MARI held the second interdisciplinary academic salon from 15:30 p.m. to 17:00 p.m. on April 17th, 2019. The invited lecturers, who shared crop yield estimation methods in different disciplines， were Dr. Nanyan Deng of College of Plant Science and Technology, Professor Xiaoheng Zhang of College of Economics and Management, Professor Baodong Xu of College of Resources and Environment and Professor Lingfeng Duan of College of Engineering. The commentators were Professor Ming Zhan of College of Plant Science and Technology and Professor Yuanyun Ling of College of Economics and Management. This activity attracted more than 30 teachers and students from College of Plant Science and Technology, College of Economics and Management, College of Resources and Environment and College of Engineering.
Dr. Nanyan Deng explained the principles and methods of predicting final crop yield by crop growth model from three processes: determining plant growth period, simulating biomass allocation in different periods and predicting final yield. Professor Lingfeng Duan reported on the nondestructive measurement of field plots based on deep learning. She proposed using Panicle-SEG and PancleNet technology to achieve accurate segmentation of multi-variety rice panicles in complex field environment, and combined it with rice agricultural knowledge to extract rice image features directly related to yield, so as to establish a more accurate yield measuring model. Professor Baodong Xu introduced the method of estimating crop yield in wide-range and long time-series by using new remote sensing technology. He explained three methods of estimating crop yield by remote sensing: empirical relationship methods, remote sensing-plant physiological process method and remote sensing-crop growth model assimilation method, and put forward relevant opinions on the challenges and future development direction they are facing. Professor Xiaoheng Zhang reported on the content of production function model. He pointed out that agricultural production was the interweaving of natural reproduction and economic reproduction, and analyzed the impact of input of production factors on grain yield.
Two commentators commented on and extended the reports of the four lecturers. From the peasant’s point of view, Professor Ming Zhan shared his views on how to improve the actual yield of the field by using yield estimation methods. Professor Yuanyun Ling higly appreciated the necessity of combining different disciplines in yield measurement. He suggested that remote sensing technology can be applied to crop growth model to predict yield more accurately and efficiently. In the process of free discussion, Professor Hang Xiong pointed out that four estimation methods can be classified into two categories: measurement-based and model-based. Model estimation can be classified into natural and economic aspects, while remote sensing estimation can be classified into low-altitude and high-altitude remote sensing. He proposed that production estimation under the background of big data can improve the accuracy and applicable space of estimation by applying big data to calibration of model parameters. Professor Shi Min believed that in yield estimation, it is necessary to combine the production function and crop model, consider natural and economic factors, and make full use of geographical big data. After discussion, it was agreed that data is very important in achieving yield estimation, and the obtainment of data at household level need long-term solid field investigation.
The comfortable seats and lean environment in digital reading space of the library made the atmosphere relaxed and warm. Every teacher and student was absorbed in the rich interdisciplinary knowledge. At the end of the meeting, everyone expressed the recognition and appreciation of this interdisciplinary exchange. They hoped for more opportunities to participate in similar activities in the future.
Written by Ziwei Li | Review by Hang Xiong