Rancang Bangun Monitoring Lebar Daun Berbasis Segmentasi Citra

Rahmadani, Nadia Sari Dwi (2026) Rancang Bangun Monitoring Lebar Daun Berbasis Segmentasi Citra. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221046-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Pertanian perkotaan (urban farming) membutuhkan metode pemantauan pertumbuhan tanaman yang efektif dan non destruktif. Salah satu parameter yang dapat digunakan untuk mengevaluasi kondisi tanaman adalah lebar daun karena berkaitan dengan kemampuan fotosintesis dan kondisi fisiologis tanaman. Penelitian ini bertujuan untuk merancang dan mengevaluasi sistem monitoring lebar daun berbasis segmentasi citra menggunakan Fast Segment Anything Model (FastSAM) yang diimplementasikan pada Raspberry Pi 4. Sistem menggunakan kamera Logitech C615 untuk memperoleh citra tanaman secara top view, kemudian citra diproses menggunakan FastSAM untuk melakukan segmentasi daun. Mask daun target dipilih berdasarkan ukuran objek dan kedekatannya terhadap pusat citra. Lebar daun diukur menggunakan Principal Component Analysis (PCA) dan dikonversi ke satuan sentimeter menggunakan persamaan regresi linear yang dikalibrasi secara terpisah untuk tanaman bayam dan pakcoy. Monitoring dilakukan dengan membandingkan hasil pengukuran sistem terhadap pengukuran menggunakan penggaris sebagai alat ukur acuan sebanyak 50 kali pengambilan data pada masing-masing tanaman. Hasil monitoring menunjukkan bahwa pada tanaman bayam dengan jarak kamera 21 cm diperoleh range pengukuran sebesar 2,169–3,961 cm, span sebesar 1,792 cm, rata-rata error sebesar 4,05%, dan akurasi sebesar 95,95%. Sementara itu, pada tanaman pakcoy dengan jarak kamera 25 cm diperoleh range pengukuran sebesar 4,731–6,094 cm, span sebesar 1,363 cm, rata-rata error sebesar 1,51%, dan akurasi sebesar 98,49%. Secara keseluruhan, sistem memperoleh rata-rata error sebesar 2,78% dengan rata-rata akurasi sebesar 97,22%.
===============================================================================================================================================
Urban farming requires effective, non-destructive methods for monitoring plant growth. Leaf width is a key parameter for evaluating plant condition, as it correlates with photosynthetic capacity and physiological status. This study aimed to design and evaluate an image segmentation-based leaf width monitoring system using the Fast Segment Anything Model (FastSAM) implemented on a Raspberry Pi 4. The system captures top-view plant images using a Logitech C615 camera, which are then processed via FastSAM for leaf segmentation. Target leaf masks are selected based on object size and proximity to the image center. Leaf width is measured using Principal Component Analysis (PCA) and converted to centimeters via linear regression equations calibrated separately for spinach and pakcoy. System performance was evaluated by comparing measurements against a ruler (reference standard) across 50 data acquisition trials for each plant type. Results for spinach (at a camera distance of 21 cm) showed a measurement range of 2.169–3.961 cm, a span of 1.792 cm, an average error of 4.05%, and an accuracy of 95.95%. For pakcoy (at a camera distance of 25 cm), the results showed a measurement range of 4.731–6.094 cm, a span of 1.363 cm, an average error of 1.51%, and an accuracy of 98.49%. Overall, the system achieved an average error of 2.78% and an average accuracy of 97.22%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Monitoring, Segmentasi Citra, Fast Segment Anything Model, Raspberry Pi 4, Principal Component Analysis, Monitoring, Image Segmentation, Fast Segment Anything Model, Raspberry Pi 4, Principal Component Analysis
Subjects:	S Agriculture > SB Plant culture T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7888.3 Digital computers
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Nadia Sari Dwi Rahmadani
Date Deposited:	05 Aug 2026 10:15
Last Modified:	05 Aug 2026 10:15
URI:	http://repository.its.ac.id/id/eprint/144099

Actions (login required)

View Item