Strawberry Biomass and Leaf Area Index Modeling Using Ground-Based Image Analysis

Guan Zhen

Guan Zhen Plant City Campus of the Gulf Coast Research and Education Center
Amr Abd-Elrahman Plant City Campus of the Gulf Coast Research and Education Center
Vance Whitaker University of Florida, Institute of Food and Agricultural Sciences
Shinsuke Agehara University of Florida, Institute of Food and Agricultural Sciences
Ali Gonzalez Plant City Campus of the Gulf Coast Research and Education Center
Katie Britt Plant City Campus of the Gulf Coast Research and Education Center

15G

Remote sensing imagery has been used extensively in agricultural applications such as plant stress detection, yield estimation, and field management. A ground-based mobile mapping system utilizing two consumer-grade fixed focal plane cameras (RGB and IR) was used to retrieve leaf area index and biomass of strawberries. Imaging missions were conducted weekly along the strawberry growing season (from 11-16-2017 to 02-27-2018 ). Leaf area, fresh weight, and dry weight of the plants were measured using destructive methods several hours after image acquisition. The captured close-range images were analyzed to produce high spatial resolution (1mm) RGB and IR image mosaics as well as dense point clouds and digital surface models. The orthoimages, digital surface models, and point cloud information were used in object-based image analysis to classify the field into canopies, soils, plastic beds, shadows, and weeds. Surface leaf area, planimetric canopy area, surface leaf angle, and canopy volume were derived from the remotely sensed data, and regression analysis (R^2>80%) was carried out to model the leaf area and biomass measured through in-situ, destructive plant data sampling. Other independent variables describing canopy structure such as standard deviation of leaf object elevation, canopy volume, LAD , canopy density, and point cloud metrics were tested to improve models’ goodness of fit as canopies grow larger. Furthermore, multiple cultivars and breeding selections were tested in the regression model to examine how the fit would change when genotype varies.

15:55 Strawberry Biomass and Leaf Area Index Modeling Using Ground-Based Image Analysis, Guan Zhen

January 29 @ 16:00
16:00 — 16:05 (5′)

Quartz AB

Guan Zhen

Add to Google Calendar