Articles
Leaf area estimation of strawberry plants using commercial low-cost LiDAR
Article number
1360_3
Pages
23 – 28
Language
English
Abstract
Phenotyping technology still lacks implementation in commercial production of fruit and vegetables so far, possibly due to high investment costs for the sensors.
In the present study, two commercial low-cost LiDAR systems (Intel® RealSense LiDAR Camera L515, USA and Nimbus 3D ToF camera, Pieye GmbH, Germany) were applied for estimating the leaf area of strawberry plants.
Geometric accuracy of sensors was tested with reference white spheres of known diameter located in a growth chamber at different height and position.
Six strawberry plants were measured with both systems on three measuring dates. 3D point clouds obtained with the low-cost sensors were analysed at different grid size (5×5, 10×10, 18×15 cm), where maximum plant height and leaf area were calculated.
Results were compared to manually measured plant properties and state-of-the-art LiDAR (SICK LMS 511 Pro, Germany). For plant height analysis, R2=0.74 and R2=0.53 were obtained by RealSense and Nimbus, respectively.
Both the sensors performed rather poorly in estimating leaf area of strawberry plants with R2<0.17. In comparison, analysis with SICK LMS 511 resulted in R2=0.95 and R2=0.77 for plant height and leaf area estimation, respectively.
Consequently, simple plant growth monitoring can be approached with low-cost sensors, whereas phenotyping questions request enhanced geometric data quality.
In the present study, two commercial low-cost LiDAR systems (Intel® RealSense LiDAR Camera L515, USA and Nimbus 3D ToF camera, Pieye GmbH, Germany) were applied for estimating the leaf area of strawberry plants.
Geometric accuracy of sensors was tested with reference white spheres of known diameter located in a growth chamber at different height and position.
Six strawberry plants were measured with both systems on three measuring dates. 3D point clouds obtained with the low-cost sensors were analysed at different grid size (5×5, 10×10, 18×15 cm), where maximum plant height and leaf area were calculated.
Results were compared to manually measured plant properties and state-of-the-art LiDAR (SICK LMS 511 Pro, Germany). For plant height analysis, R2=0.74 and R2=0.53 were obtained by RealSense and Nimbus, respectively.
Both the sensors performed rather poorly in estimating leaf area of strawberry plants with R2<0.17. In comparison, analysis with SICK LMS 511 resulted in R2=0.95 and R2=0.77 for plant height and leaf area estimation, respectively.
Consequently, simple plant growth monitoring can be approached with low-cost sensors, whereas phenotyping questions request enhanced geometric data quality.
Authors
N. Singh, K.K. Saha, P. Makaram, M. Zude-Sasse
Keywords
3D point cloud, closed environmental agriculture, resilient food systems
Groups involved
- Division Precision Horticulture and Engineering
- Division Plant-Environment Interactions in Field Systems
- Division Vine and Berry Fruits
- Division Tropical and Subtropical Fruit and Nuts
- Division Plant Genetic Resources, Breeding and Biotechnology
- Division Temperate Tree Nuts
- Division Temperate Tree Fruits
- Division Vegetables, Roots and Tubers
- Division Sustaining Horticulture in a Changing World
- Working Group Mechanization, Digitization, Sensing and Robotics
Online Articles (45)