op calcHist

The FireSight calcHist stage generates a histogram for the current pipeline image. This stage is primarily diagnostic and provides only a subset of the functionality of OpenCV calcHist.

• accumulate If true, bin values for separate channels will be summed. Default is false
• channels is a JSON vector of channel indexes. The default is [], which will use all image channels. For an RGB image, this will be identical to [0 1 2]. For a monochrome image, this will be identical to [0]. The maximum number of channels in the array is 4.
• dims histogram dimensions. Default is 1. At this time, only 1 dimension is supported.
• rangeMax maximum (exclusive)( value to use for range. Default is 256
• rangeMin minimum (inclusive) value to use for range. Default is 0

The current FireSight implementation of calcHist only supports 1-dimensional histograms. Future implementations may support the full capability of OpenCV.

Stage Model

For single channel images or if accumulate is true, the histogram model is a sparse JSON object of integers:

{
"hist":{
   "0":40751,
   "1":60157,
...
   "216":2
  }
}

For convenience, FireSight will generate one histogram for each channel of a color image if accumulate is false. The histogram values are combined into JSON arrays for each histogram bin:

{
"hist":{
   "0":[28983,40751,35947],
   "1":[44643,60157,53446],
...
  "236":[0,0,1]
  }
}

Example 1: multi-channel threshold histogram pipeline

firesight -i img/threshold64-color.png  -p json/calcHist.json

Pixel:1.3ms

The input color image is the result of a [threshold](op threshold) pipeline →

The resulting calcHist model demonstrates that the threshold operation concentrated all pixel values for each channel into either 0 or 255:

{
  "FireSight":{"version":"0.8.5","url":"https://github.com/firepick1/FireSight"},
  "s1":{
    "hist": {"0":[159711,159329,159260],"255":[289,671,740]}
  }
}