The mechanical hill wet-seeded rice machine is benefits to establish uniform seedling, and ditches were established by using this machine. However, little knowledge is known on the effect of the establishment of ditches on growth, lodging and yield, and their relationship with root traits. In this study, two field experiments were conducted during 2012 and 2013 with using two super rice varieties (i.e. hybrid rice ‘Peizataifeng’ and inbred rice ‘Yuxiangyouzhan’) grown under three ditches establishment treatments (i.e. T1: both water ditches and seed ditches were established by the machine, T2: seed ditches were established by the machine, T3: neither water nor seed ditches were established by the machine). The lodging index and lodging resistance traits, the grain yield and above-ground dry weight and the root traits were measured. The results showed that the lodging index was significantly affected by the treatments with ditches. The strongest lodging resistance was detected in mechanical hill wet-seeded rice with ditches treatment in both 2012 and 2013. The lodging resistance was strongly related to the breaking resistance, the root volume and root superficial area at the heading stage and maturity stage and the total root length at the heading stage. No significant difference was investigated in grain yield or dry weight of mechanical hill wet-seeded rice. Yuxiangyouzhan showed higher grain yield, dry weight and better lodging resistance but unfavorable root growth attributes than Peizataifeng. Therefore, the mechanical hill wet-seeded rice with ditches treatment increased rice lodging resistance is related to root traits.

Automated machine learning (AutoML) has been heralded as the next wave in artificial intelligence with its promise to deliver high performance end-to-end machine learning pipelines with minimal effort from the user. However, despite AutoML showing great promise for computer vision tasks, to the best of our knowledge, no study has used AutoML for image-based plant phenotyping. To address this gap in knowledge, we examined the application of AutoML for image-based plant phenotyping using wheat lodging assessment with UAV imagery as an example. We compared the performance of an open-source AutoML framework, AutoKeras in image classification and regression tasks to transfer learning using modern convolutional neural network (CNN) architectures. For image classification which classified plot images as lodged or non-lodged, transfer learning with Xception and DenseNet-201 achieved best classification accuracy of 93.2%, whereas Autokeras had 92.4% accuracy. For image regression which predicted lodging scores from plot images, transfer learning with DenseNet-201 had the best performance (R2=0.8303, RMSE=9.55, MAE=7.03, MAPE=12.54%), followed closely by AutoKeras (R2=0.8273, RMSE=10.65, MAE=8.24, MAPE=13.87%). Interestingly, in both tasks, AutoKeras models had up to 40-fold faster inference times compared to the pretrained CNNs. The merits and drawbacks of AutoML compared to transfer learning for image-based plant phenotyping are discussed.