Auto Waypoint Deteksi Batas Sawah Menggunakan Visi Komputer Pada Drone Pertanian

Farras, Naufal Muhammad (2026) Auto Waypoint Deteksi Batas Sawah Menggunakan Visi Komputer Pada Drone Pertanian. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022211143-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Pemanfaatan drone pertanian menjadi salah satu solusi untuk meningkatkan efisiensi kegiatan pemantauan dan penyemprotan lahan sawah. Namun, salah satu permasalahan utama dalam pengoperasian drone pertanian adalah keterbatasan sistem navigasi dalam mengenali batas petak sawah secara otomatis, sehingga berpotensi menimbulkan tumpang tindih penyemprotan atau area yang terlewat. Oleh karena itu, penelitian ini bertujuan untuk mengembangkan sistem deteksi batas sawah berbasis visi komputer yang diintegrasikan pada drone pertanian sebagai acuan navigasi. Metode yang digunakan dalam penelitian ini adalah pengolahan citra digital menggunakan OpenCV dengan pendekatan rule-based tanpa penerapan machine learning. Citra video diperoleh secara real-time dari kamera yang terpasang pada drone dan diproses menggunakan ruang warna HSV untuk melakukan segmentasi warna hijau sebagai representasi area tanaman padi. Proses pengolahan citra meliputi konversi ruang warna, thresholding, filtering menggunakan median blur, serta ekstraksi kontur untuk menentukan batas sawah. Kontur terbesar digunakan sebagai acuan utama dalam pengambilan keputusan navigasi drone. Hasil deteksi kemudian diterjemahkan menjadi perintah kendali yang dikirim ke pemancar RC melalui Arduino Nano menggunakan sinyal PPM. Pengujian sistem dilakukan pada berbagai kondisi pencahayaan, yaitu pagi, siang, dan sore hari, serta pengujian penerbangan di area persawahan. Hasil pengujian menunjukkan bahwa sistem mampu mendeteksi batas sawah secara stabil dengan nilai frame per second (FPS) rata-rata sekitar 21-23 FPS pada resolusi 960×540 piksel. Sistem kendali yang dikembangkan juga mampu mengarahkan drone untuk bergerak maju serta melakukan manuver belok kanan dan kiri sesuai dengan hasil deteksi batas sawah. Dengan demikian, sistem yang diusulkan dapat menjadi solusi awal untuk mendukung navigasi drone pertanian berbasis visi komputer dalam rangka optimalisasi area penyemprotan.
==============================================================================================================================
The use of agricultural drones has become a promising solution to improve the efficiency of monitoring and spraying activities in rice fields. However, one of the main challenges in operating agricultural drones is the limited navigation capability to automatically recognize rice field boundaries, which may result in overlapping spraying or untreated areas. Therefore, this research aims to develop a computer vision–based rice field boundary detection system integrated into an agricultural drone as a navigation reference. The method used in this study is digital image processing using OpenCV with a rule-based approach without employing machine learning techniques. Real-time video images are acquired from a camera mounted on the drone and processed using the HSV color space to perform green color segmentation representing rice crop areas. The image processing stages include color space conversion, thresholding, filtering using median blur, and contour extraction to determine rice field boundaries. The largest contour is selected as the main reference for drone navigation decision making. The detection results are then translated into control commands and transmitted to the RC transmitter via an Arduino Nano using PPM signals. System testing was conducted under various lighting conditions, including morning, noon, and afternoon, as well as through flight tests over rice field areas. The results show that the system is able to detect rice field boundaries stably with an average frame per second (FPS) value of approximately 21–23 FPS at a resolution of 960×540 pixels. The developed control system is also capable of directing the drone to move forward and perform left and right turning maneuvers according to the detected rice field boundaries. Thus, the proposed system can serve as an initial solution to support computer vision–based navigation for agricultural drones in order to optimize spraying coverage.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	drone pertanian, deteksi batas sawah, visi komputer, OpenCV, pengolahan citra digital agricultural drone, rice field boundary detection, computer vision, OpenCV, digital image processing
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > TR Photography > TR810 Aerial photography
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Naufal Muhammad Farras
Date Deposited:	29 Jan 2026 09:55
Last Modified:	29 Jan 2026 09:55
URI:	http://repository.its.ac.id/id/eprint/130568

Actions (login required)

View Item