Image analysis enables the on-line measurement of essential values such as bean color and surface area. However, it is difficult to apply this technique to coffee roasting. In this industry, color (i.e. brightness) is a key variable which determines the quality of the final product, but it is evaluated experimentally in the laboratory by the roast master. In order to control the process, it is necessary to develop a technique which enables a real-time assessment of product quality. This work proposes a method to determine brightness and surface area using image analysis, on-line during coffee roasting. The changes to brightness and surface area that affect coffee bean during the process were studied experimentally in a roaster equipped with a CCD video camera, a lighting system and two small fiber optic spotlights. Arabica green coffee beans of Colombian origin were roasted using different air temperatures (190,200,210, . . . 300 degrees C) for 10 min. The experimental results provided knowledge of the process, and also, allowed us to observe the behavior of color and surface kinetics. It was also possible to perform experiments real-time using image analysis system, stopping the roasting process when the results were optimum. When the bean temperature reached 280 degrees C, the surface area stopped increasing, so that roasting was also complete. Brightness and surface area may be two indicators to ensure on-line quality during coffee roasting. (c) 2008 Elsevier Ltd. All rights reserved.