Search and Access Archived Conference Presentations

2018 ASHS Annual Conference

Automated and Non-Destructive Measurement of Plant Growth Characteristics Using a Multispectral Image Based Technique in Controlled Environment Agriculture (CEA)

Thursday, August 2, 2018
International Ballroom East/Center (Washington Hilton)
Alexander Miller, Purdue University, West Lafayette, IN
Ranjeeta Adhikari, Graduate student, Purdue University, WEST LAFAYETTE, IN
Krishna Nemali, Purdue University, West Lafayette, IN
Intensive farming, high plant densities, and multi-level production makes it challenging to monitor plant growth and input use in controlled environment agriculture (CEA, greenhouses and vertical farms). With the advancements of imaging technology, remote observations using drones is already used in outdoor agriculture systems. Image based measurements using cameras in CEA can make monitoring crops more efficient. Our objective was to test whether a multi-spectral image analysis technique can be used to remotely and automatically measure plant growth characteristics in CEA. Experiments were conducted in a greenhouse maintained at 26/20 ºC (day/night) temperature and daily light integral (DLI) of 10-20 mol·m-2·d-1. Leaf lettuce (Lactuca sativa L. var. Black Seeded Simpson) and tomato (Solanum lycopersicum var. Early Girl) were grown in a peat-based substrate (Sungro Professional Mix) and supplied with a liquid fertilizer comprised of 15-5-15 and 21-5-20 mixes in 3:1 ratio. Leaf lettuce and tomato plants were grown under non-stress conditions in the experiment. A multi-spectral imaging station with image analysis software was used for image acquisition, plant pixel segmentation, and canopy area estimation. Image based canopy area (LAimage) was measured for groups of plants on different days. Immediately after imaging, direct measurements included total leaf area (LAactual), shoot dry weight (SDWactual), and relative growth rate (RGRactual). There was a linear relation between LAactual / SDWactual and LAimage, and RGRactual and image derived RGR (RGRimage) in leaf lettuce and tomato. These results indicate that image-based measurements can non-destructively measure plant growth characteristics. Automated non-destructive growth measurements using camera based image systems can be mounted on moving systems (ex: booms) or placed at stationary locations in CEA. These systems can aid in early detection of production related issues affecting growth, thereby leading to increased productivity in CEA.