Subtropical fruit cultivation system based on IoT and evolutionary algorithm optimization.

Abstract

Crop cultivation in Okinawa relies on farmers' unique experience and intuition because the climate is different from that of the mainland. In recent years of farming, there is a limit to carrying out crop cultivation without the experience and knowledge of farmers. As a means to solve this problem, IoT × AI is an effective means, and it becomes possible to find a growth environment suitable for Okinawa by analyzing the information accumulated in consideration of regional characteristics. In this research, research and development of a growth support system is conducted focusing on mango, which is a paid cash crop in Okinawa Prefecture. We will apply evolutionary computation algorithm in the data analysis part. Mango is difficult to predict stable yield because it is repeated good harvest and poor crop, stable environmental control is difficult. We analyze the statistical effects of environmental parameters from meteorological data on mango yield to clarify the characteristics. Moreover, although the grade of mango is determined by the color, since it is performed by the farmer's visual observation, discrimination is uneven, and the present condition is that the standard is not unified and the quality is uneven. The purpose is to identify the indicator of grading of mango from image data.