Artificial Intelligence and Autonomous Vehicles in Smart Agriculture: A Case Study of Pineapple Heart Detection

Smart agriculture is based on the concrete implementation of industrial technology arising from recent research and development in the sciences. Through an interdisciplinary combination of traditional agriculture and emerging technologies, the problems of population aging and climate change can be overcome. Furthermore, technologies related to intelligent agricultural mobile vehicles can be utilized to reduce labor costs, prevent environmental hazards, and enhance agricultural development. Pineapples are one of the most commercially important agricultural crops in Southeast Asia. During their cultivation, calcium carbide water is used for flower induction, and this must be directly placed in the heart of the pineapple plant in order to adjust the pineapple growing period and meet the increasing annual demand. This research paper proposes new pineapple heart detection and farm lane level positioning technologies by means of an intelligent agricultural vehicle to achieve a precise pineapple flower induction process. Pineapple heart detection technology integrates deep learning image recognition and circle Hough transform to detect the range of the black circular feature in the center of a pineapple plant that can be used as a basis for the location of spraying calcium carbide water by the intelligent agricultural vehicle. Farm lane level positioning technology integrates the farm lane image in front of the mobile vehicle and the coordinate information on both sides of the edges. Based on the position of the mobile vehicle in a farm lane, the GPS positioning information is corrected to the position of the mobile vehicle, and can be used to control the mobile vehicle. In sum, this research paper proposes an intelligent agricultural vehicle with pineapple heart detection and farm lane level positioning that can meet the needs of automated calcium carbide water spraying. The experimental results show that the recognition accuracy of pineapple heart detection reaches 94.5%, and the correction efficiency of farm lane level positioning is increased by 61.4%.