Mapping distributions of components in food through image analysis
Many foods and food materials have a non-uniform structure and composition. They may contain multiple components such as chocolate or caramel in a confectionery product, fruit pieces in yoghurt, structures such as marbling fat in meat, or oil and water in an emulsion, layers in packaging materials, or gradients of fat, water or salt. For measurement of average composition, methods such as grinding and blending are used to prepare representative subsamples for bulk analysis, or it may be possible to separate large components for individual analysis. However, the distribution of composition is also relevant in some cases, for example to identify specific components within the structure, to understand the function of particular components, to check that ingredients are uniformly dispersed or to study the migration of components such as water within a food product.
This white paper looks at the various ways in which this can be achieved. To discuss any of the techniques described or to find out more about how we can help you with any specific issue.
A wide range of imaging and microscopy methods are used to study the structure and appearance of food products. Several methods are also available to identify the composition of features within such images and to map the distribution of particular compounds. These include staining and fluorescence methods to highlight the location of features of interest, and advanced spectroscopic methods that provide compositional information about each pixel.
Colour is a familiar basis for recognising different components, for example to discriminate lean regions from fat in meat for assessment of marbling. Chromameters and spectrophotometers are widely used in the food industry for accurate measurement of average colour over an area of a sample. A common measurement scale is the CIELAB system, which provides a standardised description of colour by three parameters L*, a* and b*:
L* Lightness (0-100)
a* Variation from green (-100) to red (+100)
b* Variation from blue (-100) to yellow (+100)
Methods are available to calibrate colour imaging devices against such scales, enabling accurate colour measurements to be made for individual regions. Approaches include hyperspectral imaging using spectral cameras or a series of narrow-band illumination sources (e.g. Videometer), or calibration of a digital camera. At Campden BRI, we use a DigiEye imaging system (Verivide Ltd) for this purpose. The system includes an imaging cabinet with standardised lighting. Images are taken with a high resolution digital camera and the system is calibrated against a reference colour chart. The system can be used for colour measurement, and provides accurate images for documentation of process trials and production of specification images.