CN201393267Y - High dynamic range image acquisition device - Google Patents

Abstract

The utility model provides a high dynamic range image acquisition device comprising an image collecting module, an embedded processor and a display module. The image collecting module is used for acquiring at least one picture; the embedded processor is electrically connected with the image collecting module and synthesizes a high dynamic range image according to the picture; and the display module is electrically connected with the embedded processor and displays the high dynamic range image. The high dynamic range image acquisition device can directly acquire and synthesize the high dynamic range image by utilizing the high performance embedded processor, thereby fully exerting the calculation and processing capability of the embedded processor so as to directly, rapidly and conveniently acquire the high dynamic range image, therefore, a user can see the effect of the acquired high dynamic image while directly shooting a picture.

Description

The high dynamic range images deriving means

Technical field

The utility model relates to a kind of image acquiring device, and particularly relevant for a kind of high dynamic range images deriving means.

Background technology

A lot of people use camera, and such cognition may once be arranged: when people's backlight stood by the window, the people's face that comes of breaking forth was more black, if want people's face photograph clearly, scenery outside window again may be over-exposed.This be because the illumination dynamic range of actual environment greater than the dynamic range of camera.

Under many circumstances, among the width of cloth figure, the radiant light intensity rate of the bright dark-part of scene was greater than 100000: 1.Amount to dynamic range＞100db.And the dynamic range of common camera has only 60db, can not cover the dynamic range of whole scene, therefore, image of breaking forth or the part overexposure that some is bright or the part that some is dark are dark excessively, these situations all make the missing image details, can't fully reduce picture material further to analyze.

Another example of high dynamic range is the part of existing sunlight direct projection in the scene, during also hypographous part.At this moment the contrast of image light and shade part is also very big.

Eyes are windows that people experience the world, from obtaining the angle of external image, are equivalent to a video camera.For human eye, because of it has local autoregulation, the dynamic range of acquisition is more taller than common camera system.So the scenery that the people directly sees is abundanter than ordinary camera picture shot, has more details and abundanter color.Whether can use digital camera, obtain an energy and the similar picture of people's visual experience, be that a lot of camera design persons and user improve a problem of being concerned about simultaneously in photo resolution.

(not only useful for the picture that obtains the optimum visual effect of a width of cloth, to Medical Image Processing, and other proximate analysis has very big use for High Dynamic Range, image HDR) to obtain high dynamic range.In medical science and industrial picture control, also often run into the occasion of high dynamic range.Synthetic for high dynamic range images, some mechanism has begun research, but mainly at present still concentrates on computers synthetic.Present processing mainly is to take a picture group by camera earlier, synthesizes on computers by algorithm then, exports a high dynamic range images through reduction.For obtaining high dynamic range images, some companies have designed some special cameras, the increase that brings cost for this reason.

The utility model content

The purpose of this utility model is to provide a kind of high dynamic range images deriving means, to improve the disappearance of prior art.

For reaching above-mentioned purpose, the utility model proposes a kind of high dynamic range images deriving means and comprise image capture module, flush bonding processor and display module.Image capture module is used to capture at least one pictures; Flush bonding processor electrically connects image capture module, according to the synthetic high dynamic range images of picture; And display module electrically connects flush bonding processor, the demonstration high dynamic range images.

High dynamic range images deriving means of the present utility model utilizes high-performance embedded processor, directly carry out obtaining of high dynamic range images with synthetic, can give full play to the calculating and the disposal ability of flush bonding processor, make obtaining of high dynamic range images more directly quick, make the user directly when taking, see the effect of shooting.

For above-mentioned and other purpose, feature and advantage of the present utility model can be become apparent, preferred embodiment cited below particularly, and conjunction with figs. are described in detail below.

Description of drawings

Figure 1 shows that schematic diagram according to the high dynamic range images deriving means of the utility model one embodiment.

Embodiment

Figure 1 shows that schematic diagram according to the high dynamic range images deriving means of the utility model one embodiment.As shown in Figure 1, high dynamic range images deriving means 1 comprises image capture module 11, flush bonding processor 12, display module 13 and memory 14.

The picture of the different factors of image capture module many Same Scene of 11 acquisitions.Flush bonding processor 12 is by electrically connecting image capture module 11, according to the synthetic high dynamic range images of picture.Display module 13 electrically connects flush bonding processor 12, shows high dynamic range images.

In the present embodiment, high dynamic range images deriving means 1 more comprises memory 14 and touch key-press 15.Holder 14 is used to store the picture of the different factors of many Same Scene and synthetic high dynamic range images.Shooting instruction, processing instruction, idsplay order and system that touch key-press 15 is used for input picture acquisition module 11 are provided with instruction.And the image capture module in the present embodiment 11 is charge-coupled device camera, charge coupled device camera, CMOS (Complementary Metal Oxide Semiconductor) camera or CMOS (Complementary Metal Oxide Semiconductor) video camera, display module 13 is a touch control type LCD display screen, yet the utility model is not limited to this.

Further, in the present embodiment, flush bonding processor 12 is

533 processors.

Processor is that a class aims at the calculation requirement that satisfies current embedded audio, video and communications applications and power constraints condition and novel 16～32 flush bonding processors of designing.

Processor is based on the micro-signal framework of being developed jointly by ADI and Intel Company (MSA), and it has been combined in 32 risc type instruction set and two 16 multiplication (MAC) signal processing function that adds up with the ease for use that universal microcontroller had.The combination of this processing feature makes

Processor can all be brought into play excellent effect in signal processing and control and treatment application---and exempted the needs of setting up independent foreign peoples's processor in many occasions, this ability has greatly been simplified hardware and software design realization task.At present,

Processor can provide the performance up to 756MHz in the nucleus product in single. Novel symmetric multiprocessor member in the processor family has realized being doubled of performance under identical frequency condition.

Processor family also provides the leading power consumption performance that is low to moderate 0.8V.For satisfying the current and following signal processing applications (comprise broadband wireless, have the internet appliance and the mobile communication of audio/video function), the combination of this high-performance and low-power consumption is absolutely necessary.Wherein,

The 533rd, ADI company

A high-performance video frequency processing chip in the series, the high energy of its dominant frequency reaches 600MHz, and per second can be handled multiply-add operation 1200M time, has in a large number at the special instructions of image, can many instructions of parallel processing.

In the present embodiment, the picture of the different factors of image capture module 11 output Same Scene via the PPI interface enter flush bonding processor 12 (

533 processors).Flush bonding processor 12 (

533 processors) according to the picture of the different factors of above-mentioned Same Scene can obtain image capture module 11 the camera response curve (Camera Response Function, CRF).The camera response curve has reflected the exposure of each point in the gray value of image and the scene and the relation of illumination.

The exposure of each point is designated as X=E*T, (1)

It is the product of illumination E and time for exposure T.

The gray value of image and the relation of exposure, available following formula is represented:

Z _ij＝f(E _iΔt _j) (2)

Wherein, Zij is expressed as the gray value of image.

Formula (2) promptly is the CRF curve, and it is the intrinsic characteristic curve of each camera, and is non-line line function.

Debevec has proposed a kind of method of recovering the camera response curve, and it only need be clapped a picture group sheet to Same Scene and take a picture group sheet under difference exposure ratio.By view data is analyzed, can recover corresponding response curve, simple to environmental requirement, almost can under any scene, recover the CRF curve.In the present embodiment, flush bonding processor 12 (

533 processors) will adopt the method for Debevec to calculate the CRF curve.

According to the light sensitivity principles of image collecting device 11, f can think a monotonically increasing function, so f has invertible function f ^-1, (2) formula is rewritten as

f ^-1(Z _ij)＝E _iΔt _j (3)

Take the logarithm in both sides:

lnf ^-1(Z _ij)＝lnE _i+lnΔt _j (4)

Make g=lnf ^-1(Z _Ij), following formula can be noted by abridging and is:

g(Z _ij)＝lnE _i+lnΔt _j (5)

Wherein, i represents each point, and j represents the exposure of each picture.

When recovering the CRF curve, situation is Z _IjWith Δ t _jBe known, illumination Ei and function are unknown.We from the square error angle with smallest by formula (3) recover illumination Ei value and CRF function g (Z).

In fact can convert problem to and solve an equation, make following quadratic term equation get minimum value.

$O=\underset{i}{\overset{N}{\mathrm{\Σ}}}\underset{j}{\overset{P}{\mathrm{\Σ}}}{[g\left(Z_{\mathrm{ij}}\right)-\ln E_i-\ln \mathrm{\Δ}{t}_j]}^2+\mathrm{\λ}\underset{z=Z_{\min }+1}{\overset{Z_{\max }-1}{\mathrm{\Σ}}}{g}^{\′\′}{\left(z\right)}^2---\left(6\right)$

First of formula makes and separates the requirement of satisfying the variance minimum, and second quadratic sum minimum of asking the secondary inverse of g makes curve satisfy level and smooth requirement.By separating the equation group that (4) formula is determined, can obtain g (Z) function.

The value of noticing Z is limited, when recovering the CRF curve, as long as recover g (Z) value of limited point, as g (0)～g (255).

After the camera response curve was determined, the illumination of image can be recovered by following formula,

lnE _i＝g(Z _ij)-lnΔt _j (7)

Wherein Ei is the intensity of illumination of corresponding points in the scene that will recover.Z _IjBe the pixel value under certain time for exposure, Δ t _jBe the time for exposure.

Consider Zij and the Δ t of every width of cloth figure _jMay comprise some sum of errors noises in the value, also can make a weighted average to the Ei value that every width of cloth figure recovers and recover the lnEi value, the result who obtains like this has better signal to noise ratio.

$\ln E_i=\frac{{\mathrm{\Σ}}_{j=1}^{P}w\left(Z_{\mathrm{ij}}\right)(g\left(Z_{\mathrm{ij}}\right)-\ln \mathrm{\Δ}{t}_j)}{{\mathrm{\Σ}}_{j=1}^{P}w\left(Z_{\mathrm{ij}}\right)}---\left(8\right)$

Like this flush bonding processor 12 (

533 processors) can recover the Ei value of each point, obtain high dynamic range images.

For high dynamic range images, obtain the Ei value of image after, if directly deliver on the regular display, the effect of demonstration is usually bad.If simply adopt the method for Linear Mapping, entire image contrast of display degree is poor, looks like that washing is the same excessively, in order on general display, to show high dynamic range images preferably, flush bonding processor 12 ( 533 processors) must carry out tone mapping to high dynamic range images and handle (Tone Mapping), make it on regular display, realize demonstration near human eye perceives.At this moment, the image that obtains is the high dynamic range images that can show on it in common display.

High dynamic range images deriving means of the present utility model utilizes high-performance embedded processor, directly carry out obtaining of high dynamic range images with synthetic, can give full play to the calculating and the disposal ability of flush bonding processor, make obtaining of high dynamic range images more directly quick, make the user directly when taking, see the effect of shooting.

Though the utility model discloses as above with concrete preferred embodiment; right its is not in order to limit the utility model; anyly have the knack of this skill person; in not breaking away from spirit and scope of the present utility model; still can do a little change and retouching, therefore protection range of the present utility model is as the criterion when looking claims person of defining.

Claims (6)

1. high dynamic range images deriving means is characterized in that comprising:

Image capture module captures at least one pictures;

Flush bonding processor electrically connects above-mentioned image capture module, according to the synthetic high dynamic range images of above-mentioned picture; And

Display module electrically connects above-mentioned flush bonding processor, shows above-mentioned high dynamic range images.

2. high dynamic range images deriving means according to claim 1, it is characterized in that above-mentioned high dynamic range images deriving means more comprises storage device, electrically connect above-mentioned image capture module and above-mentioned flush bonding processor, and said storage unit is used to store above-mentioned picture and above-mentioned high dynamic range images.

3. high dynamic range images deriving means according to claim 1, it is characterized in that above-mentioned high dynamic range images deriving means more comprises touch key-press, electrically connect above-mentioned flush bonding processor, and shooting order, processing instruction, idsplay order and system that above-mentioned touch key-press is used to import above-mentioned image capture module are provided with instruction.

4. high dynamic range images deriving means according to claim 1 is characterized in that wherein above-mentioned image capture module is charge-coupled device camera, charge coupled device camera, CMOS (Complementary Metal Oxide Semiconductor) camera or CMOS (Complementary Metal Oxide Semiconductor) video camera.

5. high dynamic range images deriving means according to claim 1 is characterized in that wherein above-mentioned display module is a LCDs.

6. high dynamic range images deriving means according to claim 5 is characterized in that wherein above-mentioned LCDs is the touch control type LCD display screen.