Smart Agricultural Technology (Dec 2024)

Development of low-cost multifunctional robotic apparatus for high-throughput plant phenotyping

  • Jiří Mach,
  • Lukáš Krauz,
  • Petr Páta,
  • Martin Halecký

Journal volume & issue
Vol. 9
p. 100654

Abstract

Read online

The assessment of plant health and quality is a critical aspect of plant biology, agriculture, and the food industry. With the global population continuously increasing, the demand for high-quality plant products is expected to surge. Consequently, there is a necessity for the development of automated systems that are user-friendly, accurate, affordable, and capable of rapid evaluation of plant health in a multitude of settings, including fields, farms, and laboratories. Such systems would apply to various applications, including investigating the impact of environmental conditions or developing new potential biostimulants and biopesticides. This paper introduces a novel, low-cost, and innovative multifunctional high-throughput plant phenotyping system that addresses an unmet need in the market. The system was designed to be affordable, scalable, applicable, and reliable to diverse customers. The key components of the system include a robotic arm AR4 (Annin Robotics, USA), a three-dimensional (3D) scanner POP 3 (Revopoint, USA), and a multispectral (MS) visible near-infrared (VNIR) camera FS 3200D 10GE (JAI Ltd., Japan). The paper describes these devices, their calibration, performance evaluation, and final applicability assessment. In particular, the motion characteristics of the AR4 are evaluated by a pose repeatability with obtained values of Image 1 and Image 2 without load and with load, respectively. Furthermore, it is shown that the calibrated camera provides comparable NDVI index data with the content of plant pigments (R2>0.92) and also against the reference VNIR hyperspectral (HS) camera SPECIM PFD4K-65-V10E (R2>0.99). In addition, the presented 3D scanner demonstrated superior 3D models with a high degree of fit compared to the more expensive 3D scanners such as Shining 3D EinScan Pro 2X 2020 or Shining 3D EinScan-SP V2. The paper concludes with a discussion of the results, limitations, future improvements, and potential applications of the device in laboratory, educational, and field settings.

Keywords