Effects of Lighting Conditions on Imaging Quality for Agricultural Products

Machine vision is a new technique that interprets the surrounding ۳D world through ۲D image processing. Correct lighting can reduce image processing calculations and also increase the speed of operations. It also reduces costs and increases reliability. According to the importance of this issue, so far in all research, lighting has been contractually based on the experiences of others and a series of trial-and-error tests, without any justified reason for the style of lighting used. Therefore, to fundamentally solve the lighting problem, the COMSOL simulator software was applied. Three different lighting compartments including cubic, hemispherical, and vertical cylinders with two different white and black wall colors were defined based on their reflectance coefficient. Lighting was used according to the most common food and agricultural categories according to their geometrical form of an apple, egg, and cucumber. Also, three standard lighting arrangements including ۴ light sources in the corners of the compartment, circular plate, and square plate were examined in each case. For all the investigated conditions for each product, ۴ source lighting in the corners inside the white-walled hemisphere compartment was selected as the optimal lighting arrangement. Average light intensity and coefficient of variation obtained at the best conditions were measured for apple ۱۰۷.۴۸۳ W/m۲ and ۰.۵۳۷, for egg ۹۷.۳۷۲ W/m۲ and ۰.۴۷۹, and for cucumber ۱۰۰.۶۳۳ W/m۲ and ۰.۴۳۲, respectively. This study showed that the white wall color in all three compartments could provide more uniform light intensity than the black wall color

Keywords:

Lighting arrangement , Lighting compartment , Machine vision , Reflectance coefficient , Simulator software