Image analysis for detecting insect fragments in semolina |
| |
Authors: | K Bhuvaneswari Noel DG White Chandra B Singh |
| |
Affiliation: | a Post Harvest Technology Centre, Tamil Nadu Agricultural University, Coimbatore, India b 195 Dafoe Rd. Cereal Research Centre, Agriculture and Agri-Food Canada, Winnipeg, MB, R3T 2M9, Canada c Canadian International Grains Institute, Winnipeg, MB, Canada d Biosystems Engineering, University of Manitoba, Winnipeg, MB, Canada |
| |
Abstract: | Semolina is used for the manufacture of pasta (long goods and short goods) and couscous and any contrasting colored specks adversely affect the appearance of the finished product. The specks result from wheat bran, diseased wheat, ergot or weed seeds. However, there is also the possibility that insect fragments will appear as specks. Specks are currently mostly determined by a manual process or by a speck counter in milling units. We compared the speck counts from an electronic speck counter (SPX Maztech Micrco Vision), acid hydrolysis and flotation (AOAC method 993.26), and near-infrared (NIR) hyperspectral imaging in semolina seeded with insect fragments (50-300 fragments/50 g) of Tribolium castaneum (Coleoptera: Tenebrionidae). There was a significant positive correlation between the number of insect fragments added and detected by all three methods. These results underline the importance of controlling insects in flour mills producing semolina, and also in plants producing pasta and couscous, to reduce speck counts in the finished products. |
| |
Keywords: | NIR hyperspectral imaging Speck counting Insect fragments Pasta manufacture |
本文献已被 ScienceDirect 等数据库收录! |
|