15

Some webcams have an LED next to it which notifies the user it's working.

Hackers are supposed to be able to disable this LED in order to spy their victims without their knowledge, but how can you disable a component, in this case a diode, without making the rest of the circuit unusable?

If an LED is connected in series to the camera sensor, I guess damaging the diode would result in a non-voltage passing component, making the whole webcam circuit unusable.

I'm talking about domestic PC webcams, like the ones you can find in a laptop, for example. I was also thinking in a series connected LED to the webcam in my ignorance of how these systems work.

What I want to know is the way hackers can force a misbehavior of your LED to mislead you and make you think you're not being recorded.

Note that I'm asking this question from the electronic point of view, I'm interested in security, but I'm not trying to get a how-to or something similar.

migrated from electronics.stackexchange.com Nov 28 '16 at 22:32

This question came from our site for electronics and electrical engineering professionals, students, and enthusiasts.

27

Yes, they can and it has been demonstrated before.

Basically, the naive camera design uses a USB-capable microcontroller and drives the "camera on" LED with a GPIO pin. If you can figure out a way to rewrite the firmware of that microcontroller then you can do whatever you like.

I was part of a team that designed high resolution webcams. Part of the design criteria of a particular model was positive indication of a secure channel. The enclosure design included a mechanical door over the lens, and a small hall-effect sensor and magnet that forced the image sensor into reset (through a dedicated RESET# pin on the image sensor) that was forced low when the door was closed.

It was impossible to rewrite the firmware to override that security measure. It wasn't, however, totally safe: if you got physical access to the camera you could defeat this hardware interlock, but that was not part of the scope of this project.

edit

I just wanted to add that in the particular demonstration I linked, the LED was not controlled by an arbitrary GPIO on the FX2LP but rather it was attached to the GPIO that was controlling the image sensor's shutdown signal. This would have been a secure design but the authors of the paper noticed that the shutdown signal was not a "hard" shutdown for the sensor; it was actually an optional input and through careful control of the sensor configuration, the sensor was configured to ignore the shutdown signal. That's even more devious than my original answer. The particular camera I worked on was using the RESET# signal which was not able to be subverted, but it's still a very good lesson for hardware designers.

  • The caveat of "If you can figure out a way to rewrite the firmware" is a pretty big one. There are so many different kinds of cameras out there that this is a pretty big target. – Moshe Katz Nov 29 '16 at 19:02
  • Not really all that hard; a lot of cameras just use a common set of components, and using FX2LP is very common for anything requiring 480Mbps USB, as is downloading the software on boot rather than storing it in non-volatile memory. – akohlsmith Nov 29 '16 at 22:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.