Smart Boat Navigation System For River Traversing Using Computer Vision

Bachmid, Nabil (2025) Smart Boat Navigation System For River Traversing Using Computer Vision. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022222011-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (3MB) | Request a copy

Abstract

Lingkungan sungai menghadirkan tantangan signifikan dibandingkan dengan lingkungan laut karena fitur-fiturnya yang beragam dan kompleks. Pemrosesan data dari sungai, terutama data citra, secara inheren sulit karena kompleksitas ini. Penelitian ini memperkenalkan metode untuk mendeteksi area sungai pada citra menggunakan Fast Fourier Transform (FFT), deteksi kontur, dan Transformasi Garis Hough. Metode ini memanfaatkan karakteristik permukaan air yang memiliki warna seragam, yang menghasilkan komponen frekuensi rendah yang khas di domain frekuensi. Dengan memanfaatkan karakteristik ini, sistem yang diusulkan secara efektif mengidentifikasi wilayah sungai dengan akurasi tinggi.Metodologi mencakup beberapa langkah penting: pra-pemrosesan citra, penerapan FFT untuk analisis komponen frekuensi, thresholding untuk mengisolasi area frekuensi rendah, pengurangan noise melalui seleksi kontur, dan deteksi garis tepi sungai menggunakan Transformasi Garis Hough. Metode yang diusulkan telah dievaluasi pada dataset citra sungai yang beragam, dengan rata-rata nilai Intersection over Union (IoU) sebesar 90,5%, yang memvalidasi akurasinya dalam segmentasi area sungai. Selain itu, metode yang diusulkan berhasil diintegrasikan ke dalam sistem kontrol, memungkinkan unmanned surface vehicle (USV) untuk menavigasi sungai secara otonom. Sistem ini secara efektif menggunakan tepi sungai yang terdeteksi untuk menghitung heading target dan menyesuaikan lintasan USV, sehingga menunjukkan kepraktisan dari metode berbasis visi ini untuk navigasi sungai.
==================================================================================================================================
The river environment poses significant challenges compared to the marine environment due to its diverse and complex characteristics. Processing river data, particularly image data, is inherently challenging due to this complexity. This study introduces a method for detecting river regions in images using the Fast Fourier Transform (FFT), contour detection, and Hough Line Transform. The method leverages the characteristic of water surfaces exhibiting uniform color, resulting in distinctive low-frequency components in the frequency domain. Utilizing this feature, the proposed system effectively identifies river regions with high accuracy. The methodology involves several critical steps: image preprocessing, application of the Fast Fourier Transform (FFT) for frequency component analysis, thresholding to isolate low-frequencies area, noise reduction via contour selection, and riverbank line detection using the Hough Line Transform. The proposed method has been evaluated on a diverse dataset of river images, achieving an average Intersection over Union (IoU) score of 90.5%, which validates its accuracy in segmenting river areas. Moreover, the proposed method were successfully integrated into a control system, enabling an unmanned surface vehicle (USV) to autonomously navigate a river. The system effectively utilized the detected riverbanks to calculate the target heading and adjust the USV's trajectory, demonstrating the practical applicability of the proposed vision-based method for river navigation.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Frekuensi, Navigasi, Sungai, USV, Visi Komputer, Computer Vision, Frequency, Navigation, River
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Nabil Bachmid
Date Deposited:	31 Jan 2025 01:50
Last Modified:	31 Jan 2025 01:50
URI:	http://repository.its.ac.id/id/eprint/117201

Actions (login required)

View Item