Treffer: Application of Artificial Neural Networks for Classifying Earthworms (Eudrilus eugeniae) Moisture Content During the Drying Process.

Earthworms (Eudrilus eugeniae) have many benefits for the health and animal feed industries. The drying process of earthworms is necessary to extend their shelf life, yet conventional gravimetric moisture tests are slow and destructive. The purpose of this study was to classify the moisture content of earthworms using machine vision and artificial neural networks (ANN) during the drying process, with classified worms into wet (> 40% wb), semi-dry (40%--12%), and dry (< 12%) states. RGB images (n = 450) were acquired every 15 min during cabinet drying at 60 °C; reference moisture was obtained gravimetrically. Nine color and texture features were extracted and ranked in WEKA; then, the top eight features were retained. An external feed-forward ANN implemented in MATLAB with 8-40-3 architecture, TrainLM optimiser, logsig--logsig--purelin transfer functions yielded MSE = 0.0733 (training) and 0.086058 (validation) and R = 0.95309 (training) and 0.92962 (validation). The modest MSE gap reflects class imbalance rather than overfitting, as classification metrics on the unseen test set match the validation results. [ABSTRACT FROM AUTHOR]

Copyright of Pertanika Journal of Science & Technology is the property of Universiti Putra Malaysia and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)