Assessment of Long-Term Pavement Performance Plan Wall Projection-Based Distress Data Variability

Abstract

In the Long-Term Pavement Performance (LTPP) program, 35-mm, black and white, continuous-strip photographs are used as a permanent record of pavement distress development for archival purposes and to quantify the distress severity and extent for pavement performance analysis. The traditional method of interpreting distress from LTPP film utilizes a relatively small image projected onto a digitizing tablet. From quality control checks performed on the interpreted data, it was found that some low severity types of distress, identified from larger magnified images projected onto a wall or projection screen, could not be seen in the smaller image used for distress interpretation. The variability in distresses interpreted directly off of the large format, wall-image projection was assessed through analysis of interpretations performed on six asphalt concrete and six portland cement concrete pavement sections used in the LTPP distress rater accreditation workshops. The data set included distress ratings from eight individuals, four two-person rater teams, and an experienced rater team. Also available were distress ratings performed in the field by the experienced rater team, which are used as reference values which represent the best estimate of ground-truth. Statistical tests show that the film-interpreted distresses from individual raters exhibit much larger variability than those from the rating teams. The most significant contributor to this finding is outlier observations in which one of the individual raters had significantly different ratings than the rest of the group. The spread in the rating teams was much lower. The film interpreted distresses from the experienced group correlated very well with the field-derived reference values.