In this work, we present a Physics-Informed Neural Network (PINN) framework for the classification of oil palm fresh fruit bunch (FFB) ripeness using RGB images. Unlike conventional Convolutional Neural Networks (CNNs) that learn solely from visual patterns, the proposed PINN integrates a physics-based index—derived from the red-to-green pixel intensity ratio—directly into the network architecture and loss function. This hybrid design embeds wavelength-dependent physical knowledge related to chlorophyll degradation during ripening, enabling the model to learn more robust and generalizable features even with limited and imbalanced training data. The PINN model achieves a peak accuracy of 0.73, outperforming the purely data-driven CNN baseline (0.68) by a margin of 5%. Overall, the PINN demonstrates superior performance in minority-class detection and maintains stable convergence under three different lighting conditions (different light spectra). These results highlight the effectiveness of integrating domain-specific physical insights into deep learning models, offering a promising pathway toward reliable, non-destructive, and automated ripeness assessment for agricultural applications.
10 August 2023 in Advances in Agricultural and Food Research Journal
A severe disease known as Ganoderma Boninense, which attacks a group of wood-decaying fungi, causes Basal Stem Rot (BSR) in oil palm trees. One of the treatments introduced to control the spread is injecting a chemical into the palm's trunk. After receiving a hexaconazole treatment for a year, a study on 2008 Ganoderma-infected palm trees in an estate revealed that 95.6% of the infected palms were still standing. The injection process facilitates several devices that were used and invented. These machines utilise a pressure-injection technique that injects a chemical through a flexible hose into a healthy palm's tissue to prevent the disease from spreading. Hence, the development of effective trunk injector mechanisms for Ganoderma treatment is discussed in this article. The first injection method utilised the Pressure Injection Apparatus (PIA), which can deliver a chemical to the target sites to control BSR. Second, a Mechanical Trunk Injector mainly facilitated chemical injection into the palm trunk using a prime mover. Next, the Tractor Mounted Trunk Injector, comprised of an injection apparatus system, was found to perform effectively and efficiently. The distribution profile of the dye eosin in the stem revealed that it covered more than 80% of the cross-sectional and longitudinal axes of the palm tree's trunk. Instead of the previous injection apparatus control method, this machine could speed up the injection time. Compared to the 20-minute-per-hole completion time of the PIA, the machine only took 3 to 5 minutes to complete an injection, which was 65% faster, and about 100 to 120 palms could be covered in a day. Labour requirements were very minimal as it is a single-operated machine. Utilising the machine would also result in a reduction in the number of workers needed.
Oil palm fresh fruit bunch (FFB) ripeness classification is essential for improving the oil extraction rate, oil quality, and mill processing efficiency. However, RGB-based classification is often limited by insufficient colour information from dark unripe fruitlets and shadowed regions of the bunch surface. This study evaluated the effect of image input representation on convolutional neural network (CNN)-based FFB ripeness classification. Four models with the same CNN architecture were compared using RGB, RGB + infrared (IR), CIEL*a*b*, and CIEL*a*b* + IR inputs. RGB and IR images were acquired using a dual-camera setup positioned 3 m from the FFB sample. The CIEL*a*b* + IR model achieved the best overall performance, with weighted-average precision, recall, and F1-score values of 0.84, 0.81, and 0.81, respectively, compared with 0.79, 0.76, and 0.76 for RGB-only. The addition of IR improved model performance by providing complementary near-infrared reflectance information, while CIEL*a*b* colour-space transformation provided a more discriminative colour representation. Class activation heat maps showed that the models focused mainly on fruitlet regions, with the CIEL*a*b* + IR model producing more distinct activation over ripeness-relevant areas. These findings demonstrate that the proposed four-channel CIEL*a*b* + IR image representation improves CNN-based FFB ripeness classification by combining lightness-separated colour information with near-infrared reflectance features, although underripe FFB remains challenging, where it produced the highest rate of false positives because of its transitional and heterogeneous characteristics.