Interpreting coverage values in reports
Several factors affect the quality and validity of recorded video material. In addition, there may be multiple videos recorded for the same road section. When you create a report, RoadAI uses coverage values to show how well a road section is covered and whether the coverage meets the required quality.
Coverage values are available in road segment interval and group reports. In these reports, coverage is shown for each road section. In an interval report, coverage is reported for each interval. In a road segment group report, coverage is reported for each group or segment as defined in the road network geometry.
There are three coverage values that together describe how well a road section is covered:
- Coverage (total): The percentage of the road section recorded in any of the videos used for the report.
- Coverage (valid): The percentage of the road section that contains data produced from video material matching the validity criteria and filters selected for the report.
- Coverage (invalid): The percentage of the road section recorded only in video material that does not match the selected validity filters. This data is excluded from the report.
 | Coverage values are calculated using 5-meter raw data sections. |
The table shows how coverage values relate to the absolute distances covered for a road section.
| Example | Road section length | Coverage (total/valid/invalid) | Coverage distance (total/valid/invalid) |
|---|---|---|---|
| Road segment interval report, 20 m | 20 m | 100%/25%/75% | 20 m/5 m/15 m |
| Road segment interval report, 100 m | 100 m | 60%/55%/5% | 60 m/55 m/5 m |
| Road segment group report, grouped by street | 2400 m | 90%/60%/30% | 2160 m/1440 m/720 m |
In the first example, the whole 20–meter interval has been surveyed. However, 75% of the data is based on video that does not meet the selected validity filters and is excluded from the report. Only 25% of the data determines the reported values for the interval. If this high level of invalid coverage is not a common phenomenon in the report, and the interval is short, having only 5 meters represent a 20–meter interval may be acceptable, depending on your needs.
In the second example, the 100–meter interval, only 60% of the interval has been surveyed. The valid coverage is 55%, which is most of the surveyed material. However, because only 60% of the interval is covered, the valid coverage is relatively low.
Based on the third example, we can notice the pronounced impact of coverage percentages in case of longer road sections, which are typical in road segment group reports. A valid coverage of 60% translates to 1440 meters. The remaining 960 meters includes both sections with no data (10% or 240 meters) and sections with invalid data (30% or 720 meters).
To ensure that the valid coverage provides a sufficient sample for decision making, especially when considering the average road section length in a report, it is recommended to filter road sections by coverage. For example, in an Excel report, you can set a filter for the Coverage (valid) column. By setting a minimum acceptable amount of valid coverage, you can exclude road sections that do not provide a reliable sample for decision making.
| Road sections with only invalid coverage, that is, with no valid coverage, are excluded from the report. |