PENGUNAAN PENCITRAAN MULTISPEKTRAL PADA PANJANG GELOMBANG 520 NM DAN 800 NM UNTUK MENGEVALUASI TINGKAT KEMATANGAN TBS KELAPA SAWIT

Abstract

Oil palm fresh fruit brunch (FFB) are the main source of crude palm oil (CPO). Sorting and grading FFB are important in order to obtain high quality CPO. Multispecral imaging has been purposed to be implemented in high speed sorting machines due to less wavelength bandwidths used hence less processing time. This study was aimed to evaluate the ripeness levels of oil palm FFB based on relative reflectance intensity and fruit firmness. Multispectral images were acquired using two bandpass filters mounted in a filter wheel with wavelengths of 520 nm and 800 nm respectively. The image acquisition and processing were controlled using python based program. The samples consisted of 30 oil palm FFBs of Tenera varieties with three ripeness levels as unripe, ripe, and overripe. The result showed that the relatif reflectance intensity at wavelength of 520 nm is inversely proportional to the maturity level, on the other hand,  relatif reflectance intensity at wavelength of 800 nm is directly proportional to the maturity level. The relation between the firmness and ripeness level are inversely proportional. Relative reflectance intensity of the multispectral images at the wavelength of 800 nm had a better correlation to the palm fruit firmness than the image at the wavelength of 520 nm with the correlation coefficient (r) of -0.0198 at 520 nm and -0.8594 at 800 nm. it can be shown that the multispectral imaging is potensial to be implemented for FFB ripeness evaluation.