Universitas Dian Nuswantoro conducted an activity on adaptive models for determining cropland suitability and predicting total potato harvest using the Internet of Things. The activity involved Yusianto, Rindra; Shidiq, Guruh Fajar; Suprijono, Herwin; and Marimin as participants. It addressed potato production in Indonesia, where potatoes are described as the country’s fourth largest food agro-industrial commodity and a leading horticultural commodity traded internationally. The source states that, in 2021, the needs of the potato industry were met by only 85.93%.
The activity aimed to provide recommendations for adaptive models to increase productivity on the basis of agricultural cropland suitability and total-harvest prediction using IoT. The multi-thresholding method was modified by installing an SHT15 sensor to measure temperature. Rain gauge sensors were also installed, and the agricultural area was analyzed from a spatial perspective. A drone quadcopter was used for image processing and mapping to predict the total harvest. The research sample consisted of 12 randomly selected grids in an agricultural area in Wonosobo, Indonesia.
The findings classified 11.05% of the agricultural cropland as most suitable, 22.9% as suitable, and 62.01% as cropland with several inhibiting factors. The most convenient location was identified at coordinates 7° 15’ 12.1” S latitude and 109° 55’ 27.5” E longitude in Kejajar, Wonosobo District. The reported average harvest was 13.79 t ha-1. The model predicted an increase in yields, with a production prediction of 72,765 t year-1 and an accuracy rate of 89.35%. To balance supply and demand, the model recommended increasing production by 30% and fulfilling 31% from other regions. This activity relates to SDG 2 through research addressing potato productivity, harvest prediction, and supply and demand. It also provides context for SDG 10 through consideration of fulfillment from other regions. The activity is documented as a 2023 journal contribution.