Finetuned Deep Learning Models for Fuel Classification: A Transfer Learning-Based Approach

Accurately classifying petrol and diesel fuel using an image processing method is crucial for fuel-related industries such as petrol pumps, refineries, and fuel storage facilities. However, distinguishing between these fuels using traditional methods can be challenging due to their similar visual characteristics. This study aims to enhance the accuracy and robustness of existing fuel classification by utilizing the transfer learning-based finetuned pre-trained deep learning models and ensemble approaches. Specifically, we upgrade pre-trained deep models like ResNet152V2, InceptionResNetV2, and EfficientNetB7 by incorporating additional layers. Through transfer learning, these models are adapted to the specific task of classifying petrol and diesel fuels. To evaluate their performance, the upgraded deep model and an ensemble of these models are tested on a synthetic dataset. The results indicate that the ensemble of upgraded ResNet152V2, InceptionResNetV2, and EfficientNetB7 achieves recall, precision, f-score, and accuracy scores of 99.54%, 99.69%, 99.62%, and 99.67%, respectively. Moreover, a comparative analysis reveals that the upgraded models outperform state-of-the-art baseline models.