Wheel tread defects, such as flat spots, shelling, and thermal cracks, are critical concerns in railway safety and maintenance. Traditional inspection methods are often labor-intensive, time-consuming, and susceptible to human error. This paper proposes an automated defect detection system leveraging the YOLOv8 (You Only Look Once) algorithm combined with advanced image processing techniques. The system captures high-resolution images of wheel treads and applies preprocessing steps, including noise reduction and contrast enhancement, to improve defect visibility. YOLOv8 model is trained on a comprehensive dataset of annotated wheel tread images to accurately detect and classify defects such as flat spots, wheel shelling, and thermal cracks. The model's real-time detection capabilities and high precision make it suitable for integration into railway maintenance workflows for condition-based wheel maintenance. Experimental evaluations achieved high precision (97%), F1-score (93%), and mAP@0.5 (93%) with AdamW optimizer as the highest value among all the optimizers used, showing the effectiveness of the proposed approach in real-world railway inspection. The proposed approach significantly enhances inspection efficiency, reduces maintenance costs, and improves railway safety by enabling early detection of critical defects. This paper contributes to the growing body of work on intelligent transportation systems and highlights the potential of deep learning in automating critical infrastructure inspections.