Digvir S. Jayas 2017-02-22 23:39:19
A sample of harvested lentils. How would you assess the quality? When people purchase raw ingredients or processed food products, they use intuitive methods to assess the quality of the food. For example, when purchasing whole lentils, the consumer may look for the presence of non-lentil materials, variations in the lentil size and color, or damage due to insects or fungus. Similarly, a food processing plant that converts raw ingredients into packaged food products may monitor these factors and try to eliminate them through unit operations such as cleaning, sizing, and color grading to produce clean lentils ready for packaging. For proper functioning of these unit operations, the processor may monitor the quality factors manually or use instruments before and after the unit operations. Thus, high-quality raw ingredients and unit operations are the two main components that contribute to the preparation of healthy, safe, and wholesome food products for consumers. No unit operations can produce high-quality processed food products from poor-quality raw ingredients. Therefore, production, preservation, and handling of the raw ingredients must be done with great care, and the quality of the ingredients must be maintained until they are processed as food products for consumption. The Hazardous Analysis at Critical Control Points (HACCP) management program is commonly implemented to enhance food safety by identifying and eliminating biological, chemical, and physical hazards in raw materials and in the processing and distribution of food products. For proper implementation of HACCP, as well as to control the unit operations, food quality parameters must be measured at many points in the food chain from producer to consumer. Sensors (which convert a parameter of interest, such as color, pH, or moisture content, into a measurable quantity, such as electric current, resistance, or voltage) and bio-imaging (which captures images of samples in certain ranges of the electromagnetic spectrum and coverts this information into features representing product quality based on intrinsic or extrinsic factors of the product) are the main tools for determining the quality parameters of food products. Sensors Many types of sensors are used in the food industry, including electrochemical, electromechanical, physical, and biological sensors (also called biosensors). Electronic noses and electronic tongues, which incorporate an array of gas sensors, can also be included in the sensor category. Sensors are used for assessing the freshness or ripeness of fruits and vegetables, determining the aroma and color consistency of foods, and detecting insect infestation, fungal infection, foodborne pathogens, and chemical contaminants. Biosensors are becoming more popular in the food industry because they can be scaled to small sizes for incorporation into processing lines and even food packaging. This size reduction is possible due to the use of antibodies, peptides, and phages (viruses that infect bacteria) for detecting contamination. Sensors can monitor the product at many points in the process and control the unit operations to produce consistent quality from inconsistent raw ingredients, as shown in the accompanying schematic. Bio-imaging Bio-imaging means taking images of samples under proper illumination and then enhancing and processing the images to extract features that describe the quality of the food product. Using classification tools, the food product can then be separated into different classes. Food products can be imaged in the visible spectrum (400-700 nm), which is similar to what people see when they look at a product, but people cannot extract separate features to characterize food materials. How the human brain processes information is still a mystery, but computers must use quantification. To analyze an image, the computer must measure many features, such as length, width, perimeter, projected area, color, and color distribution, to create vectors that are unique to different foods or to different defects in food products. As an example of the complexity of this process, I am astonished that my two-year-old granddaughter, Priya Jayas, can identify shapes (triangles, squares, and circles) of different sizes and colors with 100% accuracy. I am certain that she is not systematically quantifying the shape features to make her decision. Somehow, she just knows the shapes. In contrast, to perform the same task, a computer would need to calculate the perimeter, area, length, and width of a bounding rectangle, and the mean and variance of the radius (the distance from the centroid to every pixel on the boundary of the image) for every object, and that would take much more time than Priya needs. When adults look at food products, they can quickly characterize any visible defects. To detect defects that are less visible, we use imaging in different spectral ranges, such as thermal, x-ray, near-infrared, and infrared, and we use the spectral characteristics of the product to quantify defects. The mean reflectance of an object is one of these spectral characteristics. We can also collect spectral information over a broad range of wavelengths to create spectral signatures that distinguish the characteristics of food products from defective components or extraneous materials. The differences in these spectral signatures are not always obvious, as shown in the accompanying graph. In such cases, we can use statistical, artificial neural network, and fuzzy logic methods to characterize the differences for defect detection. Similar to sensors, bio-imaging, on its own or in combination with sensors, can detect invisible defects in food products and control unit operations in food processing. Thus, it is critical to incorporate this useful technology into the production of high-quality food products. ASABE Fellow Digvir S. Jayas, P. Eng., Vice-President (Research and International) and Distinguished Professor, Department of Biosystems Engineering, University of Manitoba, Winnipeg, Canada, email@example.com.
Published by ASABE. View All Articles.