"A Comprehensive Dataset of Motorcycle Odometer Images (Analog and Digital)"

"This dataset contains 5,860 high-resolution images of motorcycle odometers, specifically curated to support research in automated mileage validation and predictive maintenance systems. The data collection was conducted over an intensive eight-day period, from September 15 to September 22, 2025, capturing a diverse range of motorcycle models and conditions. The primary objective of this dataset is to provide a benchmark for Optical Character Recognition (OCR) and Object Detection models in the context of vehicle durability tracking. By extracting precise mileage data, researchers can develop algorithms that correlate distance traveled with part wear-and-tear cycles, facilitating automated service recommendations. Key features of the dataset include: Diversity of Displays: A balanced representation of both analog (mechanical) and digital (LCD\/TFT) odometer types. Standardized Resolution: All images are processed at a resolution of 320 x 160 pixels, optimized for deep learning input while maintaining character legibility. Real-world Variability: Images include various lighting conditions, angles, and dashboard styles to ensure model robustness against environmental noise. The dataset serves as a foundational resource for engineers and data scientists working on Intelligent Transportation Systems (ITS) and automated vehicle maintenance diagnostics."