Arnhold Jessica, Ispizua Yamati Facundo R, Kage Henning, Mahlein Anne-Katrin, Koch Heinz-Josef, Grunwald Dennis
Institute of Sugar Beet Research, Holtenser Landstraße 77, 37079, Göttingen, Germany.
Institute of Crop Science and Plant Breeding, Kiel University, 24118, Kiel, Germany.
Plant Methods. 2024 Dec 17;20(1):183. doi: 10.1186/s13007-024-01313-0.
Root growth is most commonly determined with the destructive soil core method, which is very labor-intensive and destroys the plants at the sampling spots. The alternative minirhizotron technique allows for root growth observation throughout the growing season at the same spot but necessitates a high-throughput image analysis for being labor- and cost-efficient. In this study, wheat root development in agronomically varied situations was monitored with minirhizotrons over the growing period in two years, paralleled by destructive samplings at two dates. The aims of this study were to (i) adapt an existing CNN-based segmentation method for wheat minirhizotron images, (ii) verify the results of minirhizotron measurements with root growth data obtained by the destructive soil core method, and (iii) investigate the effect of the presence of the minirhizotron tubes on root growth.
The previously existing CNN could successfully be adapted for wheat root images. The minirhizotron technique seems to be more suitable for root growth observation in the subsoil, where a good agreement with destructively gathered data was found, while root length results in the topsoil were dissatisfactory in comparison to the soil core method in both years. The tube presence was found to affect root growth only if not installed with a good soil-tube contact which can be achieved by slurrying, i.e. filling gaps with a soil/water suspension.
Overall, the minirhizotron technique in combination with high-throughput image analysis seems to be an alternative and valuable technique for suitable research questions in root research targeting the subsoil.
根系生长最常用的测定方法是破坏性土芯法，该方法劳动强度大，且会破坏采样点的植物。另一种微根窗技术可以在整个生长季节在同一地点观察根系生长，但为了提高效率和降低成本，需要进行高通量图像分析。在本研究中，使用微根窗在两年的生长期间监测了不同农艺条件下小麦根系的发育，并在两个时间点进行了破坏性采样。本研究的目的是：（i）将现有的基于卷积神经网络（CNN）的分割方法应用于小麦微根窗图像；（ii）用破坏性土芯法获得的根系生长数据验证微根窗测量结果；（iii）研究微根窗管的存在对根系生长的影响。
现有的CNN能够成功应用于小麦根系图像。微根窗技术似乎更适合于观察土壤下层的根系生长，在那里与破坏性采集的数据有很好的一致性，而与土芯法相比，两年中土壤上层的根长结果都不尽人意。只有在安装时土壤与管之间没有良好接触的情况下，管的存在才会影响根系生长，而通过泥浆灌注（即用土壤/水悬浮液填充间隙）可以实现良好的土壤-管接触。
总体而言，微根窗技术与高通量图像分析相结合，似乎是针对土壤下层根系研究中合适研究问题的一种有价值的替代技术。
