Within any given type of lidar sensor, the most important characteristic for people-measurement is the density of the point-cloud. You’ll often see lidar devices classified by their beam count, and the reason is simple, the more beams a device has the denser the point-cloud it produces. The denser the point-cloud, the more detailed the object definition can be and the more likely it will be that the device can distinguish people in a crowd. The more crowded and complex the space, the more important beam count becomes.

In short, the density of the point-cloud is the single most important metric for gauging the quality of people-measurement you’re likely to get. But while the density of the point-cloud is typically a function of the number of beams, this simple rubric hides quite a bit of underlying complexity. It’s also about beam pattern, mounting capabilities, and scan rates.

For people measurement, you’ll usually mount the devices on ceilings to minimize the amount of blockage from crowds. That means any beams going up from the device instead of down are just wasted. You’re just throwing useless beams up on the ceiling. That’s great for counting spiders…not so much for counting people.

A fair number of puck lidar devices waste a good chunk of their beams this way – meaning you don’t get the benefit of the beams you have. This is where mounting capabilities can be important too. Sometimes you can get more effective scene coverage if you can invert a puck lidar. But not every puck lidar can be inverted – meaning you don’t have the option of flipping the beam pattern.

And, as I pointed out in my last post, the shape of the beam pattern also has a profound influence on the point-cloud you get.

Finally, some devices (both mechanical and solid-state) let you adjust the beam pattern they use. Options may include things like trading off frame rate for point density (almost always a good trade-off for our kind of people measurement),and controlling the density of the beam pattern across the field to, in essence, focus in on specific areas. This can dramatically improve the capabilities of the sensor compared to its raw specifications.

So, when you calculate point-cloud density over coverage area to compare lidars or design an implementation, make sure you use the actual, effective coverage area not a hypothetical coverage area that assumes every beam is created equal.

Which probably leaves you thinking that figuring this all out is non-trivial and that designing a lidar implementation may not be a slam dunk despite lidar’s great coverage and the considerable flexibility you get with lidar placement.

That’s true. We see a lot of people buying and installing lidars that are sub-optimal for their environment or mounted in ways that dramatically worsen their people-measurement performance. And yes…sigh…we’ve made these mistakes ourselves as we learned.

There is a problem, fortunately, where software can help. Most lidar manufacturers have simulation software that will model out coverage of their sensors in an area. These tools range from laughably primitive to pretty darn sophisticated. Third-party perception providers (like Outsight and Seoul Robotics) have simulation tools that allow you to test and compare different lidars (though only the ones they support) in a simulation:

These tools won’t figure out the optimal placement on their own and they may not handle all the customizations a sensor can provide, but they will give you a pretty good visual sense of what kind of coverage a sensor will give you out of the box given a specific mounting point and angle. This really does help a lot.

All of which should make it plain that while there is no one best lidar sensor, there is a clear path to figuring out the best sensor for your environment and people-measurement needs.

As a people-measurement analytics provider, we don’t have a vested interest in what sensors you choose (lidar or otherwise) or what type of lidar you choose. But we do have a vested interest in making sure you get the best measurement possible for your locations and spend!