An intelligent crop cultivation system based on computer vision with a multiplex switch approach

Abstract. This paper proposes a multiplex switch approach based on computer vision for the selection of crop cultivation tasks. An image processing technique was used to identify weeds, crops, and the humidity of the soil in the cultivation field. Field images were collected to offer the identifier to estimate the desired pattern. Then, the classification results would be sent to a decision module that would then set the schedules for the corresponding operations, such as weeding, watering, and chemical spraying. The fuzzy-logic inference method was used to determine the soil humidity and produce a drive signal for water irrigation operations. Information from a soil humidity sensor served as the reference signal for the inference engine. The simulated results demonstrated the effectiveness and feasibility of the proposed system.