create_bg_estiT_create_bg_estiCreateBgEstiCreateBgEsticreate_bg_esti (Operator)

Name

create_bg_estiT_create_bg_estiCreateBgEstiCreateBgEsticreate_bg_esti — Generate and initialize a data set for the background estimation.

Signature

create_bg_esti(InitializeImage : : Syspar1, Syspar2, GainMode, Gain1, Gain2, AdaptMode, MinDiff, StatNum, ConfidenceC, TimeC : BgEstiHandle)

Description

create_bg_esticreate_bg_estiCreateBgEstiCreateBgEsticreate_bg_esti creates a new data set for the background estimation and initializes it with the appropriate parameters. The estimated background image is part of this data set. The newly created set automatically becomes the current set.

InitializeImageInitializeImageInitializeImageinitializeImageinitialize_image is used as an initial prediction for the background image. For a good prediction an image of the observed scene without moving objects should be passed in InitializeImageInitializeImageInitializeImageinitializeImageinitialize_image. That way the foreground adaptation rate can be held low. If there is no empty scene image available, a homogeneous gray image can be used instead. In that case the adaptation rate for the foreground image must be raised, because initially most of the image will be detected as foreground. The initialization image must to be of type byte or real. Because of processing single-channel images, data sets must be created for every channel. Size and region of InitializeImageInitializeImageInitializeImageinitializeImageinitialize_image determines size and region for all background estimations (run_bg_estirun_bg_estiRunBgEstiRunBgEstirun_bg_esti) that are performed with this data set.

Syspar1Syspar1Syspar1syspar1syspar_1 and Syspar2Syspar2Syspar2syspar2syspar_2 are the parameters of the Kalman system matrix. The system matrix describes the system of the gray value changes according to Kalman filter theory. The background estimator implements a different system for each pixel.

GainModeGainModeGainModegainModegain_mode defines whether a fixed Kalman gain should be used for the estimation or whether the gain should adapt itself depending on the difference between estimation and actual value. If GainModeGainModeGainModegainModegain_mode is set to 'fixed'"fixed""fixed""fixed""fixed", then Gain1Gain1Gain1gain1gain_1 is used as Kalman gain for pixels predicted as foreground and Gain2Gain2Gain2gain2gain_2 as gain for pixels predicted as background. Gain1Gain1Gain1gain1gain_1 should be smaller than Gain2Gain2Gain2gain2gain_2, because adaptation of the foreground should be slower than adaptation of the background. Both Gain1Gain1Gain1gain1gain_1 and Gain2Gain2Gain2gain2gain_2 should be smaller than 1.0.

If GainModeGainModeGainModegainModegain_mode is set to 'frame'"frame""frame""frame""frame", then tables for foreground and background estimation are computed containing Kalman gains for all the 256 possible gray value changes. Gain1Gain1Gain1gain1gain_1 and Gain2Gain2Gain2gain2gain_2 then denote the number of frames necessary to adapt the difference between estimated value and actual value. So with a fixed time for adaptation (i.e. number of frames) the needed Kalman gain grows with the gray value difference. Gain1Gain1Gain1gain1gain_1 should therefore be larger than Gain2Gain2Gain2gain2gain_2. Different gains for different gray value differences are useful if the background estimator is used for generating an 'empty' scene assuming that there are always moving objects in the observed area. In that case the adaptation time for foreground adaptation (Gain1Gain1Gain1gain1gain_1) must not be too big. Gain1Gain1Gain1gain1gain_1 and Gain2Gain2Gain2gain2gain_2 should be bigger than 1.0.

AdaptModeAdaptModeAdaptModeadaptModeadapt_mode denotes, whether the foreground/background decision threshold applied to the gray value difference between estimation and actual value is fixed or whether it adapts itself depending on the gray value deviation of the background pixels.

If AdaptModeAdaptModeAdaptModeadaptModeadapt_mode is set to 'off'"off""off""off""off", the parameter MinDiffMinDiffMinDiffminDiffmin_diff denotes a fixed threshold. The parameters StatNumStatNumStatNumstatNumstat_num, ConfidenceCConfidenceCConfidenceCconfidenceCconfidence_c and TimeCTimeCTimeCtimeCtime_c are meaningless in this case.

If AdaptModeAdaptModeAdaptModeadaptModeadapt_mode is set to 'on'"on""on""on""on", then MinDiffMinDiffMinDiffminDiffmin_diff is interpreted as a base threshold. For each pixel an offset is added to this threshold depending on the statistical evaluation of the pixel value over time. StatNumStatNumStatNumstatNumstat_num holds the number of data sets (past frames) that are used for computing the gray value variance (FIR-Filter). ConfidenceCConfidenceCConfidenceCconfidenceCconfidence_c is used to determine the confidence interval.

The confidence interval determines the values of the background statistics if background pixels are hidden by a foreground object and thus are detected as foreground. According to the student t-distribution the confidence constant is 4.30 (3.25, 2.82, 2.26) for a confidence interval of 99,8% (99,0%, 98,0%, 95,0%). TimeCTimeCTimeCtimeCtime_c holds a time constant for the exp-function that raises the threshold in case of a foreground estimation of the pixel. That means, the threshold is raised in regions where movement is detected in the foreground. That way larger changes in illumination are tolerated if the background becomes visible again. The main reason for increasing this tolerance is the impossibility for a prediction of illumination changes while the background is hidden. Therefore no adaptation of the estimated background image is possible.

Attention

If GainModeGainModeGainModegainModegain_mode was set to 'frame'"frame""frame""frame""frame", the run-time can be extremely long for large values of Gain1Gain1Gain1gain1gain_1 or Gain2Gain2Gain2gain2gain_2, because the values for the gains' table are determined by a simple binary search.

Execution Information

Multithreading type: reentrant (runs in parallel with non-exclusive operators).
Multithreading scope: global (may be called from any thread).
Processed without parallelization.

This operator returns a handle. Note that the state of an instance of this handle type may be changed by specific operators even though the handle is used as an input parameter by those operators.

Parameters

InitializeImageInitializeImageInitializeImageinitializeImageinitialize_image (input_object) singlechannelimage → object (byte / real)

initialization image.

Syspar1Syspar1Syspar1syspar1syspar_1 (input_control) real → (real)

1. system matrix parameter.

Default: 0.7

Suggested values: 0.65, 0.7, 0.75

Value range: 0.05 ≤ Syspar1 Syspar1 Syspar1 syspar1 syspar_1 ≤ 1.0

Recommended increment: 0.05

Syspar2Syspar2Syspar2syspar2syspar_2 (input_control) real → (real)

2. system matrix parameter.

Default: 0.7

Suggested values: 0.65, 0.7, 0.75

Value range: 0.05 ≤ Syspar2 Syspar2 Syspar2 syspar2 syspar_2 ≤ 1.0

Recommended increment: 0.05

GainModeGainModeGainModegainModegain_mode (input_control) string → (string)

Gain type.

Default: 'fixed' "fixed" "fixed" "fixed" "fixed"

List of values: 'fixed'"fixed""fixed""fixed""fixed", 'frame'"frame""frame""frame""frame"

Gain1Gain1Gain1gain1gain_1 (input_control) real → (real)

Kalman gain / foreground adaptation time.

Default: 0.002

Suggested values: 10.0, 20.0, 50.0, 0.1, 0.05, 0.01, 0.005, 0.001

Restriction: 0.0 <= Gain1

Gain2Gain2Gain2gain2gain_2 (input_control) real → (real)

Kalman gain / background adaptation time.

Default: 0.02

Suggested values: 2.0, 4.0, 8.0, 0.5, 0.1, 0.05, 0.01

Restriction: 0.0 <= Gain2

AdaptModeAdaptModeAdaptModeadaptModeadapt_mode (input_control) string → (string)

Threshold adaptation.

Default: 'on' "on" "on" "on" "on"

List of values: 'off'"off""off""off""off", 'on'"on""on""on""on"

MinDiffMinDiffMinDiffminDiffmin_diff (input_control) real → (real)

Foreground/background threshold.

Default: 7.0

Suggested values: 3.0, 5.0, 7.0, 9.0, 11.0

Recommended increment: 0.2

StatNumStatNumStatNumstatNumstat_num (input_control) integer → (integer)

Number of statistic data sets.

Default: 10

Suggested values: 5, 10, 20, 30

Value range: 1 ≤ StatNum StatNum StatNum statNum stat_num

Recommended increment: 5

ConfidenceCConfidenceCConfidenceCconfidenceCconfidence_c (input_control) real → (real)

Confidence constant.

Default: 3.25

Suggested values: 4.30, 3.25, 2.82, 2.62

Recommended increment: 0.01

Restriction: 0.0 < ConfidenceC

TimeCTimeCTimeCtimeCtime_c (input_control) real → (real)

Constant for decay time.

Default: 15.0

Suggested values: 10.0, 15.0, 20.0

Recommended increment: 5.0

Restriction: 0.0 < TimeC

BgEstiHandleBgEstiHandleBgEstiHandlebgEstiHandlebg_esti_handle (output_control) bg_estimation → (handle)

ID of the BgEsti data set.

Example (HDevelop)

* read Init-Image:
read_image (InitImage, 'xing/init')
* initialize 1. BgEsti-Dataset with
* fixed gains and threshold adaption:
create_bg_esti(InitImage,0.7,0.7,'fixed',0.002,0.02, \
               'on',7.0,10,3.25,15.0,BgEstiHandle1)
* initialize 2. BgEsti-Dataset with
* frame orientated gains and fixed threshold
create_bg_esti(InitImage,0.7,0.7,'frame',30.0,4.0, \
               'off',9.0,10,3.25,15.0,BgEstiHandle2)

Result

create_bg_esticreate_bg_estiCreateBgEstiCreateBgEsticreate_bg_esti returns 2 ( H_MSG_TRUE) if all parameters are correct.

Possible Successors

run_bg_estirun_bg_estiRunBgEstiRunBgEstirun_bg_esti

Module

Foundation

Operators