---
name: "gpiv_errvec"
section: "1"
platform: "linux"
package: "gpivtools"
version: "0.6.0-3.1"
release: "Debian 8 (Jessie)"
platforms: ["linux"]
url: "https://new.carta.tech/gpiv_errvec"
---

# gpiv\_errvec(1)

Searches erroneous vectors in a PIV data file and substitutes with new values.

## NAME

gpiv\_errvec - Searches erroneous vectors in a PIV data file and substitutes with new values.

## SYNOPSIS

**gpiv\_errvec**  \[**-a**\]  \[**-b** \] \[**-g**\]\[**-h**\] \[**-h** \| **--help**\] \[**-i** \| **--image\_type** *S*\] \[**-n** *N*\]  \[**--no\_b** \]  \[**-p** \| **--print**\]  \[**-r** *N*\]  \[**-s** *N*\]  \[**-t** *F*\]  \[**-u** \| **--surround**\]   \[**-v** \| **--version**\]   \[*filename*\] \< *stdin* \> *stdout*

## DESCRIPTION

**gpiv\_errvec** *searches erroneous outliers in a PIV data stream and* *substitutes with new values, if possible, i.e. if (the majority) of its* *surroundings are non-erroneous.  Evaluation of erroneous vectors is* *done by looking up the Signal to Noise Ratio (SNR) value in the PIV* *data file or by calculating the median residu of the vector with its* *surroundings. A new displacement is obtained by the mean value of the* *surrounding values or by the location of the next higher peak in the* *covariance function.*

*PIV data validation is not needed anymore while this is now done* *during image interrogation with* **gpiv\_rr***. Though using* **gpiv\_errvec** *will not harm.*

*The configuration parameters (containing the* **VALID** key) may be overruled by the command line options, as explained below.

## Options

- **-a**

  Automatic threshold; calculates threshold value from residu histogram.

- **-b**

  Calculate statistics of the median residues and print a histogram of  *N* bins.

- **-g**

  Graphical visualization of residu statistics with gnuplot (needs  *filename*). **-b** is not required as this option will be  set automatically.

- **-h** \| **--help**

  On-line help.

- **-i** \| **--image\_type** *S*

  Image type or format to be used if particle displacements will have to be re-estimated at the next highest correlation peak. *S* may be png, tif, gif, bmp, pgm, r or gpi. Default: png.

- **-n** *N*

  Defines number of histogram of bins *N* for **-b** or **-a**

- **--no\_b** *N*

  **suppresses printing median residu histogram**

- **-p** \| **--print**

  Print parameters, command line options and eventually used input and output filenames to stdout. The output is  identic of *filename***.par**, in case *filename* is used.

- **-r** *N*

  **Erroneous vector determined by SNR of the correlation function (***N* = 0), median residu (*N* = 1) or normalized median (*N* = 2), i.e. the ratio between median residu with current datapoint included and between median residu of surroundigs (with datapoint excluded).

- **-s** *N*

  Substitution of the erroneous vectors by (*N* = 0) nothing, the local mean from the surrounding values (*N* = 1), median (*N* = 2) or next highest covariance peak (*N* = 3, needs *filename*).

- **-t** *F*

  **Threshold value** *F* **of SNR or median residu to** **determine validity of particle displacement.**

- **-u** \| **--surround**

  Odd number representing the surrounding grid points to be used for local mean and median values.

- **-v** \| **--version**

  Print version information on standard output, then exit successfully.

- *filename*

  Full name of PIV data file for input. Substitutes stdin and stdout. Output will be written to *filename***.erv.piv**. The parameters will be written or appended to *filename***.par** and may be used for future use by including them in *./gpivrc*. If particle displacements will have to be re-estimated at the next highest correlation peak, *filename* is required.

## SEE ALSO

gpivtools

## AUTHOR

Gerber Van der Graaf

## BUGS

The program has been tested for artificial generated DPIV images. So far, no serious bugs have been found.
