NDPluginStats¶
| author: | Mark Rivers, University of Chicago |
|---|
Contents
Overview¶
NDPluginStats computes the following.
- Basic statistics: minimum, maximum, mean, sigma, total, and net (background subtracted).
- Centroid, sigma, skewness and kurtosis values in the X and Y dimensions.
- Profiles of the array in the X and Y dimensions. A total of 8
profiles are calculated:
- The average profiles in the X and Y directions.
- The average profiles in the X and Y directions, for array elements greater than the centroid threshold.
- The profiles in the X and Y directions at the X and Y centroid position.
- The profiles in the X and Y directions at a user-defined X and Y cursor position.
- A histogram of the values (e.g. number of pixels versus intensity per pixel).
Each calculcation can be independently enabled and disabled. Calculations 1 and 4 can be perfomed on arrays of any dimension. Calculations 2 and 3 are restricted to 2-D arrays.
Time-series arrays of the basic statistics, centroid and sigma statistics can also be collected. This is very useful for on-the-fly data acquisition, where the NDStats plugin computes the net or total counts in the detector or an ROI. It can also be used to quickly plot a time-history of beam position or width, etc. The time series can also be used in a circular buffer mode where is continuously displays the last N values of the statistic.
NDPluginStats inherits from NDPluginDriver. The NDPluginStats class documentation describes this class in detail.
NDPluginStats.h defines the following parameters. It also implements all
of the standard plugin parameters from
NDPluginDriver. The EPICS database
NDStats.template provide access to these parameters, listed in the
following table. Note that to reduce the width of this table the
parameter index variable names have been split into 2 lines, but these
are just a single name, for example NDPluginStatsComputeStatistics.
| Parameter Definitions in NDPluginStats.h and EPICS Record Definitions in NDStats.template | ||||||
|---|---|---|---|---|---|---|
| Parameter index variable | asyn interface | Access | Description | drvInfo string | EPICS record name | EPICS record type |
| Basic statistics | ||||||
| NDPluginStats ComputeStatistics |
asynInt32 | r/w | Flag to control whether to compute statistics for this array (0=No, 1=Yes). Not computing statistics reduces CPU load. Basic statistics computations are quite fast, since they involve mostly double precision addition, with 1 multiply to compute sigma, per array element. | COMPUTE_STATISTICS | $(P)$(R)ComputeStatistics, $(P)$(R)ComputeStatistics_RBV | bo, bi |
| NDPluginStats BgdWidth |
asynInt32 | r/w | Width of the background in pixels to use when computing net counts. 0=no background subtraction, so the net counts is the same as the total counts. | BGD_WIDTH | $(P)$(R)BgdWidth, $(P)$(R)BgdWidth_RBV | longout, longin |
| NDPluginStats MinValue |
asynFloat64 | r/o | Minimum value in any element in the array | MIN_VALUE | $(P)$(R)MinValue_RBV | ai |
| NDPluginStats MinX |
asynFloat64 | r/o | X pixel location of minimum value in the array. This is only valid for 2-D monochromatic arrays. | MIN_X | $(P)$(R)MinX_RBV | ai |
| NDPluginStats MinY |
asynFloat64 | r/o | Y pixel location of minimum value in the array. This is only valid for 2-D monochromatic arrays. | MIN_Y | $(P)$(R)MinY_RBV | ai |
| NDPluginStats MaxValue |
asynFloat64 | r/o | Maximum value in any element in the array | MAX_VALUE | $(P)$(R)MaxValue_RBV | ai |
| NDPluginStats MaxX |
asynFloat64 | r/o | X pixel location of maximum value in the array. This is only valid for 2-D monochromatic arrays. | MAX_X | $(P)$(R)MaxX_RBV | ai |
| NDPluginStats MaxY |
asynFloat64 | r/o | Y pixel location of maximum value in the array. This is only valid for 2-D monochromatic arrays. | MAX_Y | $(P)$(R)MaxY_RBV | ai |
| NDPluginStats MeanValue |
asynFloat64 | r/o | Mean value in the array | MEAN_VALUE | $(P)$(R)MeanValue_RBV | ai |
| NDPluginStats Total |
asynFloat64 | r/o | Sum (total) of all elements in the array. This is available as an ai record. The total counts are also available as epicsInt32 values in an mca record via callbacks to the drvFastSweep driver. The mca record is very useful for on-the-fly data acquisition of the total counts in the detector or in an ROI. | TOTAL | $(P)$(R)Total_RBV, $(P)$(R)TotalArray | ai, mca |
| NDPluginStats Net |
asynFloat64 | r/o | Net (background subtracted) total of all elements in the array. The background is calculated by determining the average counts per array element in a border around the array of width NDPluginStatsBgdWidth. This average background counts per element is then subtracted from all elements inside the array. If NDPluginStatsBgdWidth is ≤ 0 then no background is computed. The net counts is available as an ai record. The net counts is also available as epicsInt32 values in an mca record via callbacks to the drvFastSweep driver. The mca record is very useful for on-the-fly data acquisition of the net counts in the detector or in an ROI. | NET | $(P)$(R)Net_RBV, $(P)$(R)NetArray | ai, mca |
| NDPluginStats SigmaValue |
asynFloat64 | r/o | Sigma (standard deviation) of all elements in the array | SIGMA_VALUE | $(P)$(R)Sigma_RBV | ai |
| Centroid statistics | ||||||
| NDPluginStats ComputeCentroid |
asynInt32 | r/w | Flag to control whether to compute the centroid statistics (0=No, 1=Yes). The centroids are computed from the average row and column profiles above the centroid threshold. These calculations are also quite fast, since they just involve addition operations for each array element. | COMPUTE_CENTROID | $(P)$(R)ComputeCentroid, $(P)$(R)ComputeCentroid_RBV | bo, bi |
| NDPluginStats CentroidThreshold |
asynFloat64 | r/w | Threshold used when computing the centroid statistics. All array elements less than this value are set to 0 for computing the centroid statistics. It is important to set this value to ignore the “background” when computing the position and size of a “beam” image, for example. | CENTROID_THRESHOLD | $(P)$(R)CentroidThreshold, $(P)$(R)CentroidThreshold_RBV | ao, ai |
| NDPluginStats CentroidTotal |
asynFloat64 | r/o | Total mass, sum of all elements above the threshold. | CENTROID_TOTAL | $(P)$(R)CentroidTotal_RBV | ai |
| NDPluginStats CentroidX |
asynFloat64 | r/o | X centroid of the array above the centroid threshold. | CENTROIDX_VALUE | $(P)$(R)CentroidX_RBV | ai |
| NDPluginStats CentroidY |
asynFloat64 | r/o | Y centroid of the array above the centroid threshold. | CENTROIDY_VALUE | $(P)$(R)CentroidY_RBV | ai |
| NDPluginStats SigmaX |
asynFloat64 | r/o | Sigma X (width) of the distribution above the centroid threshold. | SIGMAX_VALUE | $(P)$(R)SigmaX_RBV | ai |
| NDPluginStats SigmaY |
asynFloat64 | r/o | Sigma Y (height) of the distribution above the centroid threshold. | SIGMAY_VALUE | $(P)$(R)SigmaY_RBV | ai |
| NDPluginStats SigmaXY |
asynFloat64 | r/o | This is the normalized value of sigmaXY, i.e. sigmaXY/(sigmaX * sigmaY). This is often called the correlation coefficient, r. It is zero if the X and Y profiles are not correlated, meaning that the distribution is not tilted with respect to the X and Y axes. | SIGMAXY_VALUE | $(P)$(R)SigmaXY_RBV | ai |
| NDPluginStats SkewX |
asynFloat64 | r/o | Skewness X (symmetry) of the distribution above the centroid threshold, in relation
to the center of mass. == 0, symmetric distribution < 0, distribution assymetric to the left > 0, distribution assymetric to the right |
SKEWX_VALUE | $(P)$(R)SkewX_RBV | ai |
| NDPluginStats SkewY |
asynFloat64 | r/o | Skewness Y (symmetry) of the distribution above the centroid threshold, in relation
to the center of mass. == 0, symmetric distribution < 0, distribution assymetric to the top > 0, distribution assymetric to the bottom |
SKEWY_VALUE | $(P)$(R)SkewY_RBV | ai |
| NDPluginStats KurtosisX |
asynFloat64 | r/o | Excess Kurtosis X (flatness) of the distribution above the centroid threshold == 0, Gaussian (normal) distribution < 0, distribution flatter than normal > 0, distribution more peaky than normal |
KURTOSISX_VALUE | $(P)$(R)KurtosisX_RBV | ai |
| NDPluginStats KurtosisY |
asynFloat64 | r/o | Excess Kurtosis Y (flatness) of the distribution above the centroid threshold. == 0, Gaussian (normal) distribution < 0, distribution flatter than normal > 0, distribution more peaky than normal |
KURTOSISY_VALUE | $(P)$(R)KurtosisY_RBV | ai |
| NDPluginStats Eccentricity |
asynFloat64 | r/o | Eccentricity, can take values from 0 to 1. 0 means a perfectly round object and 1 mean a line shaped object. | ECCENTRICITY_VALUE | $(P)$(R)Eccentricity_RBV | ai |
| NDPluginStats Orientation |
asynFloat64 | r/o | Orientation of the object, orientation of the “long” direction with respect to horizontal (x axis). | ORIENTATION_VALUE | $(P)$(R)Orientation_RBV | ai |
| Time-Series data | ||||||
| The time series is implemented by loading an instance of the NDPluginTimeSeries
plugin for each NDPluginStats plugin, and the time series control
uses records in NDTimeSeries.template. That documentation should be consulted for
an explanation of these records. The prefix and record name macro for the time-series
plugin records from NDTimeSeries.template is $(P)$(R)TS:. NOTE: The time-series plugin is often used with drivers which sample at a fixed well-defined time interval. This cannot be guaranteed with the statistics plugin, so the averaging time records and time axis waveform record from NDPluginTimeSeries are typically not used, and the statistics data are plotted against time point #, rather than actual time. The time-series waveform records for each statistic are defined in NDStats.template. |
||||||
| NDPluginTimeSeries, TSTimeSeries | asynFloat64Array | r/o | The time series data arrays of the basic statistics and centroid and sigma statistics described above. | TS_TIME_SERIES | $(P)$(R)XXX, where XXX is: TSMinValue TSMinX TSMinY TSMaxValue TSMaxX TSMaxY TSMeanValue TSSigma TSTotal TSNet TSCentroidX TSCentroidY TSSigmaX TSSigmaY TSSigmaXY TSSkewX TSSkewY TSKurtosisX TSKurtosisY TSEccenticity TSOrientation TSTimestamp |
waveform |
| X and Y Profiles | ||||||
| NDPluginStats ComputeProfiles |
asynInt32 | r/w | Flag to control whether to compute the profiles for this array (0=No, 1=Yes). | COMPUTE_PROFILES | $(P)$(R)ComputeProfiles, $(P)$(R)ComputeProfiles_RBV | bo, bi |
| NDPluginStats ProfileSizeX |
asynInt32 | r/w | Number of array elements in the X profiles. | PROFILE_SIZE_X | $(P)$(R)ProfileSizeX_RBV | longin |
| NDPluginStats ProfileSizeY |
asynInt32 | r/w | Number of array elements in the Y profiles. | PROFILE_SIZE_Y | $(P)$(R)ProfileSizeY_RBV | longin |
| NDPluginStats CursorX |
asynInt32 | r/w | X position of a user-defined cursor for profiles. | CURSOR_X | $(P)$(R)CursorX, $(P)$(R)CursorX_RBV | longout, longin |
| NDPluginStats CursorY |
asynInt32 | r/w | Y position of a user-defined cursor for profiles. | CURSOR_Y | $(P)$(R)CursorY, $(P)$(R)CursorY_RBV | longout, longin |
| NDPluginStats CursorVal |
asynFloat64 | r/o | Value of the pixel at the current cursor position. | CURSOR_VAL | $(P)$(R)CursorVal | ai |
| NDPluginStats ProfileAverageX |
asynFloat64Array | r/o | Profile of the average row in the array, i.e. the sum of all rows in the array divided by the number of rows. | PROFILE_AVERAGE_X | $(P)$(R)ProfileAverageX_RBV | waveform |
| NDPluginStats ProfileAverageY |
asynFloat64Array | r/o | Profile of the average column in the array, i.e. the sum of all columns in the array divided by the number of columns. | PROFILE_AVERAGE_Y | $(P)$(R)ProfileAverageY_RBV | waveform |
| NDPluginStats ProfileThresholdX |
asynFloat64Array | r/o | Same as ProfileAverageX except that all array elements less than CentroidThreshold are set to zero when computing the average. | PROFILE_THRESHOLD_X | $(P)$(R)ProfileThresholdX_RBV | waveform |
| NDPluginStats ProfileThresholdY |
asynFloat64Array | r/o | Same as ProfileAverageY except that all array elements less than CentroidThreshold are set to zero when computing the average. | PROFILE_THRESHOLD_Y | $(P)$(R)ProfileThresholdY_RBV | waveform |
| NDPluginStats ProfileCentroidX |
asynFloat64Array | r/o | X profile through the array in the row defined by CentroidY. | PROFILE_CENTROID_X | $(P)$(R)ProfileCentroidX_RBV | waveform |
| NDPluginStats ProfileCentroidY |
asynFloat64Array | r/o | Y profile through the array in the column defined by CentroidX. | PROFILE_CENTROID_Y | $(P)$(R)ProfileCentroidY_RBV | waveform |
| NDPluginStats ProfileCursorX |
asynFloat64Array | r/o | X profile through the array in the row defined by CursorY. | PROFILE_CURSOR_X | $(P)$(R)ProfileCursorX_RBV | waveform |
| NDPluginStats ProfileCursorY |
asynFloat64Array | r/o | Y profile through the array in the row defined by CursorX. | PROFILE_CURSOR_Y | $(P)$(R)ProfileCursorY_RBV | waveform |
| Array histogram | ||||||
| NDPluginStats ComputeHistogram |
asynInt32 | r/w | Flag to control whether to compute the histogram for this array (0=No, 1=Yes). Not computing the histogram reduces CPU load. | COMPUTE_HISTOGRAM | $(P)$(R)ComputeHistogram, $(P)$(R)ComputeHistogram_RBV | bo, bi |
| NDPluginStats HistSize |
asynInt32 | r/w | Number of elements (bins) in the histogram | HIST_SIZE | $(P)$(R)HistSize, $(P)$(R)HistSize_RBV | longout, longin |
| NDPluginStats HistMin |
asynFloat64 | r/w | Minimum value for the histogram. All values less than this will be counted in HistBelow. | HIST_MIN | $(P)$(R)HistMin, $(P)$(R)HistMin_RBV | ao, ai |
| NDPluginStats HistMax |
asynFloat64 | r/w | Maximum value for the histogram. All values greater than this will be counted in HistAbove. | HIST_MAX | $(P)$(R)HistMax, $(P)$(R)HistMax_RBV | ao, ai |
| NDPluginStats HistBelow |
asynInt32 | r/o | Count of all values less than HistMin. | HIST_BELOW | $(P)$(R)HistBelow_RBV | longin |
| NDPluginStats HistAbove |
asynInt32 | r/o | Count of all values greater than HistMax. | HIST_ABOVE | $(P)$(R)HistAbove_RBV | longin |
| NDPluginStats HistEntropy |
asynFloat64 | r/o | Entropy of the image. This is a measure of the sharpness of the histogram, and is often a useful figure of merit for determining sharpness of focus, etc. It is defined as -SUM(BIN[i]*log(BIN[i]), where the sum is over the number of bins in the histogram and BIN[i] is the number of elements in bin i. | HIST_ENTROPY | $(P)$(R)HistEntropy_RBV | ai |
| NDPluginStats HistArray |
asynFloat64Array | r/o | Histogram array, i.e. counts in each histogram bin. | HIST_ARRAY | $(P)$(R)Histogram_RBV | waveform |
| NDPluginStats HistXArray |
asynFloat64Array | r/o | Histogram X-axis array, i.e. minimum intensity in each histogram bin. | HIST_X_ARRAY | $(P)$(R)HistogramX_RBV | waveform |
If the values of CentroidThreshold, CursorX, or CursorY are changed then the centroid and profile calculations are performed again immediately on the last array collected. Thus updated centroid statistics and profiles can be displayed even when new arrays are not being acquired. These calculations are only performed when enabled by ComputeCentroid and ComputeProfiles.
Configuration¶
The NDPluginStats plugin is created with the NDStatsConfigure command, either from C/C++ or from the EPICS IOC shell.
NDStatsConfigure(const char *portName, int queueSize, int blockingCallbacks,
const char *NDArrayPort, int NDArrayAddr,
int maxBuffers, size_t maxMemory,
int priority, int stackSize)
For details on the meaning of the parameters to this function refer to the detailed documentation on the NDStatsConfigure function in the NDPluginStats.cpp documentation and in the documentation for the constructor for the NDPluginStats class.
Screen shots¶
The following MEDM screen provides access to the parameters in NDPluginDriver.h and NDPluginStats.h through records in NDPluginBase.template and NDStats.template.
The following MEDM screen shows the average profile of an image in the X direction.
The following MEDM screen shows the profile of an image in the Y direction at the location of the user-defined cursor.
The following MEDM screen shows the histogram of intensities of an array.
The following MEDM screen combines many parameters for 5 NDPluginStats plugins on a single screen.
The following MEDM screen shows the the total counts from the Stats1 plugin. This is the total counts as a function of time.
The following MEDM screen shows the Y centroid as a function of time from the Stats1 plugin.
The following MEDM screen shows all of the basic statistics as a function of time from the Stats1 plugin.
The following MEDM screen shows all of the centroid statistics as a function of time from the Stats1 plugin.
