Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Tempe Police Release Video of Uber Accident

twitter.com

Post reply on HN
Print view View original

551–560 of 1001 posts

Re: Tempe Police Release Video of Uber Accident

#551

Post by djsumdog »

cesarb wrote: post #449
ysleepy wrote: post #73

Why is the resolution so low? Why is the FPS so low? Why don't we see the IR/LIDAR footage? Was the brightness adjusted after the fact to look favourably? Why didn't the vehicle issue a full stop once the woman was clearly visible? I feel PR handled.

My guess would be that this is an off-the-shelf dashcam, completely independent from the self-driving cameras and sensors. Being an off-the-shelf dashcam means extracting and publishing the video from it is easy, which is why it came out first. It also explains why it has video from inside the car; many off-the-shelf dashcam models have both a forward-facing and a backwards-facing camera. Personally, I'm more annoyed…

This makes the most sense. It's probably a redundant system: off the shelf dash cam not connected to anything but power. The police could easily extract the sdcard.

The actual car data will need to be collected and released by the NTSB and will probably come in a few weeks.

Re: Tempe Police Release Video of Uber Accident

#552

Post by CamperBob2 »

u801e wrote: post #519
CamperBob2 wrote: post #343

Something that's worth pointing out as well is what happens to low-light contrast sensitivity as we age. I believe there's something like a 50% loss in night vision between the ages of 20 and 50. Having watched the video, I seriously doubt I'd have seen the pedestrian in time to stop for her. I don't know that I would even feel that sorry for her, frankly -- there is only so much we can (and should) do to protect dru…

> It looks like you'd be overdriving these headlights at any speed over 15 MPH or so. The headlamps, to me, appear to be aimed way too low. They should be lighting up the roadway at least 280 feet ahead of the vehicle.

Good point, could be as simple as an aiming issue.

I'm not sure how important the headlights are to the sensor package used by Uber, but this driver needed all the help she could get.

Re: Tempe Police Release Video of Uber Accident

#553

Post by tbabb »

IIRC, high-resolution thermal infrared cameras are ITAR controlled-- which means they are not available for consumer applications above an abysmal resolution (something like 320x280).

Public safety like this seems like a great impetus to do what was done for GPS, and declassify this crucial (and now really quite mature) technology for the general public.

Re: Tempe Police Release Video of Uber Accident

#554

Post by lhorie »

jonlucc wrote: post #497
lhorie wrote: post #435

Earlier quoted context omitted.

> I do wonder if perhaps Uber made the video darker before releasing This video was released by the police department. Are you suggesting someone from Uber drove down to the scene at 10pm, somehow managed to grab the SD card from the dashcam with all the police there, ran back to their car, uploaded it to their laptop, fired up Premiere, edited the brightness, downloaded the modified video back into the SD, ran back…

Alternatively, you're assuming the police on the scene knew how to exract the video from the car without Uber's help? Edit to add: I don't really think they intentionally darken the image, but this probably isn't even the camera the car runs off of.

I was under the impression this was just footage from an off-the-shelf dashcam. I could figure out how to get video out of mine without an instructions manual. Is that not the case here?

Re: Tempe Police Release Video of Uber Accident

#555

Post by scythe »

atonse wrote: post #25

How did LIDAR and IR not catch that? That seems like a pretty serious problem. It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. When I argue for automated driving (as a casual observer), I tell people about exactly this sort of stuff (a computer can look in 20 places at t…

It seems like this perspective may come from the idea that processing camera input is a formality. But the best estimates of the practical computing power of the brain are based on its visual processing capacity, because we know that's a hard problem. CAPTCHAs all depend on humans' ability to process images semantically faster than a computer (spambot). While it probably isn't unsolvable, I don't think it's surprising that this is consistently a challenge.

Re: Tempe Police Release Video of Uber Accident

#556

Post by tbabb »

atonse wrote: post #25

How did LIDAR and IR not catch that? That seems like a pretty serious problem. It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. When I argue for automated driving (as a casual observer), I tell people about exactly this sort of stuff (a computer can look in 20 places at t…

IIRC IR is ITAR controlled and not available above abysmal resolution. I am curious if this limitation applies to self-driving cars, too.

If so, then it's really time to do what was done for GPS and declassify it for use by the general public. It's a public safety/public good issue.

Re: Tempe Police Release Video of Uber Accident

#557

Post by flatline »

heartbreak wrote: post #236
hcknwscommenter wrote: post #137

Earlier quoted context omitted.

Exactly. The dynamic range of the human eye is vastly better than a visible spectrum camera.

> The dynamic range of the human eye is vastly better than a visible spectrum camera. Certainly better than any camera mounted on a dashboard. It's honestly a bit surreal how the pedestrian appears out of the splotch of pure darkness in the frame. That's low dynamic range and resolution (or high compression) at work, not how light behaves in reality.

This was my initial reaction as well, but am not 100% certain after thinking about it some more. Take this for example:

https://i.imgur.com/AlO4h7p.gifv

I figured that light in front of the car was mostly just messing with the camera but that driver sure didn’t see that pedestrian either. I’m willing to give a human driver the benefit of the doubt here and say that even with eyes on the road and hands on the wheel the outcome would likely have been the same. The pedestrian was not highly visible - no reflectors, dark clothes, it’s really hard to see people like this.

Re: Tempe Police Release Video of Uber Accident

#558

Post by linkregister »

In Uber’s own tech blog, it describes its use of LIDAR to generate a wire frame of each object and track it. The lighting should be irrelevant for the vehicle, while it is essential for the driver.

From the below article: “Last fall, Uber officials showing off their vehicles on Mill Avenue said their radar and lidar were able to detect objects, including jaywalkers, as far as 100 yards away and avoid collisions.”

Something went really wrong here, and I look forward to the release of the rest of the data. I’m wondering why the woman, who had a large enough silhouette, wasn’t detected by the software.

This is a better source: https://www.google.com/amp/s/amp.azcentral.com/story/4476480...

Re: Tempe Police Release Video of Uber Accident

#559

Post by jozzas »

sova wrote: post #376
mirimir wrote: post #106

Earlier quoted context omitted.

Indeed! To LIDAR, she was basically standing in the lane, with a bulky bicycle. To visible light, including the driver, who was apparently half asleep or watching the dashboard, she was in shadow until just before the collision. So yes, LIDAR should have caught this. Easily. So something was clearly misconfigured. And even if the driver had been carefully watching the road, he probably wouldn't have seen her in time.…

LIDAR view on dashboard! Uber, if you are reading, take heed.

How would this have helped at 40MPH? The user would have milliseconds to react and hit the brakes. The point is that the car is self-driving. If a user has to watch a video display and intervene for every edge case it's more dangerous than just driving yourself.
Post reply on HN