Real-Time Visual Monitoring and Analysis of Fuel Cell Heavy-Duty Truck Power Systems Using Deep Learning-Enhanced Image Processing Algorithms

Abstract

Driven by global energy transition initiatives and the "dual carbon" goals, fuel cell heavy-duty trucks have emerged as a pivotal solution for the green transformation of commercial vehicles, offering advantages such as zero emissions and high energy density. However, their power systems are complex and highly susceptible to environmental and load variations, making real-time visual monitoring essential for ensuring operational safety and energy efficiency. Existing approaches largely rely on traditional sensor-based data methods or hand-crafted image processing techniques, which suffer from limitations such as high dependency on sensor precision, poor robustness in complex environments, low feature extraction efficiency, and high manual annotation costs. These limitations hinder the effectiveness of real-time fault or anomaly detection under diverse operating conditions. This study focuses on real-time visual monitoring of fuel cell heavy-duty truck power systems. It begins by clearly defining the fault and anomaly detection problem, including fault types, features, and detection objectives. Subsequently, it proposes a deep learning-enhanced image processing algorithm that leverages the ability of deep learning to automatically extract high-level image features, thereby building a robust real-time detection model suited for complex scenarios. The proposed approach aims to overcome the limitations of traditional methods in feature representation and generalization capability. The results of this research can provide technical support for the safe maintenance and performance optimization of fuel cell heavy-duty trucks, and promote the broader application of deep learning in the field of new energy vehicles.