AP0201AT datasheet(13/31 Pages) ONSEMI

13 / 31 page

AP0201AT

www.onsemi.com

IMAGE FLOW PROCESSOR

Image and color processing in the AP0201AT is

implemented as an image flow processor (IFP) coded in

hardware logic. During normal operation, the embedded

microcontroller will automatically adjust the operating

parameters. For normal operation of the AP0201AT, streams

of raw image data from the attached image sensor are fed

into the color pipeline. The AP201AT also has the option to

select a number of test patterns to be input instead of sensor

data.

Defect Correction

Image stream processing commences with the defect

correction function immediately after data decompanding.

To obtain defect free images, the pixels marked defective

during sensor readout and the pixels determined defective

by the defect correction algorithms are replaced with values

derived from the non−defective neighboring pixels.

AdaCD (Adaptive Color Difference)

The next step in the image stream processing is noise

reduction. The AP0201AT uses a noise reduction filter

called AdaCD which focuses on removing color noise while

preserving edge details. Automotive applications require

good performance in extremely low light, even at high

temperature conditions. In these stringent conditions the

image sensor is prone to higher noise levels, and so efficient

noise reduction techniques are required to circumvent this

sensor limitation and deliver a high quality image to the user.

Black Level Substraction and Digital Gain

After noise reduction, the pixel data goes through black

level subtraction and multiplication by a programmable

digital gain. Independent color channel digital gain can be

adjusted with registers. Black level subtraction (to

compensate for sensor data pedestal) is a single value

applied to all color channels. If the black level subtraction

produces a negative result for a particular pixel, the value of

this pixel is set to 0.

Positional Gain Adjustments (PGA)

Lenses tend to produce images whose brightness is

significantly attenuated near the edges. There are also other

factors causing fixed pattern signal gradients in images

captured by image sensors. The cumulative result of all these

factors is known as image shading. The AP0201AT has an

embedded shading correction module that can be

programmed to counter the shading effects on each

individual R, Gb, Gr, and B color signal.

The Correction Function

The correction functions can then be applied to each pixel

value to equalize the response across the image as follows:

Pcorrected(row, col) + Psensor(row, col)

f(row, col)

(eq. 1)

where P are the pixel values and f is the color dependent

correction functions for each color channel.

Adaptive Local Tone Mapping (ALTM)

Real world scenes often have very high dynamic range

(HDR) that far exceeds the electrical dynamic range of the

imager. Dynamic range is defined as the luminance ratio

between the brightest and the darkest object in a scene. In

recent years many technologies have been developed to

capture the full dynamic range of real world scenes. For

example, the multiple exposure method is widely adopted

for capturing high dynamic range images, which combines

a series of low dynamic range images of the same scene

taken under different exposure times into a single HDR

image.

Even though the new digital imaging technology enables

the capture of the full dynamic range, low dynamic range

display devices are the limiting factor. Today’s typical LCD

monitor has contrast ratio around 1,000:1; this contrast ratio

is not enough for an HDR image (the contrast ratio for an

HDR image is around 250,000:1). Therefore, in order to

reproduce HDR images on a low dynamic range display

device, the captured high dynamic range must be

compressed to the available range of the display device. This

is commonly called tone mapping.

Tone mapping methods can be classified into global tone

mapping and local tone mapping. Global tone mapping

methods apply the same mapping function to all pixels.

While

global

tone

mapping

methods

provide

computationally simple and easy to use solutions, they often

cause loss of contrast and detail. A local tone mapping is thus

necessary in addition to global tone mapping for the

reproduction of visually more appealing images that also

reveal scene details that are important for automotive safety

applications. Local tone mapping methods use a spatially

variable mapping function determined by the neighborhood

of a pixel, which allows it to increase the local contrast and

the visibility of some details of the image. Local methods

usually yield more pleasing results because they exploit the

fact that human vision is more sensitive to local contrast.

ON Semiconductor’s ALTM solution significantly

improves the performance over global tone mapping.

ALTM is directly applied to the Bayer domain to compress

the dynamic range from 20−bit to 12−bit. This allows the

regular color pipeline to be used for HDR image rendering.

Color Interpolation

In the raw data stream fed by the external sensor to the IFP,

each pixel is represented by a 20− or 12−bit integer number,

which can be considered proportional to the pixel’s response

to a one−color light stimulus, red, green, or blue, depending

on the pixel’s position under the color filter array. Initial data

processing steps, up to and including ALTM, preserve the

one−color−per−pixel nature of the data stream, but after

ALTM it must be converted to a three−colors−per−pixel

stream appropriate for standard color processing. The

conversion is done by an edge−sensitive color interpolation

module. The module pads the incomplete color information

available for each pixel with information extracted from an

AP0201AT Datasheet(PDF) 13 Page - ON Semiconductor

AP0201AT Datasheet(HTML) 13 Page - ON Semiconductor

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Datasheet Download

Link URL