Year 2026, Volume 72 Issue 1
| Year : | 2026 |
| Volume : | 72 |
| Issue : | 1 |
| Authors : | Sonay DUMAN, Abdullah ELEWİ, Erdinç AVAROĞLU |
| Title : | A COMPREHENSIVE MULTISENSORY LEARNING APPROACH FOR OYSTER MUSHROOM MATURITY DETECTION IN GREENHOUSES |
| Abstract : | In the mushroom industry, analysing growth and detecting maturity during cultivation is crucial for smart mushroom farming. This study introduces a comprehensive multisensory approach employing diverse machine learning algorithms to automate the classification of oyster mushroom (Pleurotus ostreatus) maturity classes in a greenhouse setting. Three unique datasets were created, comprising manually extracted features of oyster mushroom node sizes (height and width) alongside size features derived from bounding boxes automatically identified by trained Detectron2 and YOLOv8 object detectors utilising annotated mushroom images. The manual and automatic datasets were used to train classification algorithms separately, their performance was evaluated and compared. The impact of environmental setting within the mushroom greenhouse on mushroom growth was examined by integrating environmental variables as features for classification algorithms. The results indicated a significant enhancement in classification accuracy, especially incorporating the temperature and humidity values. The findings indicate that the HistGBDT classifier achieved the highest testing accuracies of 91.2%, 90.1%, and 92.7% on the automatic YOLOv8, Detectron2, and manual datasets, respectively, employing extracted mushroom node’s width and height, along with temperature and humidity features in the greenhouse. The results of other classifiers demonstrated minor differences between manual and automatic datasets, suggesting that the YOLOv8 object detector generated mushroom bounding boxes significantly more aligned with those drawn by humans compared to the Detectron2 framework. These results demonstrate an invaluable multisensory approach for detecting mushroom maturity classes in greenhouse settings, surpassing existing camera-based systems. |
| For citation : | Duman, S., Elewi, A., Avaroglu, E. (2026). A comprehensive multisensory learning approach for oyster mushroom maturity detection in greenhouses. Agriculture and Forestry, 72 (1): 203-220. https://doi:10.17707/AgricultForest.72.1.12 |
| Keywords : | Oyster mushroom, YOLOv8, Detectron2, agriculture, classification, object detection |
| download paper | |