# PCI Calculation Methodology

Roadly’s PCI calculation methodology is based on the [ASTM D6433-20](https://www.astm.org/d6433-20.html) Standard Practice for Roads and Parking Lots Pavement Condition Index Surveys. We literally follow the estimation process from the standard, but replace the manual distress evaluation procedures with AI-based models.

PCI is calculated from different statistics of defects and distresses present in a Sample Unit of road. The index is affected by the number of defects detected, their severity level, and corresponding metric measurements (length, width, or area). We followed the [Distress identification manual for the Long-Term Pavement Performance Program](https://www.fhwa.dot.gov/publications/research/infrastructure/pavements/ltpp/13092/13092.pdf) to train our AI to detect different types of distresses and their severity level. Linear measurements are performed by our visual-based algorithm from input frames.

Roadly currently exhibits high accuracy in detecting and classifying five primary distress types: transverse, longitudinal, edge, alligator/fatigue, and block cracking, as well as potholes. These distresses are predominant contributors to the overall PCI score in most instances. Ongoing research and development efforts are focused on expanding Roadly's capabilities to encompass additional distress types, including patch deterioration, rutting, and shoving.

Please note that Roadly's algorithms are specifically designed for [asphalt concrete](https://en.wikipedia.org/wiki/Asphalt_concrete) pavement surfaces, commonly known as asphalt, blacktop or tarmac. Currently, other surface types are not supported, therefore PCI estimates based on data including non-asphalt surfaces may be inaccurate.

### Approach Comparison

<table><thead><tr><th width="176.33333333333334">Feature</th><th width="260">Expert-based PCI calculation</th><th>Roadly PCI calculation approach</th></tr></thead><tbody><tr><td><strong>General</strong></td><td></td><td></td></tr><tr><td>Method</td><td>Expert-based + special equipment (lidars, geodetic equipment)</td><td>Only video from a phone installed in vehicle</td></tr><tr><td>PCI calculation methodology</td><td>ASTM</td><td>ASTM</td></tr><tr><td>Reporting</td><td>ASTM</td><td>PCI + distress statistics per sample unit</td></tr><tr><td><strong>Sample units</strong></td><td></td><td></td></tr><tr><td>Road splitting into sample units</td><td>225 ± 90 square meters</td><td>Arbitrary, defined by the customer, may be 225 square meters</td></tr><tr><td>Sample unit area estimation</td><td>Based on direct width/length estimation of road</td><td>Visual estimated model + GIS data</td></tr><tr><td>Sample units sub-sampling (number of sample units estimated from total road area)</td><td><p>Sub-sampling</p><p>(All sample units in the section may be inspected but this is usually precluded for routine management purposes by available manpower, funds, and time. Total sampling, however, is desirable for project analysis to help estimate maintenance and repair quantities)</p></td><td>All</td></tr><tr><td><strong>Distress estimation</strong></td><td></td><td></td></tr><tr><td>Types of defects</td><td>Based on survey provider</td><td>5 types of cracks + Potholes, the rest under development</td></tr><tr><td>Measurements</td><td><p>Width: std error from 1mm</p><p>Length: std error from 1mm</p><p>Depth: std error from 1mm</p><p>Area: based on simplified contours (circles, rectangles)</p></td><td><p>Length: std error from 3cm</p><p>Width/Depth: not estimated directly, AI-based approach estimation by visual context</p><p>Area: std error from 9cm2</p><p> </p></td></tr><tr><td>Classification</td><td>Methodology based expertise</td><td>AI-based approach combined with distress topology modeling</td></tr><tr><td>Severity</td><td>Methodology based expertise</td><td>Methodology based models + AI</td></tr></tbody></table>