RANCANG BANGUN ALAT FLUORESCENCE IMAGING BERBASIS IoT TERKOMBINASI DEEP LEARNING UNTUK MENDETEKSI KLOROFIL PADA SAYURAN HIJAU

Deyan Triyanda, NIM.: 22106020034 (2026) RANCANG BANGUN ALAT FLUORESCENCE IMAGING BERBASIS IoT TERKOMBINASI DEEP LEARNING UNTUK MENDETEKSI KLOROFIL PADA SAYURAN HIJAU. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (RANCANG BANGUN ALAT FLUORESCENCE IMAGING BERBASIS IoT TERKOMBINASI DEEP LEARNING UNTUK MENDETEKSI KLOROFIL PADA SAYURAN HIJAU) 22106020034_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (6MB) | Preview
	Text (RANCANG BANGUN ALAT FLUORESCENCE IMAGING BERBASIS IoT TERKOMBINASI DEEP LEARNING UNTUK MENDETEKSI KLOROFIL PADA SAYURAN HIJAU) 22106020034_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf - Published Version Restricted to Registered users only Download (14MB) | Request a copy

Abstract

An Internet of Things (IoT)-based fluorescence imaging system integrated with deep learning has been successfully developed for chlorophyll detection in green vegetables. This study represents an advancement of the previous generation of fluorescence imaging systems, which still had limitations in system flexibility, image quality, fluorescence area segmentation, and software development. To address these limitations, an IoT-based fluorescence imaging system was developed using the Python programming language and implementing the Real Enhanced Super-Resolution Generative Adversarial Network (Real-ESRGAN) and U²-Net deep learning models. The system was designed through circuit schematic development using Proteus, casing design using SolidWorks, and system flowchart development using Draw.io, and was realized using a 365 nm UV LED as the fluorescence excitation source and an ESP32-CAM as the image acquisition device. Image processing employed the Real-ESRGAN model for noise reduction and image resolution enhancement, while automatic fluorescence area segmentation was performed using the U²-Net model. The samples consisted of spinach, lettuce, mustard greens, and pakcoy. Testing was conducted by two operators using the developed fluorescence imaging system to evaluate chlorophyll fluorescence area proportion, repeatability precision, intermediate precision, and data communication performance. The results showed that the chlorophyll fluorescence area proportion in spinach decreased from 65.97%–66.80% on the first day to 37.04%–38.74% on the third day, in lettuce from 73.84%–74.90% to 61.72%–62.65%, in mustard greens from 72.82%–73.68% to 34.87%–35.78%, and in pakcoy from 82.80%–84.83% to 27.89%–28.78%. The repeatability precision values ranged from 98.79% to 99.81%, while the intermediate precision values ranged from 99.20% to 99.87%, thereby meeting the acceptance criteria specified in SNI ISO/IEC 17025:2017. Furthermore, data communication testing showed throughput values ranging from 1150 kBps to 1777 kBps, which fall within the good to very good categories according to the Telecommunications and Internet Protocol Harmonization Over Networks (TIPHON) Quality of Service (QoS) standard. These results indicate that the developed system is capable of providing consistent measurements, supporting automatic fluorescence image analysis, and maintaining stable data transmission performance, thereby demonstrating its potential as a rapid and non-destructive alternative method for chlorophyll detection.

Share this knowledge with your friends :

Actions (login required)

View Item