G. Carlomagno, L. Capozzo, G. Attolico, and A. Distante (Italy)
Intelligent Data Systems and Computing, Signal and Image Processing
The paper describes the design of a low-cost workstation, based on near-infrared (NIR) spectrometry, for non destructively grading the ripeness of fruits. The system enables non-destructive measures that can be repeated several times and be applied to the whole production instead that on the few samples normally selected for destructive analysis. Destructive methods are normally applied for estimating the maturity in terms of internal sugar content and firmness. Four different set-ups have been designed and realized to estimate these parameters by measuring the NIR radiation transmitted through the fruit. They differ for some hardware components but share the general architecture and the algorithms used for processing the acquired signals. The data provided by the system are pre-processed using a noise-reducing method based on a packets-wavelet transform. In addition, an outlier detection schema has been used for identifying irregular behaviors inside each of the classes that need to be separated. A straightforward minimum distance classifier has been applied for assigning the data to their proper classes. The obtained results show that even this early version of the system allows the correct grading of peaches with a percentage of 82.5%.
Important Links:
Go Back
