress, so that every kinds of damages can identified. Identification process using Multiclass SVM model with polynomial kernel and One vs All strategy, also using cell size 16×16 pixels when HOG extraction feature process and gives accuracy 85,71% , also needs 4,375029 seconds for computing.

Item Type: Thesis (Undergraduate)
Additional Information: RSMa 518.1 Put p-1 • Putri, Shinta Amalia Muchlis
Uncontrolled Keywords: kerusakan jalan, pengenalan, Histogram of Oriented Gradients, Support Vector Machine Bertingkat ====================================================================================== pavement distress, recognition, Histogram of Oriented Gradients, Multiclass Support Vector Machine
Subjects: Q Science > QA Mathematics > QA76.6 Computer programming.
Divisions: Faculty of Mathematics and Science > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Shinta Amalia Muchlis Putri
Date Deposited: 23 Aug 2020 05:28
Last Modified: 14 Oct 2020 09:24
URI: https://repository.its.ac.id/id/eprint/79979

Actions (login required)

View Item View Item