Cam-e-lot - Round 12 - Man vs light

In this post I will explain how the Cam-e-lot sensor sends the level of light in the room to the OpenHAB system.

There two steps involved: calculate the level and light from the image captured by the RaspiCam and send the value to OpenHAB by means of the enOcean transceiver.

Determine level of light in the room

Thanks to the OpenCV library, this task can be accomplished quite easily. The key of the problem is to use the proper color space. The RaspiCam saves images using the widely-used RGB (red-green-blue) format. It's difficult to extract information about light from this color space. However, if you transform the image from the RGB color space to the HSL (hue-saturation-light) color space, everything turns out to be very easy. In the HSL color space, the information about the amount of light can be read from a specific channel: the L channel

So, let's first convert image from RGB to HSL color space

    CvSize imgSize = cvGetSize(image);

    IplImage* hlsImage = cvCreateImage(imgSize, image->depth, 3);

cvCvtColor(image, hlsImage, CV_RGB2HLS);

Then split the image into its three channels

    IplImage* h = cvCreateImage(imgSize, image->depth, 1);

    IplImage* l = cvCreateImage(imgSize, image->depth, 1);

    IplImage* s = cvCreateImage(imgSize, image->depth, 1);

cvSplit(hlsImage, h, l, s, NULL);         

Finally, let's average all the values of the L channel to have a measure of the amount of light in the image

    CvScalar avg = cvAvg(l);

Send the measurement to OpenHAB

Now let's face the next challenge.

The enOcean profile I selected (D2-01) includes the possibility to send a measure express on 4 bytes.

This can be accomplished by sending the command 0x07 (Actuator measure response)

The format of the telegram for this command is as follow

where

• CMD
• Dim value
• I/O Channel
• Measured value

To support the new command, I made some changes to the Create method.

eoReturn eoD2EEProfile::Create(eoMessage &m)

{

// New code begins here

  if (command == CMD_MEASUREMENT_RESPONSE)

  {

       tel.RORG = rorg;

       tel.SetDataLength(6);

       tel.destinationID = destinationID;

       tel.sourceID = 0;

// bit 0: power failure (0 -> not supported)

// bit 1: power failure detected (0)

// bit 2..3: not used

// bit 4..7: cmd

       tel.data[0] = command;

// bit 8: over current switch off (0 -> not supported)

// bit 9..10: error level (0 -> hardware OK)

// bit 11..15: IO channel

       tel.data[1] = 0x00 | (channel & 0x1F);

       // bit 16: local control (0 -> not supported)

// but 17..39: output value

       tel.data[2] = msg.data[2];

       tel.data[3] = msg.data[3];

       tel.data[4] = msg.data[4];

       tel.data[5] = msg.data[5];

       return EO_OK;

  }

// New code ends here

  return NOT_SUPPORTED;

}

Build the EOLink library as I wrote in my second post and copy the libEOLink.a file (located in EOLink\ReleaseLib folder) to the /usr/lib folder of your Raspberry Pi board

Sending commands

To send commands without breaking up the interface implemented by the eoProfile class, I decided to set the command to perform and the I/O channel of interest by invoking the SetValue method with E_COMMAND for  the  channelType parameter and 0 or 1 for the index parameter to set respectively the command and the I/O channel. So the code to send an actuator status change is the following

eoProfile *sendProf = eoProfileFactory::CreateProfile(0xD2,0x01,0x00);

eoMessage mytel = eoMessage(6);

// set command to perform (Actuator measurement update)

sendProf->SetValue(E_COMMAND, (uint8_t)0x07, 0);

// set I/O channel (always 0 for measurement)

sendProf->SetValue(E_COMMAND, (uint8_t)0, 1);

// set new value

sendProf->SetValue(E_IO_CHANNEL, (float)value, 0);

sendProf->Create(mytel);

enocean_gateway.Send(mytel);

OpenHAB configuration

To see the measured amount of light, just add a Number item in the items definition file using the ENERGY_WS parameter as in this example

Number LightLevel "Light [%d]" (gCamelot) {aleoncean="LOCALID=01:8E:39:72,REMOTEID=01:85:3A:3A,TYPE=RD_D2-01-00,PARAMETER=ENERGY_WS"}

At this moment in time, the measured value ranges from 0 to 255. I would need a lux meter or another calibration tool to determine a linear mapping between such a value and the actual luminous flux