Application of computer vision systems in colour evaluation of kunda: a heat desiccated dairy product.

摘要

Two computer vision methods were compared for measuring the colour of kunda, a traditional Indian dairy product with a brown colour that is prepared by thermal drying of sweetened milk. Kunda was placed onto a suitable transparent material and its surface was scanned using the scanner, and also photographed using a digital camera under defined conditions. Images obtained using either scanning or photography were visualized using Adobe Photoshop software and colour parameters (red, green and blue) were measured using the Graphical Editor of Photoshop. Overall, the lightness, red, green and blue values obtained using the scanner method were in the ranges 124.43-127.7, 162.13-163.3, 116.24-119.94 and 67.51-73.71, respectively, on a scale of 256. Scanner variables had little impact on the magnitude of these colour parameters but the material on which the kunda was scanned had to be specified. For the photographic method the lightness, red, green and blue values were in the ranges 98.82-100.24, 137.99-141.26, 89.97-91.26 and 34.87-36.82, respectively. These values depended on the resolution of the camera used, the distance between the object and the lens and the illumination. The accuracy of the methods was indicated by low coeff. of variation values. It is concluded that computer vision can be used as an objective, rapid and non-contact method to evaluate the colour of kunda.