Crop failures among chili farmers are often caused by viruses that infect plants during their growth. The yellow virus, or Pepper Yellow Leaf Curl Indonesian Virus (PYLCIV), is one of the viruses that infects chili plants and causes distinctive symptoms, including yellowing leaves. Treatment is often carried out too late because symptoms may only become apparent when chili plants are approaching harvest.
In response to this issue, five UGM students participating in Team SPECTRA under the Student Creativity Program for Exact Sciences Research (PKM-RE) have developed an early detection innovation for chili plants as an initial step to determine disease status using a multispectral sensor and machine learning analysis. The research was initiated by five students from different disciplines, namely Hasan Rabbani from the Undergraduate Physics Program as the team leader, along with four members: Nur Johan Pratama and Priamitra Aditiya Satriamukti from the Undergraduate Plant Protection Program, and Saktian Hutama and Zaskia Fakhira Priatna from the Undergraduate Biology Program. The research is conducted under the supervision of Prof. Dr. Eng. Kuwat Triyana, M.Si., a lecturer and Dean of the Faculty of Mathematics and Natural Sciences (FMIPA) UGM.
The initial stage of the research involved growing 30 chili plants as samples in Cangkringan, Sleman, Special Region of Yogyakarta. The location was selected because it is situated on the slopes of Mount Merapi and provides suitable conditions for chili plant growth. “Chili plants should be grown in an environment with normal daily temperatures, neither too high nor too low, to produce healthy leaves that can be analyzed as samples. This is intended to ensure accurate research results,” said Priamitra Aditiya Satriamukti, a member of the team, on Thursday (11 June).
Half of the plants were grown under healthy conditions in cages with routine care, including watering and fertilization. The remaining plants were used as samples for three main stages of testing. The first stage involved maintaining whiteflies inside cages. The second stage involved transmitting the yellow virus carried by whiteflies to chili plants. The final stage involved analysis using a multispectral sensor and machine learning.
The preliminary results showed that whiteflies carried the yellow virus and were able to transmit it to healthy chili plants. Leaves from both virus-infected and non-infected chili plants were subsequently tested using polymerase chain reaction (PCR) to validate their disease status. Once the disease status was confirmed, the next step was to collect data from the chili leaves using a multispectral sensor, followed by machine learning analysis. The sensor is relatively simple to operate: a leaf is clipped using the multispectral device for approximately five seconds until a reading graph appears. The characteristics of light reflected by the leaves can then be used to distinguish between healthy chili plants, which are not infected with the yellow virus, and diseased plants infected with the virus. The developed model achieved an accuracy of nearly 80% in distinguishing between the two conditions.
Hasan Rabbani, the team leader, explained that the research is important because disease detection in chili plants, particularly yellow virus infection, is still commonly conducted through direct visual observation of the leaves. Although PCR analysis can also be used, it generally involves relatively high costs and requires specialized expertise. At the same time, farmers need detection methods that are efficient, flexible, and affordable. “The purpose of this four-month study is to determine the extent to which multispectral sensors can detect chili plants infected with the yellow virus. PCR testing is conducted to validate that the multispectral sensor used is functioning properly,” said Hasan Rabbani on Wednesday (22 July).
The research findings indicate that the combination of multispectral sensing and machine learning has the potential to become a new approach for detecting viral infections in chili plants more rapidly and without damaging the plants. Team SPECTRA hopes that this innovation can be further developed and eventually benefit the wider community, particularly chili farmers.
The research conducted by Team SPECTRA through the development of early virus-detection technology for chili plants contributes to the achievement of the Sustainable Development Goals (SDGs), particularly SDG 2 (Zero Hunger) by supporting agricultural productivity and food security, SDG 9 (Industry, Innovation and Infrastructure) through the development of innovation based on multispectral sensing and machine learning, and SDG 12 (Responsible Consumption and Production) through the development of a rapid and non-destructive disease detection method. This innovation is expected to contribute to more efficient and accessible technological solutions that can help farmers detect plant diseases at an earlier stage.
Writer: Saktian Hutama
Photo: PKM Team
