Journal of Food Engineering 61 (2004) 1–2

www.elsevier.com/locate/jfoodeng
Editorial

Computer vision––an objective, rapid and non-contact

Computer vision, or machine vision, is a novel tech- This issue starts with a major review by Brosnan
nology for acquiring and analysing an image of a real and Sun on the latest advances in the applications of
scene by computers and other devices in order to obtain computer vision in the agri-food industry. This sum-
information or to control machines or processes. The marises the fundamentals of computer vision and the
core technique in computer vision is always related to applications of computer vision in quality inspection of
image analysis/processing, which can lead to quantifi- bakery products, meat and fish, vegetables, fruit,
cation and classification of images and objects of inter- prepared consumer foods, grain and some other prod-
est within images. ucts.
The technology of computer vision is relatively young Segmentation is a very important step in image
with its origin traced back to the 1960s. Following an analysis/processing. Sun and Du developed a new stick
explosion of interest during the 1970s, computer vision growing and merging algorithm for segmenting complex
has experienced continued growth both in theory and food images such as pizza. This new algorithm devel-
application and has now expanded to various areas such oped has attractive strengths in computational effi-
as medical diagnostics, automatic manufacturing and ciency, speed, and generality. Application of computer
surveillance, remote sensing, and autonomous vehicle vision in meat quality evaluation is an active research
and robot guidance. With the developments in hardware area in recent years. Tan presents a review paper on the
and software for image analysis/processing, the appli- main results from several recent studies including char-
cations of computer vision have been extended to the acterisation of meat quality attributes such as colour,
quality evaluation of diverse and processed foods, which marbling, maturity, texture and tenderness, and pre-
has illustrated great advantages of using the technology diction of sensory score and grades. Cheese is one of the
for objective, rapid, non-contact and automated quality most widely used food products. Cheese shredding fur-
inspection and control. With these developments, com- nishes improved functionalities because of faster melt-
puter vision is expected to replace human vision for the ing. Ni and Guansekaran use a modified sequential
purpose of providing quantitative information and au- thinning algorithm for cheese shred evaluation. The
tomating certain processes according to their visual algorithm performs well in recognising and merging oc-
appearance. cluded segments in shred-like objects. Melting, brown-
In recent years, computer vision has been attracting ing and oiling off are the main functional properties of
much R&D attention from the agri-food industry, and cheese, further to their previous research work on
rapid development has been increasingly taking place on melting and browning, Wang and Sun reports their
quality inspection, classification and evaluation of a wide latest research results using computer vision to evaluate
range of agri-food products. Therefore, it is timely that the oiling off property of cheese and the results are
this special issue on ‘‘Applications of Computer Vision in correlated well with the fat ring test traditionally used in
the Food Industry’’ for Journal of Food Engineering the industry.
focuses on these advances. In this issue, each of the 14 One of the early applications of computer vision is
papers is written by an international expert (or experts), the detection of fruits, among which apple is widely
addressing in detail one aspect of the relevant technology studied. The grading of apples into quality classes is a
and therefore highlighting the truly international nature complex task. Mehl, Chen, Kim and Chan discuss using
of the work. Therefore this issue should provide valuable a high spatial resolution hyperspectral imaging system
information for the food engineer. It should also serve as as a tool for selecting better multispectral methods
an essential and comprehensive reference source to un- to detect defective and contaminated surfaces of four
dergraduate and postgraduate students and researchers varieties of apples. Leemans and Destain adopt the
in universities and research institutions. computer vision technology for grading of apple based

0260-8774/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0260-8774(03)00182-1
2 Editorial / Journal of Food Engineering 61 (2004) 1–2

on features extracted from surface defects. Besides ogy, they develop a prototype automated inspection
apples, computer vision is also employed to study the system for colour grading of oil palms; they also apply
quality of potatoes and table olives. Thybo, Szczypinski, computerised radar tomography to map moisture con-
Karlsson, Donstrup, Stodkilde-Jorgensen and Andersen tent of grain. This special issue is concluded with a paper
perform NMR-imaging on five raw potato varieties in by Yam and Papadakis, in which a simple method using
combination with different image analysis methods to computer vision to measure and analyse the surface
predict the sensory texture quality attributes of cooked colour of food products is described.
potatoes; and Diaz, Gil, Serrano, Blasco, Molto and Finally, as the Guest Editor, I would like to thank all
Blasco compare three algorithms for classifying table the authors for their valuable contributions, and to all
olives in four categories depending on the surface defects the referees for their diligence and helpfulness in re-
of the fruit. viewing these papers. I am also grateful to Prof. Brian
New areas of computer vision applications in the McKenna for his support and encouragement while
food industry are being developed. Hatcher, Symons developing this special issue.
and Manivannan develop an image analysis system to
determine appearance and colour of oriental noodles. Da-Wen Sun
Measurement of bubble haze during beer dispense pro- FRCFT Group
cess is of significant value to the brewing industry for Department of Agricultural and Food Engineering
quality control and product development. For this, National University of Ireland, Dublin
Hepworth, Hammond and Varley study using computer (University College Dublin)
vision for measuring bubble size distributions in beer. Earlsfort Terrace
Furthermore, Abdullah, Guan, Lim and Karim dem- Dublin 2 Ireland
onstrate the applications of imaging technologies for Tel.: +353-1-7165528; fax: +353-1-7167493/4752119
quality control of agricultural materials in a visible and E-mail address: [email protected]
non-visible spectrum. Using computer vision technol- URL: http://www.ucd.ie/~refrig