Applications of deep learning in high-throughput seed phenotyping for quality assessment: A comprehensive review

Seed quality plays a critical role in achieving successful agricultural production. High-throughput phenotyping (HTP) techniques facilitate the rapid and non-destructive assessment of seed traits. The integration of HTP techniques with deep learning (DL) methods enables rapid and accurate evaluation of seed quality. The advent of DL technologies has created new opportunities for processing large and complex datasets more efficiently, leading to enhanced precision in seed quality assessment. This article provides a review of the principles underlying various HTP methods used for non-destructive seed analysis. It also summarizes recent advancements in DL-based approaches for seed quality inspection. The findings highlight that integrating DL with HTP offers a robust framework for extracting detailed phenotypic information, which can be effectively applied in agriculture for crop genetic improvement, quality parameter evaluation, and the identification of high-performing seeds.

کلیدواژه ها:

Deep learning ، HTP technique ، Seed phenotyping ، Seed quality