Disable automatic brightness control of the Leap Motion LEDs

abhiram1989 2021-01-06 12:29:15 UTC #1

Hi,

I am trying to use the raw images from the Leap Motion cameras to get depth information.

I could get the raw images using the camera API from Leap Motion SDK, and then display them using OpenCV (C++)
However, I noticed this annoying automatic brightness adjustment, which I believe is a feature introduced for a certain purpose. But I do not want it for my application. Please can anyone help me to disable it?

By the way, I saw this library called LeapUVC (https://github.com/leapmotion/leapuvc) which claims to have options to configure the camera. However, I am using a Macbook, and unfortunately it seems LeapUVC is not supported for Macs

I saw a similar question in this post, but there are no answers yet:

Cheers!
AB

dankirby 2021-01-08 09:21:41 UTC #2

You could use UVCLite to read the Leap motion camera images, Mac support for this was added last year. In theory you should be able to set the same configurations such as brightness as in LeapUVC. You will need to work out how to send the data through UVCLite however.

If you are able to get OpenCV to talk to the camera without any intermediate changes then you may be able to use the configurations settings from LeapUVC to work as an alternative:

github.com

leapmotion/leapuvc/blob/master/Python/leapuvc.py#L110


    return self.setLEDsHDRorRotate(3, 1 if enabled else 0)
def setRightLED(self, enabled):
    '''Controls the Right LED. \n(enabled) -> (ret)'''
    return self.setLEDsHDRorRotate(4, 1 if enabled else 0)
def setVerticalCenter(self, value):
    '''Changes the Vertical Center. \n(value) -> (ret)'''
    return self.setLEDsHDRorRotate(5, value)
def crashDevice(self, value):
    '''Crashes the device by trying to set the Vertical Zoom.  Replugging the device in will reset it. \nDon't use this\n(value) -> (ret)'''
    return self.setLEDsHDRorRotate(6, value)
def setDigitalGain(self, value):
    '''Changes the Digital Gain, between 0 and 16. \n(value) -> (ret)'''
    return self.set(cv2.CAP_PROP_BRIGHTNESS, value)
def setGain(self, gain):
    '''Specifies the analog gain as a scalar, between 16 and 63. \n(enabled) -> (ret)'''
    return self.set(cv2.CAP_PROP_GAIN, gain)
    #return self.setGainFPSRatioOrDFrameInterval(0x4000, gain)


# These parameters seem to be out-of-date
#def setGainFPSRatioOrDFrameInterval(self, selector, value):
#    '''Sets Analog Gain (0x4000), FPS Ratio (0x8000), or the Dark Frame Interval (0xc000). \n(selector, value) -> (ret)'''

abhiram1989 2021-01-08 11:22:45 UTC #3

Hi @dankirby, thank you for the optimistic reply!

At the moment I am able to read images from the camera using the camera API from the Leap Motion SDK. I am then using OpenCV to extract information from the images (to get the disparity map for depth sensing). So I am not sure what you mean by getting OpenCV to talk to the camera without intermediate changes. Can you please clarify?

Also, I am using C++ and not Python. Can we somehow use UVCLite with C++?

Thanks,
AB

abhiram1989 2021-01-08 11:45:59 UTC #4

@dankirby Apologies if my first post gave you the impression that I am using OpenCV to get the raw images. Sorry for saying that. I should have said "I could get the images using the camera API from Leap Motion SDK, and then display them using OpenCV". Edited the original post now.