CN104079848A - Numeric field L-type pixel binning method - Google Patents

Abstract

The present invention relates to a kind of binning method of L-type pixel in digital domain, comprising the following steps: when selecting L-type pixel binning mode, every three pixels are synthesized into one pixel output; the selection mode of three pixels is in the original In some 2*2binning methods, one pixel is removed, and the remaining three pixels are synthesized into one pixel as an output. The removed pixel is a pixel at any position in the upper left, lower left, upper right, or lower right. The positions of the pixels removed by the binning method are relatively consistent. The binning method of the L-type pixel in the digital domain of the present invention constructs the L-type pixel in the digital domain to perform binning mode sampling to achieve higher resolution of the pixel.

Description

A kind of binning method of numeric field L-type pixel

Technical field

The invention belongs to electronics for imaging, computer vision, technical field of image processing, be specifically related to a kind of binning method of numeric field L-type pixel.

Background technology

Large area array CCD requires the occasion of high-resolution, the high image quality demand that has a wide range of applications at aeroplane photography, digital photo studio, medical treatment biochemistry, defence and military etc.The working frame frequency of area array CCD detector not only depends on the operating frequency that CCD electric charge shifts, and is also limited to total pixel number of ccd detector. large area array CCD can not meet the application of some high frame frequency.Conventionally improve frame frequency and can take two kinds of methods: improve CCD pixel clock or divide a plurality of passages to export ccd signal simultaneously.Modern area array CCD camera just towards large visual field, high-resolution future development, corresponding high frame frequency, large area array CCD are also produced gradually, application comprises full frame type area array CCD and frame transfer type area array CCD more widely.Full frame type area array CCD highest resolution can reach 4k * 5k, but frame frequency only has 4 frame left and right per second; When frame transfer type area array CCD resolution is 1k * 1k, frame frequency can reach 30 frames per second, and therefore, the view data rate of area array CCD camera output is very high.Visible high frame frequency is a conflicting body with large face battle array, thereby the binning pattern of conventional camera reduces pixel number, obtains higher frame frequency.

Binning of the prior art is a kind of image readout mode, and the electric charge of adjacent picture elements induction is added together, and with the pattern of a pixel, reads.Binning is divided into horizontal direction Binning and vertical direction Binning, horizontal direction Binning is added together the electric charge of adjacent row to read, and vertical direction Binning is added together the electric charge of adjacent row to read, the advantage of this technology of Binning is several pixels can be joined together to use as a pixel, improve sensitivity, output speed, reduce resolution, when row and column adopts Binning simultaneously, the aspect ratio of image does not change, when adopting 2:2Binning, the resolution of image will reduce 75%.

Summary of the invention

The present invention will solve the existing frame frequency of area array CCD of the prior art and improve with face battle array and increase afoul technical problem, proposes a kind of details of maintenance image as much as possible when can improve frame frequency, the binning method of numeric field L-type pixel.

In order to solve the problems of the technologies described above, technical scheme of the present invention is specific as follows:

A binning method for numeric field L-type pixel, comprises the following steps:

When selecting L-type pixel binning mode, every three pixels synthesize a pixel output; The mode of choosing of three pixels is under original 2*2binning mode, to remove a pixel, three remaining pixels synthesize a pixel as output, the pixel removing is the pixel of any one position in upper left, lower-left, upper right or bottom right, and the pixel position that every kind of binning mode is removed is relatively consistent.

The present invention has following beneficial effect:

The binning method of numeric field L-type pixel of the present invention, the method for carrying out the sampling of binning pattern at numeric field structure L-type pixel realizes higher pixel resolution capability.Adopt numeric field L-type pixel binning mode of the present invention, three pixels are exported as a pixel, data transfer rate is identical with traditional approach, but can find out by MTF curve, the binning method of the numeric field L-type pixel that the present invention proposes has higher image resolution ability at nyquist frequency place, and cut-off frequency place, the image resolution ratio that the image ratio tradition binning mode that numeric field L-type pixel binning mode obtains obtains is higher, and picture quality is better.

Accompanying drawing explanation

Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.

Fig. 1 is the schematic diagram of the binning method of numeric field L-type pixel of the present invention.

Fig. 2 is the contrast schematic diagram that the binning method of traditional binning method and numeric field L-type pixel of the present invention flies transfer function.

Embodiment

Invention thought of the present invention is:

Because pixel integration sampling process is further understood that single pixel, be sampled continuous moving in target face, obtaining the image information of pixel continuous wave output.The MTF formula of rectangle pixel imaging is:

$\mathrm{MTFpixels}=\frac{1}{p}{\∫}_{p/2}^{p/2}h\left(x\right)e^{\mathrm{tkx}}\mathrm{dx}=\frac{\sin \left(\frac{\mathrm{kp}}{2}\right)}{\frac{\mathrm{kp}}{2}}=\sin c\left(\frac{\mathrm{kp}}{2}\right)$

According to MTF basic calculating formula, the L shaped unfilled corner abnormity pixel shown in Fig. 2 is carried out to subsection integral, can obtain it at the MTFpixelL of x direction, for simplifying, adjusting integrating range is [0, P/2] ∪ [P/2, P].

$\begin{array}{c}\mathrm{MTFpixelL}\left(f\right)=\frac{1}{P^2}({\∫}_0^{P/2}\frac{P}{2}{e}^{2\mathrm{\πtfx}}\mathrm{dx}+{\∫}_{P/2}^{P}P{e}^{2\mathrm{\πtfx}}\mathrm{dx})\\ =-\frac{1(-1+e^{1\mathrm{\πPf}})(1+2e^{1\mathrm{\πPf}})}{4\mathrm{\πPf}}\end{array}$

According to Euler's formula e ^tx=cos (x)+isln (x), has can be by above formula abbreviation:

$\begin{array}{c}\mathrm{MTFL}\left(f\right)=-\frac{1(-1+e^{1\mathrm{\πPf}})(1+2e^{1\mathrm{\πPf}})}{4\mathrm{\πPf}}\\ =-\frac{1e^{1\mathrm{\πPf}/2}(-e^{-1\mathrm{\πPf}/2}+e^{1\mathrm{\πPf}/2})(1+2e^{1\mathrm{\πPf}})}{4\mathrm{f\πP}}\\ =-\frac{1e^{1\mathrm{\πPf}/2}\left(2\mathrm{iS}\ln (\mathrm{\πPf}/2)\right)(1+2e^{1\mathrm{\πPf}})}{4\mathrm{\πPf}}\\ =\frac{e^{1\mathrm{\πPf}/2}\mathrm{Sin}(\mathrm{\πPf}/2)(1+2e^{1\mathrm{\πPf}})}{4(\mathrm{\πPf}/2)}\\ =\frac{\left|e^{1\mathrm{\πPf}/2}(1+{2e}^{1\mathrm{\πPf}})\right|}{4}\mathrm{Sinc}\left(\frac{\mathrm{\πPf}}{2}\right)\end{array}$

Its curve is (together with being drawn in MTFs) as shown in Figure 2: by contrast, can find that the cut-off frequency (curve and first intersection point of x axle) of special-shaped pixel is twice than the cut-off frequency of unidimensional square pixel.

Therefore, the invention enables CMOS camera Binning pattern is that numeric field L-type pixel gathers image, contrasts traditional Binning pattern, and identical pixel number can obtain the image that spatial frequency is higher.

Binning is a kind of image playback mode, and the electric charge that is about to respond in ccd image sensor adjacent picture elements is added together, and with the pattern of a pixel, reads, and binning mode can improve sensitivity and output speed, but has reduced resolution.Traditional binning mode is divided into vertical direction and horizontal direction, vertical direction binning is added together the electric charge of adjacent column to read, and horizontal direction binning is added together the pixel charge inducing of adjacent lines to read, because each pixel of cmos sensor has an amplifier that electric charge is converted into electronic signal, therefore cmos sensor can carry out signal amplification on each pixel basis, therefore output be take pixel as unit, data are used more flexible, make the binning mode of L-type on hardware, be easy to realize.

Below in conjunction with accompanying drawing, the present invention is described in detail.

The binning method of numeric field L-type pixel of the present invention, shown in Fig. 1, when selecting L-type pixel binning mode, every three pixels synthesize a pixel output; The mode of choosing of three pixels is under original 2*2binning mode, to remove a pixel, three remaining pixels synthesize a pixel as output, the pixel removing can be the pixel of upper left, lower-left, upper right or any one position, bottom right, but the pixel position that every kind of binning mode is removed is relatively consistent.The nyquist frequency of the synthetic pixel of four pixels of binning image ratio obtaining is higher, and image resolution ability is better.And output frame frequency is consistent with conventional method.

When adopting traditional binning mode, when row and column all adopts binning mode, data transfer rate reduces to 25%, four pixels are exported as a pixel, if and adopt numeric field L-type pixel binning mode of the present invention, three pixels are exported as a pixel, data transfer rate is identical with traditional approach, but can find out by MTF curve, the binning mode of the numeric field L-type pixel that the present invention proposes has higher image resolution ability at nyquist frequency place, and cut-off frequency place, the image resolution ratio that the image ratio tradition binning mode that numeric field L-type pixel binning mode obtains obtains is higher, picture quality is better.

Obviously, above-described embodiment is only for example is clearly described, and the not restriction to execution mode.For those of ordinary skill in the field, can also make other changes in different forms on the basis of the above description.Here exhaustive without also giving all execution modes.And the apparent variation of being extended out thus or change are still among the protection range in the invention.

Claims (1)

1. a binning method for numeric field L-type pixel, is characterized in that, comprises the following steps:

When selecting L-type pixel binning mode, every three pixels synthesize a pixel output; The mode of choosing of three pixels is under original 2*2binning mode, to remove a pixel, three remaining pixels synthesize a pixel as output, the pixel removing is the pixel of any one position in upper left, lower-left, upper right or bottom right, and the pixel position that every kind of binning mode is removed is relatively consistent.