Non-Invasive Detection of Early Germination in Beans Using Laser Speckle Temporal Analysis and Deep Learning | ||
| Journal of Agricultural Science and Technology | ||
| Articles in Press, Accepted Manuscript, Available Online from 16 September 2025 PDF (1.13 M) | ||
| Author | ||
| Alessandro Deana* | ||
| Uninove | ||
| Abstract | ||
| The visual quality control of bean grains is currently conducted manually and relies heavily on identifying surface-level defects to classify the product type. Among these defects, germinated grains pose a unique challenge: in their early stages, germination is not externally visible, requiring invasive methods such as physically breaking the seed for detection. This study proposes a non-invasive and automated approach that combines laser speckle imaging with deep learning to identify early-stage germinated defects in beans. Speckle images were captured under coherent laser illumination (λ = 633 nm) for samples subjected to germination periods of 0, 6, 12, and 24 hours. Two contrast analysis techniques were evaluated: Laser Speckle Spatial Contrast Analysis (LASCA) and the more advanced Laser Speckle Temporal Contrast Analysis (LASTCA), the latter using temporal intensity fluctuations across 120 video frames. From the resulting contrast maps, regions of interest (50×50 pixels) were extracted and used to train a Convolutional Neural Network (CNN) for binary classification. The proposed system achieved high performance, with an accuracy of 92.33% and sensitivity of 98.21%, successfully detecting germinated defects invisible to conventional inspection. By integrating temporal biospeckle analysis with deep learning, this method offers a scalable solution for intelligent, non-destructive grain inspection—addressing a critical gap in current computer vision systems and contributing to the advancement of Agribusiness 4.0. | ||
| Keywords | ||
| Laser Speckle image; Computer Vision; Convolutional Neural Network; Bean; Germinated | ||
| References | ||
|
| ||
|
Statistics Article View: 219 PDF Download: 120 |
||
| Number of Journals | 45 |
| Number of Issues | 2,171 |
| Number of Articles | 24,674 |
| Article View | 24,437,053 |
| PDF Download | 17,551,649 |