#include "MisrUtil.h"
#include "MisrError.h"
#include "MisrProjParam.h"
#include "MisrMapQuery.h"

Data Structures
struct	MTKt_RegressionCoeff

Macros
#define	MTKT_REGRESSION_COEFF_INIT

Functions
MTKt_status	MtkRegressionCoeffAllocate (int nline, int nsample, MTKt_RegressionCoeff *regressbuf)
	Allocate buffer to contain regression coefficients. More...

MTKt_status	MtkRegressionCoeffFree (MTKt_RegressionCoeff *regressbuf)
	Free memory for regression coefficients. More...

MTKt_status	MtkLinearRegressionCalc (int Size, const double X, const double Y, const double YSigma, double A, double B, double Correlation)
	Use linear regression to fit a set of observations (x,y) to the model: y(x) = a + b * x. The values of 'x' are assumed to be known exactly. The values of 'y' may have an associated uncertainty, 'y_sigma'. Measurements with larger uncertainty are given less weight. Uncertainty must be greater than 0. More...

MTKt_status	MtkSmoothData (const MTKt_DataBuffer Data, const MTKt_DataBuffer Valid_mask, int Width_line, int Width_sample, MTKt_DataBuffer *Data_smoothed)
	Smooth the given array with a boxcar average of the specified width. The algorithm is similar to the IDL smooth routine. Smoothed data is type float. More...

MTKt_status	MtkRegressionCoeffCalc (const MTKt_DataBuffer Data1, const MTKt_DataBuffer Valid_mask1, const MTKt_DataBuffer Data2, const MTKt_DataBuffer Data2_sigma, const MTKt_DataBuffer Valid_mask2, const MTKt_MapInfo Map_info, int Regression_size_factor, MTKt_RegressionCoeff Regression_coeff, MTKt_MapInfo Regression_map_info)
	Calculate linear regression coefficients for translating values in data buffer 1 to corresponding values in data buffer 2. More...

MTKt_status	MtkResampleRegressionCoeff (const MTKt_RegressionCoeff Regression_coeff, const MTKt_MapInfo Regression_coeff_map_info, const MTKt_MapInfo Target_map_info, MTKt_RegressionCoeff Regression_coeff_out)
	Resample regression coefficients at each pixel in the target map. Resampling is by cubic convolution. More...

MTKt_status	MtkApplyRegression (const MTKt_DataBuffer Source, const MTKt_DataBuffer Source_mask, const MTKt_MapInfo Source_map_info, const MTKt_RegressionCoeff Regression_coeff, const MTKt_MapInfo Regression_coeff_map_info, MTKt_DataBuffer Regressed, MTKt_DataBuffer *Regressed_mask)
	Apply regression to given data. Uses MtkResampleCubicConvolution to resample regression coefficients to resolution of the input data. Output data is the same size and resolution as the input data. More...

MTKt_status	MtkDownsample (const MTKt_DataBuffer Source, const MTKt_DataBuffer Source_mask, int Size_factor, MTKt_DataBuffer Result, MTKt_DataBuffer Result_mask)
	Downsample data by averaging pixels. More...

MTKt_status	MtkUpsampleMask (const MTKt_DataBuffer Source_mask, int Size_factor, MTKt_DataBuffer Result_mask)
	Upsample a mask by nearest neighbor sampling. More...

Macro Definition Documentation

◆ MTKT_REGRESSION_COEFF_INIT

#define MTKT_REGRESSION_COEFF_INIT

Value:

{ \
    MTKT_DATABUFFER_INIT, \
    MTKT_DATABUFFER_INIT, \
    MTKT_DATABUFFER_INIT, \
}

Definition at line 32 of file MisrRegression.h.

Function Documentation

◆ MtkApplyRegression()

MTKt_status MtkApplyRegression	(	const MTKt_DataBuffer *	Source,
		const MTKt_DataBuffer *	Source_mask,
		const MTKt_MapInfo *	Source_map_info,
		const MTKt_RegressionCoeff *	Regression_coeff,
		const MTKt_MapInfo *	Regression_coeff_map_info,
		MTKt_DataBuffer *	Regressed,
		MTKt_DataBuffer *	Regressed_mask
	)

Apply regression to given data. Uses MtkResampleCubicConvolution to resample regression coefficients to resolution of the input data. Output data is the same size and resolution as the input data.

Returns: MTK_SUCCESS if successful.

Example:: In this example we apply a regression based on the supplied source, source_mask, and source_map_info using regression_coeff and regression_coeff_map_info. The result is regressed and regressed_mask.

status = MtkApplyRegression(&source, &source_mask, &source_map_info, &regression_coeff, &regression_coeff_map_info, &regressed, &regressed_mask);

Note

Parameters

[in]	Source	Input data. (float)
[in]	Source_mask	Valid mask for input data. (uint8)
[in]	Source_map_info	Map info for input data
[in]	Regression_coeff	Regression coefficients
[in]	Regression_coeff_map_info	Map info for regression coefficients
[out]	Regressed	Output data. (float)
[out]	Regressed_mask	Valid mask for output data. (uint8)

Definition at line 36 of file MtkApplyRegression.c.

◆ MtkDownsample()

MTKt_status MtkDownsample	(	const MTKt_DataBuffer *	Source,
		const MTKt_DataBuffer *	Source_mask,
		int	Size_factor,
		MTKt_DataBuffer *	Result,
		MTKt_DataBuffer *	Result_mask
	)

Downsample data by averaging pixels.

Returns: MTK_SUCCESS if successful.

Example:: In this example we downsample source with source_mask by size_factor. The result is stored in result with result_mask.

status = MtkDownsample(&source, &source_mask, size_factor, &result, &result_mask);

Note

Parameters

[in]	Source	Source data. (float)
[in]	Source_mask	Valid mask for source data. (uint8)
[in]	Size_factor	Number of pixels to aggregate along each dimension.
[out]	Result	Downsampled result. (float)
[out]	Result_mask	Valid mask for result. (uint8)

Definition at line 36 of file MtkDownsample.c.

◆ MtkLinearRegressionCalc()

MTKt_status MtkLinearRegressionCalc	(	int	Size,
		const double *	X,
		const double *	Y,
		const double *	Y_Sigma,
		double *	A,
		double *	B,
		double *	Correlation
	)

Use linear regression to fit a set of observations (x,y) to the model: y(x) = a + b * x. The values of 'x' are assumed to be known exactly. The values of 'y' may have an associated uncertainty, 'y_sigma'. Measurements with larger uncertainty are given less weight. Uncertainty must be greater than 0.

Returns: MTK_SUCCESS if successful.

Example:

double x[4] = {1,5,7,8};
double y[4] = {12,14,19,21};
double y_sigma[4] = {0.1,0.2,0.2,0.3};
double a, b;
status = MtkLinearRegressionCalc(4, x, y, y_sigma, &a, &b)

Note

Parameters

[in]	Size	Size of X and Y arrays
[in]	X	X array
[in]	Y	Y array
[in]	Y_Sigma	Uncertainty in Y
[out]	A	A
[out]	B	B
[out]	Correlation	Correlation

Definition at line 39 of file MtkLinearRegressionCalc.c.

◆ MtkRegressionCoeffAllocate()

MTKt_status MtkRegressionCoeffAllocate	(	int	nline,
		int	nsample,
		MTKt_RegressionCoeff *	regressbuf
	)

Allocate buffer to contain regression coefficients.

Returns: MTK_SUCCESS if successful.

Example:: In this example, we allocate a buffer with a size of 5 lines by 10 samples of type MTKe_int16

status = MtkRegressionCoeffAllocate(5, 10, &regressbuf);

Note: The caller is responsible for using MtkRegressionCoeffFree() to free the memory used by regressbuf

Parameters

[in]	nline	Number of lines
[in]	nsample	Number of samples
[out]	regressbuf	Data Buffer

Definition at line 36 of file MtkRegressionCoeffAllocate.c.

◆ MtkRegressionCoeffCalc()

MTKt_status MtkRegressionCoeffCalc	(	const MTKt_DataBuffer *	Data1,
		const MTKt_DataBuffer *	Valid_mask1,
		const MTKt_DataBuffer *	Data2,
		const MTKt_DataBuffer *	Data2_sigma,
		const MTKt_DataBuffer *	Valid_mask2,
		const MTKt_MapInfo *	Map_info,
		int	Regression_size_factor,
		MTKt_RegressionCoeff *	Regression_coeff,
		MTKt_MapInfo *	Regression_coeff_map_info
	)

Calculate linear regression coefficients for translating values in data buffer 1 to corresponding values in data buffer 2.

Returns: MTK_SUCCESS if successful.

Example:: In this example we determine coefficients to map values in data1 with mask valid_mask1 to data2 having mask valid_mask2 and uncertainty data2_sigma. Linear regression will use pixel neighborhood size specified by regression_size_factor and output the results to regression_coeff with regression_coeff_map_info.

status = MtkRegressionCoeffCalc(&data1, &valid_mask1, &data2, &data2_sigma, &valid_mask2, &map_info, regression_size_factor, &regression_coeff, &regression_coeff_map_info);

Note

Parameters

[in]	Data1	Data buffer 1
[in]	Valid_mask1	Valid mask for data buffer 1
[in]	Data2	Data buffer 2
[in]	Data2_sigma	Uncertainty for data buffer 2
[in]	Valid_mask2	Valid mask for data buffer 2
[in]	Map_info	Map info for input data.
[in]	Regression_size_factor	Number of pixels to aggregate along each axis when generating regression coefficients.
[out]	Regression_coeff	Regression coefficients.
[out]	Regression_coeff_map_info	Map info for regression coefficients

Definition at line 36 of file MtkRegressionCoeffCalc.c.

◆ MtkRegressionCoeffFree()

MTKt_status MtkRegressionCoeffFree ( MTKt_RegressionCoeff * regressbuf )

Free memory for regression coefficients.

Returns: MTK_SUCCESS

Example:: In this example, we have a MTKt_RegressionCoeff called regressbuf that was previously allocated with MtkRegressionCoeffAllocate() and we wish to free it.

status = MtkRegressionCoeffFree(&regressbuf);

Parameters

[in,out] regressbuf Data Buffer

Definition at line 33 of file MtkRegressionCoeffFree.c.

◆ MtkResampleRegressionCoeff()

MTKt_status MtkResampleRegressionCoeff	(	const MTKt_RegressionCoeff *	Regression_coeff,
		const MTKt_MapInfo *	Regression_coeff_map_info,
		const MTKt_MapInfo *	Target_map_info,
		MTKt_RegressionCoeff *	Regression_coeff_out
	)

Resample regression coefficients at each pixel in the target map. Resampling is by cubic convolution.

Returns: MTK_SUCCESS if successful.

Example:: In this example we use cubic convolution to resample regression_coeff with regression_coeff_map_info to target_map_info. Output is at regression_coeff_out.

status = MtkResampleRegressionCoeff(&regression_coeff, &regression_coeff_map_info, &target_map_info, &regression_coeff_out);

Note: MtkRegressionCoeffCalc() can be used to generate regression_coeff and regression_coeff_map_info from 2 data sets and masks.

Parameters

[in]	Regression_coeff	Regression coefficients.
[in]	Regression_coeff_map_info	Map info for regression coefficients.
[in]	Target_map_info	Map info for target grid.
[out]	Regression_coeff_out	Regression coefficients resampled to target grid.

Definition at line 38 of file MtkResampleRegressionCoeff.c.

◆ MtkSmoothData()

MTKt_status MtkSmoothData	(	const MTKt_DataBuffer *	Data,
		const MTKt_DataBuffer *	Valid_mask,
		int	Width_line,
		int	Width_sample,
		MTKt_DataBuffer *	Data_smoothed
	)

Smooth the given array with a boxcar average of the specified width. The algorithm is similar to the IDL smooth routine. Smoothed data is type float.

Returns: MTK_SUCCESS if successful.

Example:: In this example we smooth data with mask valid_mask and smoothing window widths width_line and width_sample. Output is data_smoothed.

status = MtkSmoothData(&data, &valid_mask, width_line, width_sample, &data_smoothed);

Note

Parameters

[in]	Data	Data. (float)
[in]	Valid_mask	Mask inidicating where data is valid. (uint8)
[in]	Width_line	Width of smoothing window along line dimension. Must be and odd number.
[in]	Width_sample	Width of smoothing window along sample dimension. Must be an odd number.
[out]	Data_smoothed	Smoothed data. (float)

Definition at line 36 of file MtkSmoothData.c.

◆ MtkUpsampleMask()

MTKt_status MtkUpsampleMask	(	const MTKt_DataBuffer *	Source_mask,
		int	Size_factor,
		MTKt_DataBuffer *	Result_mask
	)

Upsample a mask by nearest neighbor sampling.

Returns: MTK_SUCCESS if successful.

Example:: In this example we upsample source_mask by a factor of size_factor. The output is result_mask.

status = MtkUpsampleMask(&source_mask, size_factor, &result_mask);

Note

Parameters

[in]	Source_mask	Source mask. (uint8)
[in]	Size_factor	Number of pixels to expand along each dimension.
[out]	Result_mask	Upsampled result. (uint8)

Definition at line 36 of file MtkUpsampleMask.c.

Data Structures

Macros

Functions

Macro Definition Documentation

◆ MTKT_REGRESSION_COEFF_INIT

Function Documentation

◆ MtkApplyRegression()

◆ MtkDownsample()

◆ MtkLinearRegressionCalc()

◆ MtkRegressionCoeffAllocate()

◆ MtkRegressionCoeffCalc()

◆ MtkRegressionCoeffFree()

◆ MtkResampleRegressionCoeff()

◆ MtkSmoothData()

◆ MtkUpsampleMask()