Measuring image analysis attributes and modelling fuzzy consumer aspects for tomato quality grading

A mapping of fuzzy image analysis attributes to an overall visual quality reduces destroying tests and can easily be rearranged and optimised according to varying consumers' expectations.

Abstract

Quality grading is done by the consumer in a fuzzy way according to his senses sight, touch and smell. Visual appearance is the main source of information and can be broken down by image analysis to attributes like size, colour, shape, defects and abnormalities. Moreover, these attributes are often correlated with nutritional or other sensual quality parameters. Starting with such basic quality attributes, with the example of tomatoes a reasoning is proposed, mapping various fuzzy consumer aspects to overall quality classes. The objective is to achieve an automatic rating of fruit quality, modelling consumer aspects and producer needs. Such a mapping of fuzzy image analysis attributes to an overall visual quality reduces destroying tests. The reasoning can easily be rearranged and optimised according to varying consumers' expectations.