| 摘要 |
PURPOSE:To divide a rail contact face of a wheel into a plurality of areas, compare a plurality of photographed images with each other and automatically detect a flat produced on the contact face. CONSTITUTION:A wheel 1 passes a plurality of specified positions on a rail, for instance, a wheel position detection means 5 transmits a signal to a photographing means 4 such as a CCD camera, photographs a plurality of pictures of a contact face of a wheel 1 and transmits them to an image storing means 6. Each conversion coefficient for correcting change of a distance and an angle of a view at the time when the whole circumference of the rail contact face is photographed, is calculated by a deformation quantity calculation means 7 and outputted to an image conversion means 8. The means 8 corrects a coordinate system of images Gr taken in for the second time by means of each conversion coefficient, and image data H[Gr] of the same space position relation as the image data of the images Go taken in at the first time are transmitted to an image comparison means 9. The means 9 finds a difference between it and the image data of the images Go and detects specified images or more as a flat 3. A flat size measurement means 10 calculates an area, vertical width and horizontal width of the flat 3 therefrom. |
