Skip to content
TrackPodcasts
historyMar 17, 202616:17

How stereoscopic cameras calculate vehicle speed

pplpod

About this episode

How stereoscopic cameras calculate vehicle speed

Get every episode summarized

Each time pplpod publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.

Email me new episodes

Free for 3 shows. No card needed.

Hosts & guests

Transcript ready

389 searchable segments. Every word is indexed and playable.

How stereoscopic cameras calculate vehicle speed

pplpod

0:00
16:17

Full transcript

pplpodHow stereoscopic cameras calculate vehicle speed. Machine-transcribed; use the interactive transcript above to jump the player to any line.

You're listening to a podcast right now, driving, working out, walking the dog. If you're into podcasts, chances are you have something to say too. With RSS.com, starting your own is free and easy. Upload an episode, and we distribute it to Apple podcasts, Spotify, Amazon Music, and hundreds more. Track your listeners, see where they're from, and start earning from ads like this. Even with just 10 listeners a month. If you've been thinking about starting a podcast, this is your sign. Start free at RSS.com. Picture the camera mounted on like the highway gantry you drove under this morning. Right. Or maybe it was one of those sleek little gray boxes perched on a stuck metal pole by the interstate. Yeah, exactly. And you probably tapped your brakes, assuming it was actively shooting a radar beam at your car to clock your speed. Which is what we all do, right? For sure. Yeah. But what if it wasn't emitting anything at all? What if it was just watching you using pure geometry to calculate your velocity? That's a pretty wild thought.

It is. Welcome to today's deep dive. We have an incredibly revealing excerpt from a Wikipedia article detailing a technology called video detection and ranging or a Vidar. Vidar, yeah. And our mission today is to decode how modern speed enforcement has quietly shifted from bouncing active signals off your bumper to passively calculating your speed using advanced stereoscopic math. Okay. Let's unpack this. Well, to really appreciate the sheer leap in engineering that Vidar represents, we kind of have to look at what the source material identifies as the old guard. Right. The traditional methods it's replacing. Exactly. The text specifically points to two foundational technologies. You have radar, which relies on Doppler shifts and Liar, which calculate speed using the time of flight principle. Okay. We all know the basics of those old school systems, but let's translate that for the listener really quick. Yeah. Good idea. So radar is reading the Doppler shift, which is basically measuring how compressed or

radio wave gets when it bounces off the grill of a car moving toward the sensor. Right. The frequency change. Yeah. It's like driving a boat into oncoming waves. The faster you drive, the faster the waves hit the bow. So the radar gun measures that wave compression to find your speed. And then there is Liar, which abandons the Doppler effect entirely because it uses light instead of radio waves, right? Exactly. As the text outlines, Liar relies on time of flight. So think of time of flight like bouncing a tennis ball against a brick wall in the dark. Oh, I like that analogy. You throw it and you count the exact milliseconds it takes to return to your hand to figure out how far away the wall is. Just way faster. Right. An incredibly fast game of Marco Polo played with light instead of sound. The police scanner yells Marco by firing a laser pulse millions of times a second and it tracks the exact microsecond your bumper yells Polo back by reflecting that light. And since we know the speed of light is a constant, measuring the round trip time of that laser pulse gives you the exact distance to the car.

And you just do that thousands of times a second and boom, you have the car speed. Exactly. But what's fascinating here is the underlying operational philosophy of both radar and light are what do you mean by philosophy? Well, despite using totally different physics, you know, compression of radio frequencies versus the speed of light, they share one defining characteristic, which is they're both active systems, meaning they have to like throw energy into the environment to get a result. Yes. They're completely reliant on a mission to measure the physical world. They have to physically interact with like a light our system requires a laser diode to generate intense pulses of light optics to focus that beam and often some sort of mechanical or solid state scanning mechanism to sweep that beam across the highway. That sounds complicated. It is. And radar requires an oscillator to generate microwaves and antenna to broadcast them and a receiver sensitive enough to catch the scattered reflection. Which means they require significant power.

Yes. A lot of power. And they have moving parts or complex broadcasting hardware that can degrade over time. Plus, they're constantly shouting into the electromagnetic spectrum. Exactly. And that active emission is exactly the paradigm our source material says is ending. Wow. Really? Yeah. We're moving from a regime of active interrogation, like pinging vehicles with energy, to a regime of entirely passive observation. Which brings us to VDAR. Right. We are entering the era of VDAR. So the text defines video detection raining as a technique that measures the speed of a distant vehicle simply by tracking the object through vision cameras. It's watching it. No lasers, no radio waves, just capturing the ambient light that is already bouncing off your car. It is an incredibly elegant concept. You strip away the emitters, the diodes, the antennas, and you just replace them with the lens and a sensor. I mean, I'm hung up on the mechanics here, though. Also. Well, a camera captures flat two-dimensional images, right? Right. And a car is moving toward the lens in three-dimensional space at like 75 miles an hour.

Yeah, fast. So how does a flat grid of pixels calculate velocity without a radar wave telling it the physical distance? I mean, a single video is just a rapid sequence of flat photographs. That exact physical limitation is what kept passive video speed detection in the realm of theory for decades. Oh, so it was a known problem. Absolutely. A single camera lens, no matter how high the resolution, is inherently psycholopic. It has zero native depth perception. Right, because a large car far away looks geometrically identical to a small car close up. Exactly. But if you look closely at the source text, it specifies that this high precision measurement isn't done with the standard camera. Okay. What is it done with? It is achieved through advanced stereoscopic imaging techniques. Stereostopic imaging. So we aren't talking about one lens. No. We are talking about two. Yeah. We are talking about replicating human biology. It is the exact same biological mechanism that allows you to catch a baseball or navigate a crowded sidewalk.

Let me sketch this out mentally for the listener, because the biological parallel makes the math so much easier to grasp. Yeah. Go for it. So hold your thumb out at arm's length. Okay. Do it. Close your left eye and align your thumb with something on the wall behind it. Say a picture frame. Got it. Now, without moving your hand, open your left eye and close your right eye. Your thumb appears to jump horizontally across the background. Exactly. That visual jump is called parallax. Your left eye and your right eye are viewing the world from two slightly different angles, separated by a couple of inches. Your brain takes those two flat 2D images, measures the difference of the disparity between where objects appear in each eye, and instantly stitches them together to create a 3D map of the world. It's pretty amazing. That's how you know how far away your sum is from the wall. The Vidar applies that exact phenomenon to traffic enforcement. But instead of neurons, it uses pure geometry and silicon. So dual visual inputs. Yes. The SISAB utilizes two camera lenses mounted on that gantry or pole, staring at the exact

same stretch of asphalt. But wait, for the computer to calculate depth from that parallax, it has to have a baseline, right? I mean, the human brain knows how far apart our eyes are. Exactly. Our system established that metric. The baseline is actually the most critical variable in the entire system. The two lenses on a Vidar unit are rigidly mounted at an exact, mathematically known distance from each other. So they don't move? Never. Let's say it's exactly 24 inches. The system software is hard-coded with that 24 inch baseline. Gotcha. When a car drives into the frame, camera A captures an image of the license plate. In that flat image, the license plate is located at specific pixel coordinates. They say pixel 1000 on the x-axis. And at that exact same millisecond camera B captures its own image. Yes. But because camera B is 24 inches to the right, it sees the license plate from a slightly different angle. Precisely. So in camera B's flat image, the license plate isn't at pixel 1000. It's at like pixel 1050.

And that 50 pixel difference is the disparity. Ah, the visual jump. Exactly. Software knows the exact distance between the two lenses, the exact focal length of the optics, and the exact pixel disparity between the two flat images. It can use basic trigonometry to triangulate the exact distance to that license plate in three-dimensional space. Wow. It has mapped the z-axis purely through observation. Okay. So that establishes the depth for one single moment in time. Right. But speed is distance over time. So the Vidar system has to run that complex stereoscopic triangulation over and over again. Frame by frame by frame. That's a lot of max. It is. This is where the video aspect of video detection and ranging comes into play, because video is really just sequential photography. Sure. Let's assume the Vidar cameras are shooting at 60 frames per second. Okay. If it's shooting 60 frames a second, that means there is a gap of about, what, 16.6 milliseconds between every single photograph. Exactly. So the car enters the frame. The system uses stereoscopic disparity to calculate that the bumper is exactly 100 feet

away. Right. And in seconds later, the cameras capture the next frame. The system recalculates the disparity. The bumper is now, say, 98 feet away. So it knows the exact distance the car moved through 3D space and it knows the exact fraction of a second it took to cover that distance. And distance divided by time equals velocity. That is wild. The system calculates your speed without ever emitting a single pulse of energy. It just watched the light naturally bouncing off your car from two different angles and let the software do the heavy lifting. Which brings up a massive historical question for me. The geometry behind stereoscopic triangulation isn't new. I mean, humans have understood parallax and trigonometry for centuries. Oh, absolutely. So if this passive method is so elegant, if it saves us from building complex radar emitters and expensive spinning light-ar lasers, why wasn't this the standard from the very beginning of traffic enforcement? That is a great question. And the answer is, the limitation was never the math. What was it? The limitation was the silicon.

The source material actually points us directly to the moment this barrier was broken. Oh, right. The patent. Yes. It cites a specific architectural origin for this technology. US patent 8184863B2. Titled the video speed detection system. Issue to an inventor named Jigang Wang on May 22, 2012. 2012? I mean, that is incredibly recent when you think about the history of highway infrastructure. Very recent. And the view traffic cameras as these archaic, bureaucratic monoliths that have been around forever. But the specific patented capability to enforce speed limits using passive stereoscopic video is barely over a decade old. Yeah. And when you look at what the system actually has to do to replace a radar gun, that 2012 date makes perfect sense. Because of the processing power. Exactly. Think about the computational load we just described. Yeah. The system is pulling in two high-definition video feeds simultaneously. Right. To scan millions of pixels in real time, identify the vehicle moving in highway speeds, isolate

a specific feature like a license plate or a headlight. And find that exact same feature in the second video feed. Yes. Calculate the pixel disparity, triangulate the 3D depth, and then do it all again 16 milliseconds later. And it has to do all of that math before the car physically drives out of the camera's field of view. Right. And these are early 2000s, trying to process dual HD video feeds frame by frame for real time 3D reconstruction would have just melted a conventional roadside processor. You would have needed a server rack the size of a refrigerator sitting next to the highway. Pretty much. The hardware simply couldn't keep up with the geometry. Radar and LiDAR were the standard because computationally speaking, they are relatively simple. Because measuring the frequency shift of a bounced radio wave takes very little processing power. Exactly. But by 2012, Moore's law had done its work. High definition digital image sensors became incredibly cheap and highly accurate. Yeah, they were everywhere. And more importantly, microprocessors became fast enough to handle the sheer volume of data

required for real time disparity mapping. Spot on. Jigging Wang's patent represents the exact moment when the processing power caught up to the math. Wow. Transition where software finally overtook hardware in the realm of speed detection. And as the text outlines, that 2012 invention didn't languish in a patent office somewhere. No, not at all. The source explicitly lists its primary applications as remote sensing and traffic speed enforcement. So this passive stereoscopic observation has actively reshaped our physical environment. It really has. The text mentions the systems are deployed in gantries or on roadside poles. That's the physical footprint of this paradigm shift. It completely changes the architecture of enforcement. How so? Well, because you no longer need heavy high voltage equipment to emit radar waves or mechanical spinners for LiDAR, the footprint of a speed detection system shrinks drastically. Right. It doesn't need to be this massive rig. Exactly. It becomes just a pair of lenses in a small housing, quietly processing the ambient light.

So the next time you are driving down the interstate and you pass under one of those massive metal gantry stretching over the lanes or you catch a glimpse of a camera cluster on a roadside pole out of the corner of your eye, I want you to look at it through the lens of this deep dive. Yeah. It changes your perspective entirely. Don't just think of it as a dumb security camera waiting to take a flat flash photograph of your license plate. Because it's not. That gray box is doing high-level real-time calculus on the physical space around it. You are driving through the field of view of an incredibly sophisticated stereoscopic set of eyes. Literally watching you. It is silently taking in dual video feeds, mapping the pixel disparity of your vehicle against the background, reconstructing the three-dimensional reality of the highway and calculating your exact trajectory and velocity. It is translating the physical world into pure geometry and doing it all passively. It forces a complete reimagining of how we measure our environment. So what does this all mean?

We started with a source detailing a rather dry-sounding Wikipedia entry about video detection and ranging, but what we actually uncovered is a fundamental evolution in how technology interacts with reality. It really is a massive shift. We are witnessing the end of the brute force era, an era where we had to actively hurl radar ways and laser pulses out into the world and wait for them to bounce back just to understand how fast something was moving. Yeah, shouting into the dark. And we have transitioned into an era of elegant, passive, visual tracking where simply gathering ambient light through two lenses is enough to decode the complex physics of a 4,000 pound object moving at 80 miles an hour. If we connect this to the bigger picture, knowing the mechanics behind Vidar completely changes how we should interact with our built environment. How so? Well, the mundane highway commute isn't just a stretch of asphalt anymore. It is a live showcase of advanced physics, stereoscopic biology and cutting edge engineering. That's a great way to put it. When you understand the baseline map and the frame rate calculations happening inside

those roadside poles, you realize that our infrastructure is becoming increasingly perceptive. Perceptive? I like that word. It doesn't need to ping us with energy to know we are there. It is passively, silently understanding the geometry of everything moving through it. I think that is the perfect concept to land on. It fits, right? It really does. We've spent this entire deep dive discussing how stereoscopic technology applies to cars on a highway because, well, that is exactly where the 2012 patent and the source material focus. Right. But I want to leave you with a final thought to mull over long after you finish listening. Oh, this is a good one. If stereostopic cameras mounted on a simple roadside pole can calculate exact speed, trajectory, and vehicular data entirely through passive observation, what happens when this passive tracking capability shrinks even further? Right. Because tech always gets smaller. Exactly. What happens when the processing power gets even cheaper, the lenses get even smaller, and the stereoscopic math scales down to track human movement.

That's, yeah, that's wild to think about. How does our world change when the environment around us can calculate the speed, distance, and trajectory of people moving through a public square, a shopping mall, or a simple city sidewalk, just as effortlessly as it tracks a car on the interstate? It completely changes privacy for one. It really does. Just something to think about the next time you walk past a camera. You're listening to a podcast right now driving, working out, walking the dog. If you're in a podcast, chances are you have something to say too. With rss.com, starting your own podcast is free and easy. Upload an episode and we distribute it to Apple podcasts, Spotify, Amazon Music, and more. Track your listeners, see where they're from, and start earning from ads just like this. If you've been thinking about starting a podcast, this is your sign. Start your new podcast for free today at rss.com. The sun shining, birds are singing, and all feels right in the world. Until the season changes, and suddenly you lose your motivation to get out of bed.

In fact, one in five people experience some form of depression no matter the season or time of year. At the American Psychiatric Association Foundation, our vision is to build a mentally healthy nation for all, because we want you to live your best life and be your best you all year round. Please visit mentallyhealthination.org to learn more.

More episodes

More from pplpod

View all episodes →