Notice: Function _load_textdomain_just_in_time was called incorrectly. Translation loading for the pojo-accessibility domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /home/doctoral/public_html/wp-includes/functions.php on line 6121
ICDSUPL4-T002 – University of Life Sciences in Lublin

ICDSUPL4-T002

Volume: 4, 2025
4th International PhD Student’s Conference at the University of Life Sciences in Lublin, Poland:
ENVIRONMENT – PLANT – ANIMAL – PRODUCT

Abstract number: T002

DOI: https://doi.org/10.24326/ICDSUPL4.T002

Published online: 9 April 2025

ICDSUPL, 4, T002 (2025)


Machine learning solutions for seed quality prediction

Paweł Jaskuła1*

1 Department of Computer Science, Faculty of Electrical Engineering and Computer Science, Lublin University of Technology, Nadbystrzycka 38A, 20-618 Lublin, Poland

* Corresponding author: jaskulap@gmail.com

Abstract

Seed quality is one of the key determinants of crop yield. Even under optimal growing conditions, the lack of good quality seeds can lead to disappointing harvests. Despite this, traditional quality assessment methods, mostly based on manual classification by experts, are still the main way of seed selection. This process is inefficient and prone to errors due to the subjective approach of the assessors. According to an FAO report, by 2050, global food production will have to increase by 70% compared to 2009 levels to meet the consumption needs of the growing and economically advancing population. Therefore, there is an increasing need to create new, scalable and efficient solutions that will ensure optimal seed selection. This review analyzes current machine learning methods that contribute to solving the previously described problems, offering precise and fast seed assessment. The aim of this study is to review selected research from recent years on seed assessment using machine learning techniques. The analyzed studies, sourced from databases such as ScienceDirect, Springer, MDPI, and Nature show that ML methods are already capable of detecting a wide range of defects and imperfections – from mechanical or moisture damage to pathogen infections, and their accuracy exceeds human experts in this field. Importantly, many solutions do not even require advanced equipment; some methods use ordinary cameras or scanners, which makes them possible to use even by small farms. Thanks to the automation of the entire process, the time needed for analysis is significantly reduced, which means that a larger number of seeds can be tested, and farmers thus obtain better planting material, which translates into an increase in both yields and profitability. Despite continuous development and promising prospects, the use of ML is still associated with certain limitations. The diversity of plant species and working conditions in the field makes the creation of universal classification methods extremely difficult. It is certain, however, that agriculture and advanced technologies, such as machine learning, are a combination that will permanently inscribe themselves in the future of agriculture, and the challenges currently faced by existing methods will likely be solved in the near future.

Keywords: machine learning, seed classification, non-destructive testing, image analysis


How to cite

P. Jaskuła, 2025. Machine learning solutions for seed quality prediction. In: 4th International PhD Student’s Conference at the University of Life Sciences in Lublin, Poland: Environment – Plant – Animal – Product. https://doi.org/10.24326/ICDSUPL4.T002

Skip to content