I just read this article comparing LiDAR and Visual Cameras for detecting lanes on highways. It sparked my interest, because we have successfully used digital cameras to extract lanes and pavement markings for many years.
The research conducted in the article was for automatic vehicle navigation. Unintended lane changes are a major cause of traffic accidents. These accidents are responsible for 50% of all fatalities on American roads. If we could equip cars with sensors that warn the driver when a lane changes occurs, we could save lives and reduce the number of accidents.
The same technology is very useful for roadway infrastructure inventories. Transmap’s ON-SIGHT™ mobile mapping van captures both LiDAR and HD digital images of the roadways. Both sensors are available and could be used to detect pavement markings and extract lanes. Which sensor will yield the best, most reliable results?
Which Method is Superior?
Researchers quoted in the article found that cameras performed better during day-light and during rain, while LiDAR had an advantage at dusk and when visibility deteriorated.
What does this mean for inventorying pavement markings?
Surveys to inventory pavement markings are mostly conducted during the day. If the visibility is good, the visual approach, using HD video cameras, is certainly appropriate. Digital images make it easy for our engineers to rate the condition of the pavement markings, measure their dimensions and clearly categorize each feature. Just like pavement condition surveys, a hybrid approach yields reliable and repeatable results.
A combination of visual images and LiDAR is beneficial when you try to automate the data extraction process. When visibility deteriorates, LiDAR enhances the results of visual imagery and improves the reliability of the data.
When Is the Best Time to Survey Pavement Markings?
It may be desirable to run pavement marking surveys at night. Imagine a busy city street with lots of cars obstructing the view of your cameras. Running the survey at night will give you a clear view of the street surface. However, there won’t be enough contrast to identify and measure the pavement markings.
This is where LiDAR comes in. LiDAR is an active sensor that measures signals reflected from the road surface. As pavement markings are made of retro-reflective material, the response is much stronger when LiDAR hits that surface, independent of sunlight.
Here it is appropriate to mention a third approach that combines infrared strobe lights with digital infrared cameras. The illumination by the strobe is used to increase the contract between the marker and the road surface. This technique enables the user to measure the retro-reflectivity of the pavement marking material, as well.