Scholarlink

<p>Rapid advances in plant phenomics technologies have led to their application in recent breeding and genetic studies. Although various phenomics platforms have been developed, those that can obtain independent and precise 3D images for each plant without moving them have rarely been designed. Therefore, we developed a mobile platform to fill this technological gap. It comprises three high-resolution action cameras that rotate around a plant to take a video. A foldable background panel was installed to prevent other plants besides the subject from being captured in the images, eliminating the need for any post‑processing to remove or distinguish them. The platform can pass by each plant after the cameras have rotated and thus can record continuous video of multiple plants simultaneously. The video was segmented for each plant based on the QR code and then utilized to extract multiple 2D images at regular angular intervals. The images were merged using the Structure-from-Motion (SfM) workflow to reconstruct a precise and reproducible 3D image for each plant. An analysis of the 3D images of pepper plants at different growth stages indicated only a small discrepancy between the counts of floral buds and fruits in the 3D images and those obtained by direct observation. We expect this system to enable the 3D analysis of stationary plants in diverse environments, providing a strong foundation for breeding and genetic research. In particular, it is expected to have high practicality as it is inexpensive compared to existing systems.</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

A mobile and rotary phenotyping platform developed to obtain high-resolution 3D images of peppers

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • Sung Jae Lee,
  • Yeong Deuk Jo

摘要

Rapid advances in plant phenomics technologies have led to their application in recent breeding and genetic studies. Although various phenomics platforms have been developed, those that can obtain independent and precise 3D images for each plant without moving them have rarely been designed. Therefore, we developed a mobile platform to fill this technological gap. It comprises three high-resolution action cameras that rotate around a plant to take a video. A foldable background panel was installed to prevent other plants besides the subject from being captured in the images, eliminating the need for any post‑processing to remove or distinguish them. The platform can pass by each plant after the cameras have rotated and thus can record continuous video of multiple plants simultaneously. The video was segmented for each plant based on the QR code and then utilized to extract multiple 2D images at regular angular intervals. The images were merged using the Structure-from-Motion (SfM) workflow to reconstruct a precise and reproducible 3D image for each plant. An analysis of the 3D images of pepper plants at different growth stages indicated only a small discrepancy between the counts of floral buds and fruits in the 3D images and those obtained by direct observation. We expect this system to enable the 3D analysis of stationary plants in diverse environments, providing a strong foundation for breeding and genetic research. In particular, it is expected to have high practicality as it is inexpensive compared to existing systems.