Atmospheric Noise-Resilient Image Classification in a Real-World Scenario: Using Hybrid CNN and Pin-GTSVM
This addresses the challenge of reliable parking space detection in real-world hazy environments for smart parking infrastructure, representing an incremental improvement over existing methods by focusing on atmospheric noise resilience.
The paper tackled the problem of parking space occupation detection in hazy conditions, where existing deep learning methods perform poorly, by proposing a hybrid model with a pre-trained feature extractor and Pin-GTSVM classifier that eliminates the need for dehazing systems and is insensitive to atmospheric noise, achieving significant accuracy improvements on hazy parking datasets.
Parking space occupation detection using deep learning frameworks has seen significant advancements over the past few years. While these approaches effectively detect partial obstructions and adapt to varying lighting conditions, their performance significantly diminishes when haze is present. This paper proposes a novel hybrid model with a pre-trained feature extractor and a Pinball Generalized Twin Support Vector Machine (Pin-GTSVM) classifier, which removes the need for a dehazing system from the current State-of-The-Art hazy parking slot classification systems and is also insensitive to any atmospheric noise. The proposed system can seamlessly integrate with conventional smart parking infrastructures, leveraging a minimal number of cameras to monitor and manage hundreds of parking spaces efficiently. Its effectiveness has been evaluated against established parking space detection methods using the CNRPark Patches, PKLot, and a custom dataset specific to hazy parking scenarios. Furthermore, empirical results indicate a significant improvement in accuracy on a hazy parking system, thus emphasizing efficient atmospheric noise handling.