| The potential of non-destructive spectroscopic technique for the detection of potato internal defect (Hallow Heart) using artificial neural networks |
| Paper ID : 1126-NICAME1402 |
| Authors: |
|
Kamal Imanian * Agricultural Engineering Research Department, West Azarbaijan Agricultural and Natural Resources Research and Education Center, AREEO |
| Abstract: |
| Potato is a widely consumed product all over the world, and the detection of internal defects in the potato tuber is very important for quality control and reduction of post-harvest losses. The present research has been done in order to detect apparently healthy potatoes but with internal defects ((Hallow Heart) using non-destructive spectroscopic method. In this regard, a VIS-NIR-SWIR spectrometer was used. With the help of hybrid neural network-cultural algorithm, effective wavelengths were determined in two spectral regions: VIS-NIR (1100-350 nm) and SWIR (2500-1100 nm). The effective wavelengths in the VIS-NIR region included 998, 883, 861, in the SWIR region included 1858, 1539, and 1896 nm. An ANN model was developed based on the spectral data obtained from healthy potato tubers and He had been trained. The research results show the ability of spectroscopy combined with artificial neural network (ANN) as a non-destructive and efficient method to detect pitting defects in potato tubers, which enables online grading and minimizing economic losses in the potato industry. |
| Keywords: |
| : Potato, spectroscopy, internal defect, Hallow Heart, non-destructive |
| Status : Paper Accepted (Poster Presentation) |
