2D·

LiDAR destroys camera

And thus probably becomes inadmissible - what do you think?


https://x.com/Strombock/status/1956332289392083265


The problem cannot be completely solved physically: LiDAR lasers (e.g. 1550 nm) generate intense IR beams that thermally damage CMOS sensors and burn out pixels. Reduction by filters or weaker lasers reduces range/accuracy. Recent cases (e.g. Volvo 2025) show permanent damage to cameras/mobile phones. No perfect fix without compromise.


Tesla's solution therefore seems the only sensible one. ($TSLA (+0.85%) )

5
12 Comments

profile image
That's nonsense.....

Hardware protection on the camera system

Place IR blocking filters in front of the lens to block wavelengths >900 nm.
OR
Use special laser protection filters that only block the LiDAR-specific range (e.g. 905 ± 10 nm).
This also works with the night vision camera.

Optimize camera position and alignment.
Align slightly downwards or upwards to avoid direct "point hits".

Use a wide-angle lens instead of a telephoto lens, as this distributes the laser over more pixels (less point energy).

There is AI-based camera software that detects laser reflections and immediately adjusts the exposure in affected areas to prevent overexposure damage.

Mechanical shutters or variable ND filters could automatically stop down when laser light is detected.

In many cities, security camera operators are now using dual camera systems:
One camera for visible light (protected with IR blocker)
A second camera for pure IR night vision (for surveillance at night)
In this way, the visible camera cannot be damaged by LiDAR and the night vision function is not lost.
10
profile image
@7Trader and then you do this millions of times on all+ smartphones simply so that you can approve a new tech, right?
profile image
@7Trader Boy, if thousands of vehicles drive around and simply destroy the clans of iPhones, I don't want to know what compensation must be behind it - that's unacceptable imo
profile image
@7Trader if you really think everyone is retrofitting hardware protection... you're completely lost
profile image
@7Trader absolutely unfuckingbelievable that people still like hardware protection on billions of cams 🤣🤣🤣🤣🫷
profile image
@PaulPK You seem to be very upset because your problem is not the 1,000,000 dollar per share Tesla bull case.
profile image
@Nick-investing I certainly see it quite differently - and you will have to see it that way too
profile image
@PaulPK Fortunately, the world doesn't need to rely on you, but there are clever people who can solve these little problems. There are already several approaches on various sites, so you don't need to be afraid.... 🤗

1. changes at the LiDAR manufacturers

Shift wavelength → Go to 1550 nm instead of 905 nm (some premium LiDARs already do this).
→ Advantage: 1550 nm is almost completely blocked by camera IR filters and is significantly less dangerous for sensors.

Optimize power → Control the laser so that the energy per pulse is below the damage threshold for cameras without losing range (through better detectors).

Adjust the scattering pattern → No point laser beam, but slightly defocused spots that emit less energy per pixel.

2. changes at the camera manufacturers

Install standard IR protection against 905 nm in all sensors (even in surveillance cameras with night vision by blocking only the harmful area).

Apply laser protective coatings to front lenses that reflect or absorb high-energy IR pulses.

Automatic detection of laser pulses in the image signal → camera could reduce exposure in microseconds before damage occurs.

3. legal & standards solutions

Regulation of laser power for civil LiDAR systems, similar to laser pointers.

Establish uniform standards for LiDAR wavelengths that do not endanger camera sensors.

4. transitional solutions for existing cameras

For surveillance cameras: retrofit IR blocking filters or special 905 nm blockers.
2
profile image
@7Trader you are a clever ruski 🤣
profile image
@PaulPK better keep playing with your mushrooms and stay away from the big boys 😉
2
Tesla to the 🌛
1
profile image
Firstly, this has been known for months and secondly, it only happens if you film too close up. You simply don't have to take close-ups
Join the conversation