Improving oil palm fresh fruit bunch grading system via software and hardware modifications

An improved technique is proposed on how to increase the quality of oil palm ripeness grading in the Real Time Fresh Fruit Bunch (FFB) Oil Palm Grading System. This technique improvised the existing prototype grading system into a higher level “towards commercialization” grading machine. The improved grading machine changed the hardware design and system. Previously, the grading system used two software platforms which were MATLAB and LabVIEW and this was time consuming problem. This problem is due to the size of the image captured which is 1Gb per image. Therefore, the algorithm was migrated to a standalone software using LabVIEW. In this improved implementation, correctly human graded samples of oil palm bunches were image captured and analyzed for two categories. As a result, two sets of low resolved intensity images are captured by the Charged Coupled Device (CCD) camera. The grading system involves the hardware component which is the CCD camera and the software algorithm that is, the LabVIEW software for imaging purposes. The image analysis uses Artificial Neural Network (ANN) technique which includes training and testing of data. Model for the ANN is created based on the training data which is stored in the software memory. The ANN model is then used in the testing process where the software decides the grade of the oil palm fruit bunch. A significant improvement in the design specifications is made between the prototype and the new grading machine, which include weight measurement, sorting process, grip belting and feeder system. In the new machine, the speed for grading 60 bunches per minute is obtained compared to the existing system which is 10 bunches per minute. The design specification shows that the machine completes this process in one minute 33 seconds.

Preview Text FS 2016 75 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item