Improvement of photogrammetric JRC data distributions based on parabolic error models

• Practical guidelines for photogrammetric surveys for JRC estimation.
• Performance of lab tests for JRC estimations focusing on the influences of lens focal length and the photograph distances.
• Suggest the allowable photographic distances for JRC values according to the employed focal length of lenses.
• Suggest parabolic functions to predict the errors of photogrammetric JRC values.

This research describes the influences of the focal length of lenses as well as camera-to-object distances that control the accuracy and precision of 3D photogrammetry models used for assessing joint roughness coefficients (JRC). A parabolic error model, which is developed by a series of photogrammetric laboratory tests, is used to improve the JRC data distributions obtained from the photogrammetry 3D models. The influences of camera-to-object distances and focal lengths on the accuracy of JRC values are investigated in laboratory conditions. A series of photogrammetry tests are performed under different conditions by varying the camera-to-object distances and the camera lens focal lengths. The camera positions are controlled on a sliding device to increase the accuracy of geo-referencing. The results suggest the allowable photographic distances for the employed photogrammetry equipment (e.g. camera and focal lengths of lenses) to estimate JRC values through comparison with the manually measured values. The root-mean-square-error (RMSE) of JRC functions, which are correlated with the camera distances and the normalized JRC values, are depicted as different size parabolic curves according to the employed focal lengths. These models successfully improve on the accuracy of the originally obtained JRC distributions.