Hybrid Feature Extraction for Accurate Weather Image Classification

Accurate weather condition recognition from images is crucial for applications like autonomous driving, intelligent transportation, and environmental monitoring. This paper presents a hybrid weather image classification method, integrating ResNet50-based deep feature extraction, followed by Principal Component Analysis (PCA) for feature extraction, and Support Vector Machine (SVM) classification. The approach tackles complex visual patterns and limited annotated data through transfer learning and data augmentation. Experiments on a public Kaggle dataset show high classification accuracy across multiple weather categories, demonstrating its effectiveness and generalizability.