International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
the localized colour has to be used. Similarly, the average gray
scale for the localized colour can be calculated by the
summation of gray scales falling within the limits of the upper
and lower thresholds for the localized colour and dividing this
sum by the number of their corresponding pixels.
4. CONCLUSIONS
The goal of this study was to investigate specific phases of the
design and implementation of vision-based inspection system to
improve the quality of some baked products namely Arabic Pita
and Mexican tortilla breads, and to improve the efficiency of
production lines of these products. The investigation, in this
context, concentrated on developing and implementing methods
and procedures to measure, process and analyze visual data and
geometric characteristics of baked products that can be used to
implement a vision-based quality control system for these
baked products. From the experience gained in this project, it is
clear that the visual inspection process involves three main
distinct stages; these are visual data collection, data processing
and feature extraction, and control decision making. This
investigation was done concentrating on the first two stages.
The core of the visual system for this study was a 5-megapixel
camera, which facilitated the acquisition of high resolution
images up to (1500x3000) pixels. Although these high
resolution capabilities enable the production of high definition
images with the finest of details, however, in relation to the
available hardware and its computing capabilities, and specially
when trying to use a C++ code to carry out the image
processing operations, it was clear that the produced images
exceed the capabilities of the visual system's unit responsible
for software execution and feature extraction. In many cases
and in order to be able to process the images in near-real time, it
was unavoidable but to crop the images around the objects that
undergo feature extraction process, this cropping process results
in reducing the size of image matrices to be processed so
speeding up different processing and analysis operations.
Feature extraction procedures developed for this investigation
are suitable for many integrated shape-colour pattern
recognition applications. A preliminary investigation will be
carried out on the possibilities and problems associated with
implementing visual feature-based cashier system for food
store.
References:
Gonzales R. C., and Woods R. E. (1992): Digital Image
Processing, Addison-Wesley Pub. USA. p. 716.
ICD., (2003) Industrial Camera Directory- Camera
Manufacturers, Products and Specifications. Vision Systems
Design, vol. 8., no. 9, pp. 20-55.
MathSoft Corporation (1999): Help Text and Built in Functions
of the Image Processing Extension Package.
Parker, J.R. (1991) Gray Level Thresholding in Badly
Illuminated Images, IEEE Trans. Pattern Anal. Machine Intell.,
vol. 13, no. 8, pp. 225-233.
Perez, A, and Gonzalez, R.C. (1987): An Interactive
Thresholding Algorithm for Image Segmentation., IEEE Trans.
Pattern Aanal. Machine Intell., vol. PAMI-9, no. 6., pp. 742-
7581.
Prat, W.K. (1978): Digital Image Processing, John Wiley and
Sons, NY, USA.
Rosenfeld, A., and Kak, A. C. (1982) Digital picture
processing, 2nd ed., Academic Press, Ny, USA.
482
White, J.M., and Rohrer, G.D. (1983); Image Thresholding for
official character Recognition and Other Applications
Requiring Character Image Extraction, IBM Journal of
Research and Development, vol. 27, no. 4., pp. 400-411.
Wilson, A. (2003): Smart Cameras Embed Processor Power,
Vision Systems Design, vol. 8., no. 9, pp. 95-99.
Zhoi, L., Chalana, V., and Kim, Y. (1998): PC-based Machine
Vision System for Real-Time computer-Aided Potato
Inspection, Journal of Imaging system Technology, vol. 9, pp.
423-433.