Intelligent measurement of the colour of food products

Food colour is an important attribute of food quality. This paper discusses issues in using machine vision for the objective assessment of food colour, such as colour image segmentation, calibration and statistical parameter design, and their application to pea colour assessment. The algorithm first established a 3D histogram on RGB colour space based on an octree data structure. From the colour space, a statistical index was calculated to measure the major colour of an object simulating the characteristics of human colour perception. This method was applied to assess the major colour of peas. Accuracy and efficiency were achieved in measuring pea seed colour and its variations.