Détermination du potentiel de conservation de jeunes pousses de légumes-feuilles par imagerie hyperspectrale

Vegpro International inc. is the largest fresh vegetable producer in Canada. Vert nature inc., its corporate farm dedicated to baby greens conventional production, produces approximately 5800 acres of baby greens annually, including 3500 acres in Québec and 2300 acres in Florida. Vegpro International inc. produces and distributes up to 15 different ready-to-eat baby greens blends. Following harvest, vegetables are stored in bulk for up to 5 days, and are packed in plastic containers for sale. A shelf life of 12 days must then be guaranteed by the producer. So, each component of the blend needs to maintain its freshness during this period of time, which may represent an important challenge in stock management. Produce shelf life is influenced by many parameters specific to the commodity (respiration rate), climatic conditions (rainfall, freezing), and plant diseases incidence. Baby greens degradation is associated to a gradual change in color, as well as the increase of free water caused by cell destruction. These changes can be measured by a hyperspectral imaging technique, as it evaluates the intensity of diffusely reflected light from a surface at different wavelengths, inducing a spectral signature specific to the product. This research project aims to evaluate the potential of the hyperspectral imaging technique to predict baby greens’ shelf life in order to facilitate stocks management. Four commodities will be evaluated, arugula, spinach, kale and lettuce. Ten packages of each commodity will be analyzed within 24 h following their packaging. All 40 packages will then be stored at 4 degrees Celcius and analyzed every day for 20 days. Packages will be taken out separately in order to maintain the storing conditions as long as possible and images will always be taken the same way for each package. Hyperspectral imaging system will be adjusted to optimize the spatial resolution and the image quality for wavelengths from 400 through 1000 nm (resolution of 2.8 nm), as well as from 1000 through 2500 nm (resolution of 10 nm) according to light intensity. High quality color images will also be taken every day in order to evaluate the general state and degradation of commodities. This assay will be repeated three times during summer 2017 (July – September) and a total of 120 samples will be analyzed. Field conditions during greens production will be recorded (temperature, rainfall), as well as agronomic observations (diseases, pests) in order to establish a link with the produce spectral signatures. This project is the first step before introducing the technique in packaging and transformation plants.