Scholarlink

<p>To effectively catalogue and utilise the genetic wealth of tea (<i>Camellia sinensis</i> (L.) O. Kuntze), integrating extensive phenotypic data with existing germplasm resources is essential. While genomic tools have advanced, the application of high-throughput digital phenotyping remains limited in practice, creating a gap between genetic potential and observable traits. To address this, the study quantitatively assessed the leaf morphological diversity of a core collection of 129 accessions, representing a hybrid continuum of fundamental lineages and introgressed morphotypes, alongside six related <i>Camellia</i> species. We applied computer vision techniques to 2580 digital images of maintenance leaves, systematically sampled from replicated plants of each accession to extract a comprehensive dataset comprising linear geometric measurements, CIELAB colour coordinates, surface texture descriptors, and Elliptic Fourier Descriptors. The results revealed an extensive phenotypic continuum that effectively resolved intermediate morphotypes which are often difficult to classify manually. Static allometric analysis quantitatively confirmed distinct scaling strategies: high-yielding Assam's types showed significant negative scaling, constrained towards broader shapes as they expanded, whereas China's types maintained a consistent narrow form. Furthermore, the analysis showed that related species, such as <i>Camellia rosiflora</i> and <i>Camellia irrawadiensis</i>, possess distinct serration patterns and greater textural complexity compared to cultivars. These specific morphological features are likely associated with structural adaptations to environmental stress. The study demonstrates that these standardized digital tools offer a rigorous method to generate descriptors for genebank management, providing breeders with precise criteria to select trait donors for developing climate-resilient tea cultivars.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Computer vision-based leaf morphometry of Indian tea hybrids and related Camellia species

  • 期刊论文

  • 发表日期：

    2026年4月2日

查看全文
  • Shuvam Datta,
  • Pritom Chowdhury,
  • Sangeeta Borchetia,
  • Hemanta Saikia

摘要

To effectively catalogue and utilise the genetic wealth of tea (Camellia sinensis (L.) O. Kuntze), integrating extensive phenotypic data with existing germplasm resources is essential. While genomic tools have advanced, the application of high-throughput digital phenotyping remains limited in practice, creating a gap between genetic potential and observable traits. To address this, the study quantitatively assessed the leaf morphological diversity of a core collection of 129 accessions, representing a hybrid continuum of fundamental lineages and introgressed morphotypes, alongside six related Camellia species. We applied computer vision techniques to 2580 digital images of maintenance leaves, systematically sampled from replicated plants of each accession to extract a comprehensive dataset comprising linear geometric measurements, CIELAB colour coordinates, surface texture descriptors, and Elliptic Fourier Descriptors. The results revealed an extensive phenotypic continuum that effectively resolved intermediate morphotypes which are often difficult to classify manually. Static allometric analysis quantitatively confirmed distinct scaling strategies: high-yielding Assam's types showed significant negative scaling, constrained towards broader shapes as they expanded, whereas China's types maintained a consistent narrow form. Furthermore, the analysis showed that related species, such as Camellia rosiflora and Camellia irrawadiensis, possess distinct serration patterns and greater textural complexity compared to cultivars. These specific morphological features are likely associated with structural adaptations to environmental stress. The study demonstrates that these standardized digital tools offer a rigorous method to generate descriptors for genebank management, providing breeders with precise criteria to select trait donors for developing climate-resilient tea cultivars.