Near-infrared technique for oil palm fruit grading system

The Malaysian palm oil industry is considered to be highly regulated.A major problem faced by oil palm producers is the accurate grading of fresh oil palm fruits according to their ripeness levels before processing.Classification of oil palm fresh fruit bunch (FFB) maturity is a critical factor that dictates the quality of produced palm oil.The human eye, for example, has historically judged quality via appearances.External features and properties such as color, texture, shape, and size are good indicators for parameters like ripeness.Hence, grading system technologies offer a solution to these problems. The grading systems in general utilized improved engineering designs with image processing techniques to ensure the quality of the product.In this research, a hyperspectral oil palm grading system was built and an image processing techniques algorithm was developed based on the spectral reflectance of the external features of oil palm fresh fruit bunches (FFB). Color changes resulting from biochemical reactions in fruit texture can be related to fruit maturity . In addition, the oil palm fruit pigments such as carotenoids and chlorophylls and their ratios affect the color of the oil palm fruit. Underripe fruits have a higher proportion of chlorophyll that gradually decreases upon maturity. Similarly, carotenoids increases as the oil palm fruits mature. These color and biochemical changes can be observed utilizing the spectral reflectance of the fruit. Using the FFB spectral reflectance this research was used to modify and adapt the hyperspectral scanner to enhance its suitability for the maturity detection of oil palm FFBs at the near-infrared (NIR) range (400 nm to 1000 nm).This objective was achieved by improving the illumination system of the hyperspectral scanner. The strategic positioning of the halogen and applied security design (ASD) lamps helps provide a shadow for free illumination. Image processing approaches, such as image acquisition, image pre-processing, and image feature extraction, as well as image classification were developed to automate the ripeness grading for oil palm fruit bunches. The mathematical model was developed to determine the real value of the reflection of specific wavelengths for the three categories of oil palm FFBs through regression analysis. The results are then confirmed by a trained human grader. The application software was developed in a MATLAB 7.0 environment, and was used to classify the oil palm FFBs. The data collected by this system are subjected to the artificial neural networks (ANN), kernel nearest neighbor (KNN), support vector machine (SVM) techniques, and a number of statistical analyses such as the CHAID method and one-way ANOVA for oil palm FFB classification. The developed system showed high classification results on accuracy of the maturity detection for the three types of oil palm fruits (nigrescens, virescens, and oleifera ) with rates of 95%, 99%, and 98 %, respectively, using the ANN-MLP classifier; rates of 96%, 99%, and 98 %, respectively, using the KNN classifier; and rates of 76%, 96%, and 94%,respectively, using SVM. Based on the results of testing hyperspectral with the scientific results of bands of overripe, underripe, and ripe we fabricated the multiband sensor. This multibandactive sensor has the ability to detect the maturity of oil palm fruit at 735, 750, 780, and 940 nm). The multiband sensor was field tested and can categorize the oil palm FFB into three classes: overripe, ripe, and underripe. The system helps increase the accuracy of the oil palm FFB grading system, which will be useful for the oil palm industry, oil palm engineers, mill operators, plantation managers, small holders, and the research community.

Preview PDF FK 2013 24R.pdf Download (1MB) | Preview

Actions (login required)

View Item