Non destructive monitoring of the yoghurt fermentation phase by an image analysis of laser-diffraction patterns: Characterization of cow’s, goat’s and sheep’s milk

• Published in: Food chemistry Vol. 274; pp. 46 - 54
• Main Authors: Verdú, Samuel, Barat, José M., Grau, Raúl
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 15.02.2019
• Subjects: Animals; Cattle; Diffraction pattern; Fermentation; Food Microbiology - instrumentation; Food Microbiology - methods; Food-Processing Industry - instrumentation; Food-Processing Industry - methods; Goats; Image analysis; Image Processing, Computer-Assisted - methods; Laser; Lasers; Milk; Milk - microbiology; Monitoring; Sheep; Spectral phasor; Texture; Yogurt - microbiology; Yogurt fermentation; Image analysis; Laser; Spectral phasor; Diffraction pattern; Yogurt fermentation; Texture; Monitoring; Milk
• ISSN: 0308-8146; 1873-7072
• DOI: 10.1016/j.foodchem.2018.08.091

•Yogurt fermentation was monitored by image analysis of diffraction patterns of laser.•Cow’s, goat’s and sheep’s milks were the milk types tested.•Structure of patterns was explored from which regions of interest were determined.•The captured variance from patterns affected both milk type and texture changes.•Models of the physico-chemical parameters were obtained for each milk type. Monitoring yogurt fermentation by the image analysis of diffraction patterns generated by the laser-milk interaction was explored. Cow’s, goat’s and sheep’s milks were tested. Destructive physico-chemical analyses were done after capturing images during the processes to study the relationships between data blocks. Information from images was explored by applying a spectral phasor from which regions of interest were determined in each image channel. The histograms of frequencies from each region were extracted, which showed evolution according to textural modifications. Examining the image data by multivariate analyses allowed us to know that the captured variance from the diffraction patterns affected both milk type and texture changes. When regression studies were performed to model the physico-chemical parameters, satisfactory quantifications were obtained (from R2 = 0.82 to 0.99) for each milk type and for a hybrid model that included them all. This proved that the studied patterns had a common fraction of variance during this processing, independently of milk type.