Micro-master: A nondestructive microscopic panoramic imaging system for plants

Jianglin Wang , Jiawei Shi , Wanneng Yang
National Key Laboratory of Crop Genetic Improvement and National Center of Plant Gene Research, Huazhong Agricultural University, Wuhan, 430070, Chin

Microscopic phenotypes play a crucial role in bridging the gap between genotypes and macroscopic phenotypes in plants. To achieve real-time observation of plant tissue microstructures, we have developed a non-destructive microscopic imaging system, named Micro-Master. This system consists of a high-precision motion module and a high-resolution microscopic imaging system, enabling high-precision four-dimensional spatial movement. The Micro-Master system facilitates dynamic scanning of microscopic phenotypes in plant leaves and root systems, and performs real-time panoramic image stitching. Currently, the system has a throughput of up to 100 plants, with a complete panoramic microscopic image of a rice root system captured in just 2 minutes, and the scanning of 100 plants can be completed within 3-4 hours. For trait extraction, we have developed a deep learning-based software, Micro Traits Master (MTM), which leverages the Segformer network and image processing techniques to extract root hair and stomatal traits efficiently. The MTM software can rapidly extract 20 phenotypic traits including root hair distribution, root hair development rate, root hair density, stomatal size, and stomatal aperture. This study presents a solution for high-throughput, non-destructive microscopic phenotyping in plants