บทคัดย่องานวิจัย

หมวดหมู่ : ฐานข้อมูลงานวิจัย

Spectral imaging VIS-NIR system to forecast the chilling injury onset on citrus fruits

Menesatti P.; Carzaniga S.; Urbani G.; Stanco F.; Lanza G. and Di Martino E.

5th International Postharvest Symposium . Volume of Abstract . Verona, Italy 6-11 June 2004, p.116

2004

บทคัดย่อ

Spectral imaging VIS-NIR system to forecast the chilling injury onset on citrus fruits    Cold storage is o­ne of the most important and useful opportunities to preserve citrus fruit quality and to extend the sales period. Citrus fruits, however, are susceptible to chilling injury (CI) depending o­n species, cultivar, and the length of exposure to low temperatures. Susceptibility is also influenced by many preharvest factors and harvest date. Blood oranges and lemons may develop chilling injury when held at temperatures below about 8 and 12°C respectively. Symptoms of chilling injury include pitting, brown staining, increased decay and quality deterioration.

Aim of the work was to forecast the chilling injury development by non-contact hyperspectral VIS-NIR imaging o­n the fruit rind at least 10 days in advance before its visual appearance. The analysis was made o­n 129 fruits of ‘Femminello siracusano’ lemons (Citrus limon [L.] Burm.) stored at 5°C and high relative humidity for 94 days. Spectral images were made 12 times. The opto-electronic system for the images acquisition was made by a VIS-NIR spectral scanner (DV Padova), which can acquire bidimensional spectral images with a very good optical (250 Kpixels) and spectral (115 l) resolution. The acquired image allows readings of the standard reflectance values (white and standard illuminating or illumination) for the VIS-NIR spectral rage (400-970 nm), with a 5 nm bandwidth.

After 94 days, 79% of the fruits showed damaged areas. Partial least squares (PLS) regression analysis was applied for constructing a predictive model based o­n the spectral normalized response, constructing the model o­n a sub sample and verifying the model (test) o­n independent o­nes. Hyperspectral PLS model allows to predict chilling injury o­nset o­n lemons 20 days before the visual appearance with a percentage of 70%