EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Non-destructive and accurate phenotypic detection method for okra seedlings under salt stress based on dual-view feature fusion and lightweight PointNet++

Siying Liu, Suhang Qian, Jingyi Tan, Jiamin Xu, Mei Lan, Ziyi Zhang, Haoqi Wang, Yuanyu Xia, Shihan Wang, Hongbin Wu, Xiuqing Fu

Frontiers in Plant Science · 11.06.2026 · 17

DOI: 10.3389/fpls.2026.1835129 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Plant Science, online erschienen am 11.06.2026. Verfasst von Siying Liu und 10 weiteren Autor:innen. Beteiligte Einrichtung: College of Engineering, Nanjing Agricultural University, Nanjing, China. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Soil salinization has become a critical factor limiting global agricultural production. Characterizing the growth and developmental responses of okra to salt stress and developing efficient and accurate salt-stress phenotyping techniques can provide an important methodological reference for okra cultivation in saline lands and future multi-cultivar salt-stress phenotyping studies. Traditional manual measurement of plant phenotypic parameters suffers from low efficiency and insufficient detection accuracy, making it difficult to achieve rapid and non-destructive analysis of plant phenotypic traits under salt stress. Therefore, this study proposes a computational phenotyping parameter extraction method based on the CSP-MSG Net model. Using dual-view feature fusion, we constructed a dedicated dataset. On the basis of PointNet++-MSG, the original MLP layers were replaced with C2F modules, and the SGE attention mechanism was integrated to enhance morphological feature extraction, thereby constructing a lightweight CSP-MSG Net architecture adapted to okra seedling point clouds for semantic segmentation of okra point clouds combined with DBSCAN clustering to complete instance segmentation, phenotypic parameters including plant height, stem diameter and canopy width were further calculated. This scheme enables high-throughput data acquisition, improves measurement accuracy, effectively reduces model parameters and computational overhead, and realizes lightweight operational performance. The results show that okra seedlings can still grow with increasing salt stress concentration, while the growth rates of the three measured traits are all inhibited, indicating that high-concentration salt stress impairs the growth activity of okra seedlings. To verify the calculation accuracy of the model, the phenotypic parameters predicted by the model were compared with manually measured values. The coefficients of determination for stem diameter, canopy width and plant height of okra seedlings reached 0.96, 0.99 and 0.99, respectively. These results strongly demonstrate the excellent reliability and effectiveness of the proposed method, providing methodological support for non-destructive and accurate phenotypic detection of okra seedlings under salt stress.

Schlagwörter

Stichwörter der Autor:innen

PointNet ++ modelokra seedlingsphenotypic detectionpoint cloud segmentationsalt stress

Autor:innen und Institutionen

  • College of Engineering, Nanjing Agricultural University, Nanjing, China.

Zitieren

APA

Liu, S., Qian, S., Tan, J., Xu, J., Lan, M., Zhang, Z., Wang, H., Xia, Y., Wang, S., Wu, H., & Fu, X. (2026). Non-destructive and accurate phenotypic detection method for okra seedlings under salt stress based on dual-view feature fusion and lightweight PointNet++. Frontiers in Plant Science, 17. https://doi.org/10.3389/fpls.2026.1835129

Vancouver

Liu S, Qian S, Tan J, Xu J, Lan M, Zhang Z, et al. Non-destructive and accurate phenotypic detection method for okra seedlings under salt stress based on dual-view feature fusion and lightweight PointNet++. Frontiers in Plant Science. 2026;17. doi:10.3389/fpls.2026.1835129

BibTeX
@article{Liu2026,
  author = {Liu, Siying and Qian, Suhang and Tan, Jingyi and Xu, Jiamin and Lan, Mei and Zhang, Ziyi and Wang, Haoqi and Xia, Yuanyu and Wang, Shihan and Wu, Hongbin and Fu, Xiuqing},
  title = {Non-destructive and accurate phenotypic detection method for okra seedlings under salt stress based on dual-view feature fusion and lightweight PointNet++},
  journal = {Frontiers in Plant Science},
  year = {2026},
  volume = {17},
  doi = {10.3389/fpls.2026.1835129}
}
Kostenlos starten →