Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/158948
Title: Prediksi Kandungan Klorofil Daun Tanaman Selada Romaine (Lactuca sativa L.) Menggunakan Artifical Neural Network
Other Titles: Estimation of Chlorophyll Content in Romaine Lettuce (Lactuca sativa L.) Using Artificial Neural Network
Authors: Suhardiyanto, Herry
Messaline, Theressa
Issue Date: 2024
Publisher: IPB University
Abstract: Selada romaine merupakan sayuran daun yang selalu mengalami peningkatan konsumsi. Salah satu parameter kualitas selada romaine adalah karakteristik warna daun, sehingga dengan mengetahui kandungan klorofil dapat meningkatkan nilai jual dan produktivitas tanaman. Pengamatan secara visual terhadap kandungan klorofil sulit untuk dilakukan. Tujuan dari penelitian ini adalah membangun model Artificial Neural Network (ANN) dengan algoritma backpropagation untuk mengestimasi kandungan klorofil secara efisien tanpa merusak tanaman. Penelitian dilakukan di rumah tanaman (greenhouse) dengan mengumpulkan citra daun dari 8 perlakuan terdiri dari kontrol, -N, -P, -K, 0%, 50%, 150%, dan 200% dari kontrol mulai 12 sampai 18 hari setelah tanam (HST). Hasil citra kemudian diolah untuk mendapatkan nilai indeks vegetasi tanaman yaitu Visible Atmospherically Resistant Indeks (VARI). Selain itu, jumlah kandungan klorofil aktual diukur dengan Soil Plant Analysis Development (SPAD-502). Sebanyak 720 dataset dikumpulkan dengan parameter input terdiri dari kondisi iklim mikro lingkungan pertumbuhan yaitu nilai suhu udara (°C), kelembaban relatif udara (RH) dan intensitas radiasi matahari (?W m ?^(-2)).), parameter larutan hara/nutrisi yaitu nilai ppm dan nilai pH, dan kadar larutan hara/nutrisi N, P, K (?mg L?^(-1)) serta VARI. Parameter output adalah nilai kandungan klorofil daun dalam satuan unit. Model ANN dibangun dengan penentuan parameter model dengan hasil learning rate initial 0,4, momentum 0,93, hidden layer size 2, dan maximum iteration 100. Struktur model ANN dengan kinerja terbaik didapatkan pada struktur 9-2-1 (9 input layer–2 hidden layer–1 output layer) menghasilkan R^2 0,909 dan Root Mean Square Error (RMSE) 0,060 yang menunjukkan kinerja model sudah berjalan sangat baik dengan tingkat akurasi yang tinggi.
Romaine lettuce is a leaf vegetable that is always increasing in consumption. One of the quality parameters of romaine lettuce is the color characteristics of the leaves, so knowing the chlorophyll content can increase the selling value and productivity of the plant. Visual observation of chlorophyll content is difficult to do. This research aimed to build an Artificial Neural Network (ANN) model with a backpropagation algorithm to estimate chlorophyll content efficiently without damaging the plants. The research was conducted in a greenhouse by collecting leaf images from 8 treatments of control, -N, -P, -K, 0%, 50%, 150%, and 200% of the control from 12 to 18 days after planting (DAP). The images were then processed to obtain the plant vegetation index value, namely the Visible Atmospherically Resistant Index (VARI). In addition, the actual chlorophyll content was measured with Soil Plant Analysis Development (SPAD-502). 720 datasets were collected with input parameters consisting of microclimate conditions of the growth environment, namely air temperature (°C), air relative humidity (RH) and solar radiation intensity (?W/m?^2), nutrient solution parameters, namely ppm and pH values, and nutrient solution levels of N, P, K (?mg L?^(-1)), and VARI. The output parameter was the value of leaf chlorophyll content in units. The ANN model is built by determining model parameters with the results of the initial learning rate 0,4, momentum 0,93, hidden layer size 2, and maximum iteration 100. The ANN model structure with the best performance is obtained in the 9-2-1 structure (9 input layers-2 hidden layers-1 output layer) resulting in R^2 0,909 and Root Mean Square Error (RMSE) 0,060 which shows the performance of the model has been running very well with a high level of accuracy.
URI: http://repository.ipb.ac.id/handle/123456789/158948
Appears in Collections:UT - Agricultural and Biosystem Engineering

Files in This Item:
File Description SizeFormat 
cover_F1401201050_99fc9d97699e41f99e680987e383ced7.pdfCover561.1 kBAdobe PDFView/Open
fulltext_F1401201050_8839bce54f8e49cc8f068df674c2647b.pdf
  Restricted Access
Fulltext2.53 MBAdobe PDFView/Open
lampiran_F1401201050_6aed2ce9131247209b989bf3662b4d98.pdf
  Restricted Access
Lampiran2.34 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.