Hadi, Raditya Wisnu (2022) Kecerdasan Artifisial Untuk Sistem Pemantauan Pertumbuhan Tanaman Hidroponik Berbasis Internet Of Things. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
07111840000091-Undergraduate_Thesis.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Kelompok tani “Keputih Bersemi” memiliki lahan hidroponik yang berada di lahan kantor kelurahan keputih. Penggunaan teknik kultur hidroponik tidak dapat mengatasi permasalahan-permasalahan pertanian seperti adanya hama ataupun penyakit yang sering mengganggu pertumbuhan tanaman. Oleh karena itu dilakukan proses pemantauan secara berkala, akan tetapi jumlah orang yang sedikit dan faktor kesalahan manusia lain sehingga proses pemantauan tidak berlangsung secara rutin. Dari permasalahan tersebut dibuat sistem pemantauan yang dapat mengidentifikasi pertumbuhan tanaman hidroponik dari waktu ke waktu. Untuk mengidentifikasi pertumbuhan digunakan suatu kecerdasan artifisial yang dapat mengetahui beberapa parameter pertumbuhan seperti tinggi, lebar, dan jumlah daun dari suatu jenis tanaman hidroponik yaitu kangkung. Metode yang digunakan untuk mengetahui parameter tersebut berbasis pengolahan citra dan proses deteksi objek dengan proses deep learning oleh Tiny-YOLOv3 pada Raspberry Pi V3. Deteksi objek dilakukan oleh model dengan 2 kelas yaitu kangkung dan sawi. Dari hasil evaluasi model, didapatkan rata-rata skor dari setiap hasil deteksi 97% dan rata-rata akurasi dari perhitungan jumlah daun oleh model sebanyak 80%. Data yang didapatkan dari hasil deteksi akan dikirimkan dan disimpan melalui platform Thingspeak yang berbasis IoT. Pertumbuhan tanaman direpresentasikan oleh beberapa parameter yaitu ketinggian dan jumlah daun. Data ketinggian tanaman yang tersimpan dibandingkan dengan data ketinggian rata-rata tanaman kangkung, sehingga dapat diketahui proses pertumbuhan tanaman tersebut normal atau tidak. Berdasarkan hasil evaluasi sistem, akurasi ketinggian objek pada objek besar lebih rendah dan akurasi jumlah daun pada objek kecil juga lebih rendah. Dari sistem yang sudah dibuat, didapatkan sistem pemantauan yang dapat mengidentifikasi pertumbuhan tanaman kangkung dari waktu ke waktu dengan metode deteksi objek dari model yang sudah dievaluasi.
=====================================================================================================================================
The “Keputih Bersemi” farmer group has a hydroponic land located on the Keputih sub-district office area. The use of hydroponic techniques cannot overcome agricultural problems such as the presence of pests or diseases that often interfere plant growth. Therefore, the monitoring process is carried out regularly, but the number of people is small and other human error factors so that the monitoring process does not take place regularly. From these problems, a monitoring system was created that can identify the growth of hydroponic plants from time to time. To identify growth, an artificial intelligence is used which can determine several growth parameters such as height, width, and number of leaves of a hydroponic plant, namely kale. The method used to determine these parameters is based on image processing and object detection with a deep learning process by Tiny-YOLOv3 on Raspberry Pi V3. Detection of objects is trained with 2 classes, namely kale and mustard greens. From the results of the model evaluation, the average score for each detection result is 97% and the average accuracy of the calculation of the number of leaves by the model is 80%. The data obtained from the detection results will be sent and stored through the IoT-based Thingspeak platform. Plant growth is represented by several parameters, namely height and number of leaves. The stored plant height data is compared with the average height data of the kale plant, so that it can be seen whether the plant growth process is normal or not. Based on the results of the system evaluation, the accuracy of the height of objects on large objects is lower and the accuracy of the number of leaves on small objects is also lower. From the system that has been created, a real-time monitoring system is obtained that can identify the growth of kale plants with object detection methods from models that have been evaluated.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Hidroponik, real-time monitoring, Tiny-YOLOv3 |
| Subjects: | Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis |
| Depositing User: | Raditya Wisnu Hadi |
| Date Deposited: | 03 Feb 2022 04:22 |
| Last Modified: | 14 Apr 2024 13:34 |
| URI: | http://repository.its.ac.id/id/eprint/92680 |
Actions (login required)
 |
View Item |