High-throughput phenotyping in horticultural crops: Innovations, challenges, and the path ahead

Abstract

Image-based phenotyping has become vital in the breeding, cultivation, and quality assessment of economically important crops. This article examines its advancements and applications in horticultural crops, highlighting imaging techniques such as RGB, thermal, hyperspectral, fluorescence, and tomographic imaging. These methods enable high-throughput phenotyping across various traits, including morphology, physiology, biochemistry, diseases, pests, and abiotic stresses. In addition to accelerating breeding cycles through rapid trait measurement, image-based phenotyping supports real-time monitoring and quick decision-making for activities like pesticide application, fertilization, and harvesting, ultimately improving yield and produce quality. It also plays a crucial role in postharvest phenotyping, ensuring quality during storage and handling. With its impact on both yield and product quality, image-based phenotyping is integral to the entire horticultural chain. The integration of machine learning and deep learning technologies is key to efficiently extracting valuable insights from the extensive data generated.