Imaging technique for quantification of oil palm fruit ripeness and oil content

We investigated a non‐destructive technique for measuring the colour of oil palm fruit (Elaeis guineensis) bunches at different stages of maturity and correlating the colour data with the oil content in the fruit bunch. A digital red, green and blue (RGB) camera was used to capture the image of palm fruits and hue histograms were then constructed from mathematically transformed RGB data. The dominant hue peaks were then correlated to the oil content of fresh fruit bunches (FFBs). The study showed that the results obtained were influenced by FFB that possessed different expressions of colour changes during the ripening process. The results obtained from colour image analyses were not satisfactory – the correlation between hue peak and FFB oil content was r = 0.7933. However, when FFB having the same ripening patterns were analysed, the correlation was very much improved with r = 0.9519. Thus, statistical evaluation showed that the digital imaging technique for determining FFB oil content can be successfully used on a homogeneous population of palms that display similar change of colour during the ripening process. Practical applications : The ability to correctly identify physiological maturity and harvest maturity of palm fruits by phenological characteristics will ensure timely harvest to avoid cutting of either under‐ and over‐ripe fruits. The Malaysian Palm Oil Board MPOB] 1 has published an oil palm fruit grading manual outlining 16 classifications for grading FFB ripeness and quality. However, the FFB grading methods are still very subjective, i.e. depend on the judgement, experience and skill of a person. In addition, the methods do not provide quantitative information about the oil in the FFB. The method reported in this paper allows for rapid screening of FFB ripeness, and relate it to oil content in the FFB and accordingly the amount of oil that can be extracted from a consignment of FFB. Since it only takes a few seconds to capture signals, the system could be incorporated for on‐line monitoring of FFB ripeness and oil content thereof.