Field Robotic Systems for High-Throughput Plant Phenotyping: A Review and a Case Study

AbstractContinuous crop improvement is essential to meet the growing demands for food, feed, fuel, and fiber around the globe. High-throughput plant phenotyping (HTPP) aims to break the bottleneck in plant breeding programs where phenotypic data are mostly collected with inefficient manual methods. With the recent rapid advancements and applications of robotics in many industries, field robots are also expected to bring transformational changes to HTPP applications. This chapter presents an updated review of the infield ground-based robotic HTPP systems developed so far. Moreover, we report a case study of an autonomous mobile phenotyping robot PhenoBot 3.0 for row crop phenotyping, focusing on the development and evaluation of the navigation system for the articulated steering, a four-wheel-drive robot with an extremely tall sensor mast. Several navigation techniques were integrated to achieve robustness at different corn plant growth stages. Additionally, we briefly review the major sensing technologies for field-based HTPP and present a vision sensor PhenoStereo to show the promising potential of integrating conventional stereo imaging with the state-of-the-art visual perception techniques for plant organ phenotyping applications. As an example, we show that a highly accurate estimation of sorghum stem diameter can be achieved with PhenoStereo. With this chapter, our goal is to provide valuable insights and guidance on the development of infield ground robotic HTPP systems to researchers and practitioners.KeywordsHigh-throughput plant phenotypingField roboticsRobot Operating SystemNavigationStereo visionComputational perception