High-throughput phenotyping technologies have become an increasingly important topic of crop science in recent years. Various sensors and data acquisition approaches have been applied to acquire the phenotyping traits. It is quite confusing for crop phenotyping researchers to determine an appropriate way for their application. In this study, three representative three-dimensional (3D) data acquisition approaches, including 3D laser scanning, multi-view stereo (MVS) reconstruction, and 3D digitizing, were evaluated for maize plant phenotyping in multi growth stages. Phenotyping traits accuracy, post-processing difficulty, device cost, data acquisition efficiency, and automation were considered during the evaluation process. 3D scanning provided satisfactory point clouds for medium and high maize plants with acceptable efficiency, while the results were not satisfactory for small maize plants. The equipment used in 3D scanning is expensive, but is highly automatic. MVS reconstruction provided satisfactory point clouds for small and medium plants, and point deviations were observed in upper parts of higher plants. MVS data acquisition, using low-cost cameras, exhibited the highest efficiency among the three evaluated approaches. The one-by-one pipeline data acquisition pattern allows the use of MVS high-throughput in further phenotyping platforms. Undoubtedly, enhancement of point cloud processing technologies is required to improve the extracted phenotyping traits accuracy for both 3D scanning and MVS reconstruction. Finally, 3D digitizing was time-consuming and labor intensive. However, it does not depend on any post-processing algorithms to extract phenotyping parameters and reliable phenotyping traits could be derived. The promising accuracy of 3D digitizing is a better verification choice for other 3D phenotyping approaches. Our study provides clear reference about phenotyping data acquisition of maize plants, especially for the affordable and portable field phenotyping platforms to be developed.

Stalk lodging is an impediment to improving profitability and production efficiency in maize. Lodging resistance, a comprehensive indicator to appraise genotypes, requires both characterization of mechanical properties in laboratory and investigation of lodging percentage in field. However, in situ characterization of maize lodging resistance still remains poor. The aim of this study was to develop an indicator, named cumulative lodging index (CLI), based on lodging percentages at different wind speeds for evaluating lodging resistance for different maize cultivars, and to evaluate the accuracy and reliability of this method. Different cultivars showed different patterns of lodging percentage along with wind speeds. The failure wind speed (FWS) for maize ranged between 16 and 30 m s−1 across cultivars. The CLI differed between maize cultivars and showed favorable reliability (i.e. nRMSE of 5.38%). Mechanical properties of the third internode did not vary significantly between cultivars. Significant differences in the reduction index (RI) of wind speed sheltered by maize canopy were found between cultivars. Our findings implied that mobile wind machine is powerful in reproducing wind disaster that induce crop lodging. The newly-built CLI was demonstrated to be a more robust indicator than mechanical properties, FWS, and RI when evaluating lodging resistance in terms of both reliability and resolution. This study offers a new perspective for evaluating in situ lodging resistance of crops, and provides technical support for accurate identification of lodging-resistant phenotypic traits.

The Anthesis-Silking Interval (ASI) is a crucial indicator of the synchrony of reproductive development in maize, reflecting its sensitivity to adverse environmental conditions such as heat stress and drought. This paper presents an automated method for detecting the maize ASI index using a field high-throughput phenotyping platform. Initially, high temporal-resolution visible-light image sequences of maize plants from the tasseling to silking stage are collected using a field rail-based phenotyping platform. Then, the training results of different sizes of YOLOv8 models on this dataset are compared to select the most suitable base model for the task of detecting maize tassels and ear silks. The chosen model is enhanced by incorporating the SENetv2 and the dual-layer routing attention mechanism BiFormer, named SEBi-YOLOv8. The SEBi-YOLOv8 model, with these combined modules, shows improvements of 2.3% and 8.2% in mAP over the original model, reaching 0.989 and 0.886, respectively. Finally, SEBi-YOLOv8 is used for the dynamic detection of maize tassels and ear silks in maize populations. The experimental results demonstrate the method’s high detection accuracy, with a correlation coefficient (R2) of 0.987 and an RMSE of 0.316. Based on these detection results, the ASI indices of different inbred lines are calculated and compared.

Method for the real-time detection of tomato ripeness using a phenotype robot and RP-YolactEdge