VX6953CB datasheet(70/107 Pages) STMICROELECTRONICS

70 / 107 page

EDOF control

VX6953CB

70/107

8209023 Rev 5

The plot in Figure 30 shows the EDoF main principle:

Figure 30. EDoF main principle

8.1

EDoF capabilities

The EDoF reconstruction engine is applied to the image after defect correction. It has

denoising, sharpening, deconvolution algorithms among others that allow recovering the

resolution.

The end user should not use the EDoF IP as a sharpness block. EDoF has a depth of field

recovery role. Oversharpening the image using EDoF before the HOST image processing

might alter the final image quality.

The EDoF reconstruction engine can improve the depth of field for a coloured image in the

range of 30 cm to infinity and for a monochrome image below 30 cm.

The EDoF block works with ranges of distances for which the correction applied will be

different:

•

5 Mpixel EDoF camera:

–

Super Macro mode: < 30cm

–

Macro mode: 30cm to 59cm

–

Portrait: 60cm to 249cm

–

Landscape: 250cm to infinity

The distance estimation analyzes gradients and relative sharpness among channels.

It first returns a measure of the sharpest channel in the image giving a macro ‘vote’ when

the sharpest channel is the blue one, portrait when it’s the green and landscape when it’s

the red.

The relative sharpness among channel varies with the object distance but also in the field of

the image (field curvature), so the votes are then weighted by calibration data in order to

obtain the most appropriate mode or distance at the output.

Note:

The aim of the distance estimation is not to obtain a precise distance but to select the best

parameterization for the correction to be applied in capture, therefore the returned distance

or specified distances range is indicative only.

VX6953CB датащи(PDF) 70 Page - STMicroelectronics