HDR Class Reference

#include <peak_icv/algorithms/hdr/peak_icv_hdr.hpp>

Inheritance diagram for HDR:

Public Member Functions
	HDR ()
	Constructs an HDR processing object.
	~HDR () override
HDRAlgorithm	GetAlgorithm () const
	Returns the HDR algorithm used by this object.
void	EstimateResponseCurve (const std::vector< Image > &images) const
	Estimates the Camera Response Curve (CRC) from a set of input images.
Image	Process (const std::vector< Image > &images) const
	Generates an HDR image from the specified input images.
ResponseCurve	GetResponseCurve () const
void	SetResponseCurve (const ResponseCurve &responseCurve)

Detailed Description

Performs High Dynamic Range (HDR) image reconstruction.

This class provides functionality to generate a High Dynamic Range (HDR) image from a set of input images captured with different exposure settings. The input images are expected to depict the same scene while varying exactly one parameter: either the exposure time OR the gain.

Internally, the HDR reconstruction is performed using the Debevec algorithm. A key part of this process is determining the Camera Response Curve (CRC), which maps digital pixel values to relative radiance.

Typical usage

• Create an HDR object
• Optionally calculating the response curve using EstimateResponseCurve
• Generate the HDR image using Process

Note

All input images must have the same resolution, the same pixel format and being captured from an identical viewpoint of the same scene without any change in camera position or scene content. Any mismatch may lead to invalid HDR reconstruction results.

Limitations

• All input images must be compatible (same resolution, pixel format, and scene)
• At least tow input images must be provided
• Only the Debevec HDR algorithm is currently supported

Since

ids_peak_icv 1.2

Constructor & Destructor Documentation

HDR()

HDR ( )

inlineexplicit

Constructs an HDR processing object.

Creates and initializes an HDR object using the default HDR reconstruction algorithm (Debevec).

Since

ids_peak_icv 1.2

~HDR()

~HDR ( )

inlineoverride

Member Function Documentation

EstimateResponseCurve()

void EstimateResponseCurve ( const std::vector< Image > & images ) const

inline

Estimates the Camera Response Curve (CRC) from a set of input images.

This method calculates the response characteristics of the camera sensor. The CRC defines the relationship between scene radiance and the measured pixel values.

All input images must meet the following conditions:

• Represent the same scene.
• Be compatible in size and pixel format.
• Contain valid metadata for either exposure time or gain, and the values for the chosen parameter must vary across the given images (e.g., peak::common::MetadataKey::DeviceExposureTime or peak::common::MetadataKey::DeviceGain).

Performing this estimation step explicitly can significantly reduce the processing time of subsequent calls to Process(), as the curve does not need to be re-calculated.

If this method is not called, the response curve is estimated automatically during the first call to Process().

Supported pixel formats

Mono8 RGB8

See the Pixel and Pixel Format Specifications for a detailed description of the pixel formats.

Parameters

[in]

images

Images of the same scene with different exposure time or gain metadata. Must not be empty.

Since

ids_peak_icv 1.3

GetAlgorithm()

HDRAlgorithm GetAlgorithm ( ) const

inline

Returns the HDR algorithm used by this object.

The algorithm is defined at construction time and determines how the input images are combined during HDR reconstruction.

Returns

The HDR algorithm identifier.

Since

ids_peak_icv 1.2

GetResponseCurve()

ResponseCurve GetResponseCurve ( ) const

inline

Exceptions

NotPossibleException

No response curve is currently set. Use HDR::SetResponseCurve() or HDR::EstimateResponseCurve() to set a response curve first.

Since

ids_peak_icv 1.4

Process()

Image Process ( const std::vector< Image > & images ) const

inline

Generates an HDR image from the specified input images.

Processes the provided images and combines them into a single HDR image.

If a pixel is under- or overexposed across all input images this pixel is clamped to the HDR image's global minimum or maximum values, respectively. All clamped pixels are excluded from the image region.

All input images must meet the following conditions:

• Represent the same scene.
• Be compatible in size and pixel format.
• Contain valid metadata for either exposure time or gain, and the values for the chosen parameter must vary across the given images (e.g., peak::common::MetadataKey::DeviceExposureTime or peak::common::MetadataKey::DeviceGain).

If the camera response curve has not been explicitly estimated using EstimateResponseCurve, an internal estimation step is performed automatically when it is first called. This may increase the overall processing time for the first call.

Supported pixel formats

Mono8 RGB8

See the Pixel and Pixel Format Specifications for a detailed description of the pixel formats.

Parameters

[in]

images

Images of the same scene with different exposure time or gain metadata. Must not be empty.

Returns

The resulting HDR image with pixel format Mono32f resp. RGB32f.

Since

ids_peak_icv 1.2

SetResponseCurve()

void SetResponseCurve ( const ResponseCurve & responseCurve )

inline

Since

ids_peak_icv 1.4