tcConv5x5 Class Reference

#include <Conv5x5.h>

Collaboration diagram for tcConv5x5:

Public Types
enum	teOperatingMode { eeMonochrome = 0, eeBayerModeGreenFirst = 1, eeBayerModeGreenSecond = 2 }
enum	teWeights {   eeMono_Horizontal = 0, eeMono_Vertical = 1, eeGreen_Horizontal = 2, eeGreen_Vertical = 3,   eeRedBlue_Horizontal = 4, eeRedBlue_Vertical = 5 }

Public Member Functions
	tcConv5x5 (int anAddress)
virtual	~tcConv5x5 ()
bool	initialized ()
void	enable (bool abEnable)
void	setOperatingMode (teOperatingMode aeMode)
int	setWeight (teWeights aeWeights, int index, float weight)
int	setWeights (teWeights aeWeights, const std::vector< float > &avWeights)
int	setPassThroughWeights (void)
int	setGaussianWeightsMono (float sigmaX, float sigmaY)
int	setGaussianWeightsMono (float sigmaXY)
int	setGaussianWeightsBayer (float sigmaX_G, float sigmaY_G, float sigmaX_RB, float sigmaY_RB)
int	setGaussianWeightsBayer (float sigmaXY_G, float sigmal_XY_RB)
int	setGaussianWeightsBayer (float sigma)

Static Public Member Functions
static std::vector< float >	Gaussian (float std_dev, teWeights weight_type)

Protected Attributes
tcFPGARegister< uint32_t >	mcReg

Detailed Description

The tcConv5x5 class manages the conv_5x5 MityCAM IP core, which performs a 5x5 convolution using an X/Y separable kernel. Note the weights must be between 0.0f and 1.9f.

Member Enumeration Documentation

teOperatingMode

enum tcConv5x5::teOperatingMode

Weight application mode

Enumerator
eeMonochrome
eeBayerModeGreenFirst
eeBayerModeGreenSecond

teWeights

enum tcConv5x5::teWeights

Weight selection enum

Enumerator
eeMono_Horizontal
eeMono_Vertical
eeGreen_Horizontal
eeGreen_Vertical
eeRedBlue_Horizontal
eeRedBlue_Vertical

Constructor & Destructor Documentation

tcConv5x5()

tcConv5x5::tcConv5x5

(

int

anBaseAddress

)

Class constructor.

Parameters

anBaseAddress

Physical Address of the core, typically in the 0xFF20xxxx range

~tcConv5x5()

tcConv5x5::~tcConv5x5 ( )

virtual

Class Destructor

Member Function Documentation

enable()

void tcConv5x5::enable

(

bool

abEnable

)

Enable the core for operation.

Parameters

abEnable

true to enable the core, false to disable the core

Gaussian()

std::vector< float > tcConv5x5::Gaussian ( float  std_dev, teWeights  weight_type  )

static

Compute 5 point gaussian weights for given standard of deviation

Note

this routine will normalize the weights for unity gain

Parameters

std_dev	the standard of deviation
weight_type	the type of weights being computed

Returns

the 5 vector weights.

initialized()

bool tcConv5x5::initialized

(

)

Returns

true when unit is correctly initialized

setGaussianWeightsBayer() [1/3]

int tcConv5x5::setGaussianWeightsBayer ( float  sigma )

inline

setGaussianWeightsBayer() [2/3]

int tcConv5x5::setGaussianWeightsBayer	(	float	sigmaX_G,
		float	sigmaY_G,
		float	sigmaX_RB,
		float	sigmaY_RB
	)

Setup Gaussian shaded weights for Bayer mode

Parameters

sigmaX_G	guassian standard deviation in X dimension, green
sigmaY_G	guassian standard deviation in Y dimension, green
sigmaX_RB	guassian standard deviation in X dimension, red/blue
sigmaY_RB	guassian standard deviation in Y dimension, red/blue

Returns

non-zero on error

setGaussianWeightsBayer() [3/3]

int tcConv5x5::setGaussianWeightsBayer ( float  sigmaXY_G, float  sigmal_XY_RB  )

inline

setGaussianWeightsMono() [1/2]

int tcConv5x5::setGaussianWeightsMono	(	float	sigmaX,
		float	sigmaY
	)

Setup Gaussian shaded weights for Monochrome mode

Parameters

sigmaX	guassian standard deviation in X dimension
sigmaY	guassian standard deviation in Y dimension

Returns

non-zero on error

setGaussianWeightsMono() [2/2]

int tcConv5x5::setGaussianWeightsMono ( float  sigmaXY )

inline

setOperatingMode()

void tcConv5x5::setOperatingMode

(

teOperatingMode

aeMode

)

Sets the weight application mode of the 5x5 convolution.

Parameters

aeMode

the operating mode

setPassThroughWeights()

int tcConv5x5::setPassThroughWeights

(

void

)

Configure pass-through weights on all channels / modes.

Returns

non-zero on error.

setWeight()

int tcConv5x5::setWeight	(	teWeights	aeWeights,
		int	index,
		float	weight
	)

Set the array of 5 specified weights.

Parameters

aeWeights	the weight type / vector
index	the index (0-4) to set
Weights	array of 5 floating point weights.

setWeights()

int tcConv5x5::setWeights	(	teWeights	aeWeights,
		const std::vector< float > &	avWeights
	)

Set the array of 5 specified weights.

Parameters

aeWeights	the weight type / vector
avWeights	array of 5 floating point weights.

Member Data Documentation

mcReg

tcFPGARegister<uint32_t> tcConv5x5::mcReg

protected

---

The documentation for this class was generated from the following files:

• /tmp/mitycam_sdk/ARM/camera_software/src/Utility/Conv5x5.h
• /tmp/mitycam_sdk/ARM/camera_software/src/Utility/Conv5x5.cpp