Advancements of Artificial Intelligence in Horticulture: A Comprehensive Review

Abstract

Artificial Intelligence (AI) is rapidly transforming horticulture by introducing innovative solutions for phenotyping, cultivation practices, crop protection, postharvest handling, and breeding. Although significant advancements have been achieved in recent years, challenges such as scalability, affordability, limited farmer training, and the lack of transparency in AI models continue to restrict widespread adoption. Between 2019 and 2023, notable progress has been made, with systematic reviews of literature from databases such as Scopus, Web of Science, ScienceDirect, and IEEE Xplore highlighting promising applications. Out of 512 initially retrieved studies, 124 were shortlisted, and 72 were included after applying strict selection criteria. The findings suggest that AI holds immense potential in vision-based phenotyping, remote sensing, agricultural robotics, genomics integration, and the development of digital twin models. However, critical gaps remain in optimizing quality, diversifying datasets, improving farmer-oriented interpretability, and developing supportive policy frameworks. Emerging technologies such as Vision Transformers, neuro-symbolic AI, digital twins, and federated learning are shaping the future of AI-driven horticulture [2]. For equitable adoption, research must prioritize dataset standardization, participatory validation with farmers, cost-effective deployment strategies, and ethical governance. A practical roadmap for the future includes establishing dataset validation protocols, integrating hybrid AI with affordable robotics, and enabling ecosystem-wide adoption supported by regulatory and ethical frameworks.