Objective With the rapid development of urban rail transit, wheelsets as the critical components of vehicle bogie system, their dimension condition directly impacts train operational safety and stability. Traditional manual static detection methods suffer from high labor intensity, low efficiency, susceptibility to missed inspections, and significant subjective errors, failing to meet growing maintenance demands. To overcome the inherent defects of manual contact measurement, achieve efficient, precise, and automated online detection of wheelset dimensions, thereby ensuring operational safety and reducing O&M costs, an online wheelset detection method based on structured light image recognition is proposed.
Method A non-contact, online measurement technology is adopted, utilizing 8 sets of area-array CCD (charge-coupled device) cameras in collaboration with linear lasers. Based on the light section method, the deformation of wheel flange profile is projected with laser. Combined with precise calibration (via checkerboard fixture and perspective transformation), the multi-camera coordinate system is unified. Image processing algorithms are utilized to extract light sections while filtering noise. Finally, key flange dimensions are automatically calculated using triangulation and chord length measurement methods, eliminating manual errors. Taking a straight section of metro line entrance/exit depot in Shenzhen as an example, an onsite application test is conducted for a Type A train.
Result & Conclusion The proposed method is able to accurately measure all parameters with measurement errors meeting the technical specifications; it not only enhances dimension measurement accuracy, but also improves maintenance efficiency significantly through innovative dimension detection method, calibration and algorithm methods.
DownLoad: