CN100464568C - Image processing system - Google Patents
Image processing system Download PDFInfo
- Publication number
- CN100464568C CN100464568C CNB2005101185519A CN200510118551A CN100464568C CN 100464568 C CN100464568 C CN 100464568C CN B2005101185519 A CNB2005101185519 A CN B2005101185519A CN 200510118551 A CN200510118551 A CN 200510118551A CN 100464568 C CN100464568 C CN 100464568C
- Authority
- CN
- China
- Prior art keywords
- signal
- data
- order
- data path
- control
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Color Television Image Signal Generators (AREA)
Abstract
The invention comprises: data path unit, integrated with non-linear filter unit, color space conversion unit, for use in processing image signals; geometry conversion unit; linear filter unit; gain control unit for processing a original image signal and outputting a processed signal; and control and estimation unit for receiving the original image signal, outputting the processed signal from the data path unit and generating multi adjustment signals respectively sent to the non-linear filter unit in the data path unit, color space conversion unit, geometry conversion unit, linear filter unit and gain control unit.
Description
Technical field
The present invention is relevant for image processing system, it is used the digital signal processor (DSP) that is embodied as video camera, so that flexible data path to be provided, handles to support various digitized videos, for example, Automatic white balance (AWB), automatic exposure (AE), flaw remove, Gamma correction or the like.
Background technology
Yet the calcspar of traditional cameras system is to simplify shown in the 4th figure., actual have must have a lot of frameworks could realize camera system, for example, the selection of processing, processing sequence, the color space or the element of being handled, reach other and select function or the like for use.Simultaneously, institute's implementation method also is variation.The processing capacity of the 4th figure and order are single example, and it is just in order to illustrate the usefulness of video camera.
This camera chain mainly is made of two parts: one is the data path that is used for signal processing, and another is the non-data path in order to control. the traditional data path in camera system is combined usually or mixes with controlled function.Sometimes, each data path is controlled by a local control square, and therefore, actual camera chain is linked by little processing capacity square.In between these processing squares, not having powerful relation or relevance. this is that generation video camera IC is designed by analogy bipolar integrated circuit (IC) is finished before because traditional cameras is based on.These IC are finished by the binding with the analogy square of little function.In addition, because these separations make people understand the processing of each function easily, thereby design each standalone feature square easily.
In the 4th figure, input 60 is the initial data from image sensor. initial data is in first flaw compensation square 61, compensated with flaw location data 62. after the flaw compensation, the shade that is caused for video camera module lens is compensated with shade table data 64 at shadow compensation square 63.After these computings, initial data is transferred to RGB or YCrCb color data at conversion square 65.This processing is to be executed in these color spaces.Noise reduces square 66 for using linearity or nonlinear filtering, for example the noise of middle bandpass filter reduces, with the inhibition is the noise that image sensor and ADC etc. are caused. after noise reduces, carry out edge strengthening (or aperture compensation) at edge strengthening square 67. then, the signal after the processing is admitted to automatic exposure (AE) square 68 and proceeds to exposure via line 69 simultaneously and measures square 70.The warp as a result 72 that exposure measures square 70 is sent back to automatic exposure square 68.In the automatic exposure square, exposure control square will be controlled the gain of each colour content, form preferable image quality to adjust brightness.In simultaneously, the output that measures square 70 from exposure also is input to iris diaphgram/shutter (electronics or mechanical type) control square 71, controls to the inductor module for iris diaphgram actuator and shutter speed to produce control signal.
In one-level is Automatic white balance square 74.In this square, the color character of existing image is in color assessment square, does to measure and select the color adjustment.For example, in this square, judge that shooting is at indoor or outdoors.When outdoor, will select to be used for outdoor suitable color temperature, otherwise, if when indoor, then select another suitable color temperature.The main points of this program are the judgement that the increase of the measurement of color temperature of existing photo and each colour content changes.In order to improve applicability, also can suitably add other judgements, but this will make AWB deduction complexity to various scenes.
Output from AWB square 74 is at gamma compensation square 79, uses gamma tables 80 to be done the gamma compensation, the characteristic of this gamma tables decision gamma compensation.Sometimes, in special effects square 81, carry out special processing. after gain compensation square 82 was adjusted gain, the result 83 of video camera image processing was output to next stage.
This camera chain comprises system controller 84 usually, and it is realized by CPU/DSP, and is used to control the operation of whole system, comprises to open that the beginning is set and the conversion and control of system pattern.The part that measures and judge can be carried out on this system controller 84.
On the other hand, the someone wants to realize these image processing with software program. this can make the video camera image processing more full of elasticity.Yet each program in this software is identical with the conventional hardware imaging program mostly, and it may be implemented as the binding that little routine is handled, and its each processing is corresponding to the processing that is implemented in the traditional cameras square.
Yet, in above-mentioned prior art, still have following some problems:
1. because each function is with near in addition construction of independent mode, so each function square all needs a local controlled function. the system of this type also needs to connect similar function.Therefore, needed hardware resource increases, thereby, cause the structural device complexity of traditional cameras.
2. because local control architecture, make the structure of traditional cameras be difficult to use more powerful CPU/DPS, the preface of controlling these types to. local controller is after system development, is realized by exclusive hardware, therefore, can not upgrade deduction again.Because being special design and control assembly, realized with a lot of little local controller that disperses again data path, so can not change or increase new deduction traditional cameras.
3. because data path is to be engaged to control assembly powerfully because control assembly can not the user system in or the more powerful CPU outside the system, the clumsiness and do not have elasticity so camera system becomes too.Therefore, system needs higher-order image quality and other visual performances sometimes, and another system does not then need the high-order image, and needs lower cost.Therefore, this camera system must redesign and could satisfy considering on system cost and the power-consumption limitations.
Summary of the invention
Therefore, The present invention be directed to foregoing problems, propose solution.First purpose of the present invention is for providing a kind of intelligent image processing arrangement, it has the data path hardware that elasticity is concentrated, and data path merged, so that the function square is simpler, simultaneously, controlled function is arranged on the outside, operation is simplified and the quantity of signal path is reduced in a large number so that the majority signal in this equipment is handled, thereby, realized flexible data path.So the structure of simplified control system has also improved the reliability of video sensing system.
In addition, because the control square is shifted in to outside the shooting square, so the hardware size type of shooting square is to reduce. externally the control of the brute force on processing machine permission system can be with more powerful CPU processor, use the more control of high-order. with this powerful CPU usefulness, system can realize the wisdom deduction of each shooting processing capacity.
In order to finish these purposes, the video signal processing circuit according to the embodiment of the invention comprises:
Be used to handle the data path mechanism of signal of video signal, merging to concentrate has nonlinear filtering mechanism, color space transformation mechanism, geometric transformation mechanism, linear filtering mechanism and gain controlling mechanism, to handle an original image signal, to export a processed signal; And
Control and decision mechanism, in order to receive this original image signal and the processed signal of output, produce most nonlinear filtering mechanism, color space transformation mechanism, geometric transformation mechanism, linear filtering mechanism and gain controlling mechanisms that signal is given this data path mechanism respectively that adjust from data path mechanism.
Description of drawings
The 1st figure is the flow chart of camera processor of the present invention;
The 2nd figure is data path of the present invention and control and judges schematic diagram partly;
The 3rd figure is the process chart according to another embodiment of the present invention; And
The 4th figure is the calcspar of traditional camera system.
Embodiment
The present invention is relevant for image processing. and the present invention just makes illustrative nature in this processing selecting of describing and illustrating, certainly, the present invention can be suitable for to every other image processing program of not addressing or different image processing orders. and the major function of image processing is following listed:
-flaw compensation (defect compensating)
-inlay conversion (initial data is to RGB/YCrCb)
-automatic exposure
-Automatic white balance
-focusing automatically
-edge strengthening
The compensation of-gamma
-shadow compensation
-digital zoom
-image stabilization
-local filtering comprises noise and reduces
-distortion compensation
-dynamic range is amplified
-specific function
Because these functions are normally carried out with independent discrete function square, therefore, these function squares are by combination together, to set up whole camera system.Sometimes, each function has its data path square and attached this locality control square. and usually, these processing are to finish in a special color space; Some function is to be maintained in the original data space, and some is then in the RGB color space.Simultaneously, YCrCb (YUV) also is used the color space as data processing.
Yet, behind these squares of labor, partly function can be merged or be made up, and can make an intensive simple function square. in particular, data path can be merged into a common hardware, and control assembly and decision means one of may be implemented as on processor peculiar software.
The processing square is analyzed
The function square of camera function below will be described.These are the example of camera function.Also have other a lot of squares and deduction and can realize these functions.After the expansion of each function, common trait can be summarized as follows.
The flaw compensation
Flaw (defective) pixel of compensation CCD/CMOS image pick-up device is to white value or black value, it typically is fixed data and be not related to image. usually, this flaw compensation is the digitlization initial data of handling from the CCD/CMOS image sensor. local filter is in order to replace or to suppress the flaw pixel data, to cover flaw/defective.Sometimes, system not only supports the previous defective that detects with the position data that is preloaded into of flaw, simultaneously, is also detecting flaw instantly, detection of dynamic and compensation, and compensation immediately, and on flaw location, do not use the information that stores in advance.
Inlay conversion
Be converted to the color space transformation that RGB or YCrCb bit reflect figure by initial data.Partly the short pixel data of color component is produced by filtering local data, to finish for example three primary colors color space of RGB.With this program, the quantity of data will become three times greater than initial data.Initial data can be primary colors (R, G, B) or complementary colours (G, Cy, Mg).Colour content after inlaying conversion can be RGB or YCrCb.
Automatic exposure
With the gain of control amplifier square, iris diaphgram, and shutter speed (electric or mechanical type), adjust image exposure. this program is as follows, at first, is measured corresponding to the particular value of image exposure, then, the gain of each color component is adjusted and is used for whole image.This operation is finished with a pixel one pixel-wise.
Automatic white balance
Adjust the color of the image in the various colorful light-emitting states. after the color with the colour temperature analysis image, camera system is adjusted color by the gain of adjusting each color component.This program can be finished in the RGB color space or in the YCrCb color space, and this operation is carried out with a pixel one pixel-wise usually.
Automatically focusing
Automatically the main purpose of focusing is the change with the condenser lens position, detects the acutance or the contrast of image. and the partial parameters relevant for image frequency can be in order to the actuator of controlling lens and the moving direction of judging focus. and this program only is used for the lens module.This moment, image pickup signal is not dealt with.
Edge strengthening
This is local filtering, in order to the acutance of improvement image, to obtain image more clearly.Can in RGB colour content or YCrCb colour content, make Filtering Processing.
The gamma compensation
Each color is in the mode of a pixel one pixel, carries out non-linear conversion.Can use the RGB or the YCrCb color space.
Shadow compensation
Lens have shade, and expression is near the edge of image-region, because the feature of lens, and luminous emissivity reduces. the location of pixels that depends in image is handled in shooting, and the change amplifier gain, and compensating this reduction. colour content changes by a pixel one pixel ground and increases, and correct.
Digital zoom
Change with level and vertical filter, with change image size.
Image stabilization
This program is the action of depending on video camera, and the displacement view data. wherein, finished the displacement and the filtering of data, be done to be used for interpolation.This program is the local filtering of using with interpolation, is finished in the mode of a pixel one pixel.
Noise reduces
This is local filtering, comprises nonlinear filter, for example, middle bandpass filter. this program is that finishing in order to color component is main finish and filtering is for example to finish in 3 * 3,5 * 5 or the like the selected window size.
Distortion compensation
Geometric transformation is to be adjusted into the distortion that lens characteristics is caused.This conversion is to be main being finished with the color component. geometric transformation is to finish with interpolation filtering.
Dynamic range enlarges
Each pixel value will be according to its pixel data value change.Sometimes, use the relevant or district's city relevant procedures of data, and change the gain of each pixel data.In light-emitting zone, gain is lowered and the dark areas gain is increased, with the vision quality of improvement in two zones.This is equivalent to the amplification of image dynamic range.
Special effects
The special effects function of a lot of types is arranged, for example eliminate color, pseudo-classic color or the like.In following illustration special effects, be assumed to be the processing of pixel, comprise the change color.
These processing capacities are to be summarized in the table 1.
| Handle | The color space | Operation | Data cell |
| The flaw compensation | Former data | Filtering | Window |
| Inlay conversion | Initial data → RGB/YC | Filtering | Window |
| Automatic exposure | RGB/YC | Gain | Pixel |
| Automatic white balance | RGB/YC | Gain | Pixel |
| Automatically focusing | RGB/YC | The lens module | --- |
| Edge strengthening | RGB/YC | Filtering | Window |
| The gamma compensation | RGB/YC | Non-linear | Pixel |
| Shadow compensation | RGB/YC (initial data) | Gain | Pixel |
| Digital zoom | RGB/YC | Geometric transformation+filtering | Window |
| Image stabilization | RGB/YC | Geometric transformation+filtering | Window |
| Noise reduces | RGB/YC | Filtering | Window |
| Distortion compensation | RGB/YC (initial data) | Geometric transformation+filtering | Window |
| Dynamic range enlarges | RGB/YC | Gain | Pixel |
| Special effects | RGB/YC | Gain/color | Pixel |
Point out the part common trait of everywhere reason in the table, the chromatic number of for example being handled is according to the activity classification of space, each processing, the data cell of being handled.The color space represents to be used for the colour content of each processing.In theory, each processing can be handled with arbitrary colour content, and for example: initial data, RGB and YCrCb, other possible colour contents also are feasible.The type that the operation expression is calculated.Filtering is illustrated in the local filtering in the window, and the gain expression be multiply by gain factor and changed numerical value.Geometric transformation need produce pixel data by the specific region. after this geometric transformation, can carry out filtering.Data cell is represented handled data.Filtering needs data to be set in the window.Gain adjustment needs the individual pixel data to be used for this processing.
These factors have been arranged, and significantly, the processing on camera function is divided into several classes.This type of is used to following example 1, and with the example of Display Realization system of the present invention, wherein, the same category operation is merged into co-operate, and common hardware is implemented method.
In following table 2, in addition again line description two factors, promptly one for being treated to " relevant with the zone ", another is for being treated to " relevant with scene ".
| Handle | Scene is relevant | Cutting zone is relevant |
| The flaw compensation | √ | Fixing |
| Inlay conversion | Fixing | |
| Automatic exposure | √ | √ |
| Automatic white balance | √ | √ |
| Automatically focusing | √ | --- |
| Edge strengthening | √ | √ |
| The gamma compensation | √ | |
| Shadow compensation | √ | Fixing |
| Digital zoom | ||
| Image stabilization | ||
| Noise reduces | √ | √ |
| Distortion compensation | Fixing | |
| Dynamic range enlarges | √ | √ |
| Special effects |
The relevant expression in zone is handled can relevant improved potency with processing region, this zone be for the graphical analysis of back-end processing specified.The reason operation of the relevant expression of scene everywhere is to depend on that institute's choosing location picture is finished.Yet this two factor is to be useful on shooting to handle., in traditional camera system and this two factor that is of little use, because handle the very difficult event of the relevant processing of scene analysis and cutting zone. the present invention provides the possibility of these high-order complex process under the assistance of rear end high performance processor.Partly scene relevant treatment and regional relevant treatment example are to be described in the following example.
Example 1
Based on above-mentioned analysis, exemplary system of the present invention is to be described as follows.First step is for explaining how system of the present invention realizes having the intelligent system that merges processing capacity, and this is shown in the 1st figure.Among the 1st figure, the flow process of camera processor is described.The upper part of the 1st figure is for should be at the handled treatment element of camera processor.These treatment elements are classified into four group.Group 1 is for carrying out nonlinear filtering, and it for example comprises, and noise reduces and the flaw compensation.Group 2 comprises digital zoom, image stabilization, distortion compensation.The required common function of these operations is geometric transformation and subsequent filtering operation. and group 3 is for example edge strengthening, and it needs local filtering person. and group 4 comprises automatic exposure, Automatic white balance, gamma compensation, dynamic range expansion, shadow compensation and special effects or the like.These are with a pixel one pixel-wise, are used for the basic operation of pixel data, are that composition is relevant sometimes, and sometimes, it is to be the operation to the color vector, but its operation relates to each pixel data, does not need filtering operation.
Based on the analysis of each processing in the table 1, these processing are the centres as the 1st figure, are categorized into several classification.This classification is that the gain that comprises geometric transformation 5, is used for the filtering 6 of window and is used for pixel changes 7.This nonlinear filtering wave group 1 is directly to reflect figure to nonlinear filtering square 9.Geometric transformation 5 is the operations of extracting geometric transformation portion from the 2nd group operation out.The filtering operation that is used for the 2nd group is drawn out of and is incorporated into the filtering operation 6 that is used for window.Also be drawn out of and be incorporated into the filtering 6 of window from the 3rd group filtering operation.Last type is to change 7 for the gain that is used for pixel.This is for a pixel one pixel ground operation to each pixel data, by for example three kinds of compute type changes of multiplication value.
The bottom 14 of the 1st figure is for being implemented as the function square of hardware.Input 8 is the digitlization initial data from image sensor. these data are at first handled at nonlinear filtering square 9.At square 10, output is changed to the RGB or the YCrCb color space by initial data, and interpolation is with unexistent pixel data on image sensor, and produce RGB or the YCrCb be used for following processing and reflect diagram data. RGB that is changed or YCrCb data at first obtain and interpolation 11 by geometric transformation, to obtain proper data and interpolation to be created in the data in the geometric transformation.Level 12 in the end, it is to be combined into an operation that filtering operation 6 and gain adjust 7, after this operation, final result is outputted as 13.
Main points of the present invention are that several co-operate and common hardware are merged in several the operations that also will so classify for each shooting treatment classification. merge with this, system of the present invention can realize intelligent system, it has how flexible operation, and littler data path hardware.Clear cutting between data path and control and arbitration functions will be described as follows, and control and will judge with powerful processor, to show the effect of native system.
The shooting processing of more than selecting is a kind of example.Classification is handled and the order of operation also is an example.Might suitably change and select these to handle and order.
Shown in the 2nd figure, show real embodiment.The upper part of the 2nd figure be for the following part of data path and the 2nd figure for control and judge partly 27, it is to be implemented as software in this case. data path architecture is identical with the 1st figure person.Input initial data 15 is input to nonlinear filtering square 16, wherein carries out flaw compensation and noise and reduces.Be input to initial data → RGB from the output of square 16 and change square 18, reflect diagram data to produce RGB.The RGB that is produced reflects diagram data and is input to geometric transformation square 20, wherein, and by intensive all geometric operation of finishing.The geometric transformation square needs memory body (being shown as being regional memory body 23), the pixel data that is used for conversion process with storage. geometric transformation square 20 has address generator 21, generation is used for the address of regional memory body 23, to use for access data. this geometric transformation square 20 comprises interpolation filter, to produce the pixel data corresponding to the position of being changed.Output 24 from geometric transformation square 20 is input to filtering square 25, and it has the function of relevant local filtering and gain adjustment.Output 26 from filtering square 25 is input to the shooting treatment system.
The following part 27 of the 2nd figure is generally the software program that is embodied on CPU or the DSP for control and decision means. and the major function 30 of this formula square 27 produces parameter and control signal and gives data path for scene analysis and according to the result of scene analysis.The scene of being analyzed is not that initial data input 28 is exactly the output 29 of having handled video signal.The two all can be used as scene analysis simultaneously.After the scene analysis of image action, scene analysis device 30 produces the parameter 31,32 that is used for image stabilization and gives geometric transformation construction device 33, and it is to mix with distribution of compensation or the like.Distribute information to be stored in advance and distribute square 34, and give geometric transformation construction device 33, wherein, produced final argument and control signal 36 via circuit 35.These control signals 36 are given geometric transformation square 20 then, to carry out the geometric transformation of combination.
On the other hand, other scene analysis are also finished in scene analysis device 30 with other functions as Automatic white balance and automatic exposure. and the operation that relevant gain is adjusted is combined and is sent to fader 38.Operation relevant for filtering is combined, and is sent to local filtering combiner 40.Merged at filter structure device 43 from the output 41 of fader 38 and output 42, given filtering square 25 to set up intensive filtering operation result 44 from local filtering combiner 40.Because in this example, prop up relevant and regional relevant operation of the scene that is applied at filtering and gain adjustment, so the scene information 45 from scene analysis device 30 is to be stored in the scene information square 46, this is used to change according to scene itself and this area that is cut as the scene analysis device with being adapted to.
Sometimes, action need scene information on initial data, for example noise reduces may need for example area data of dark space. and the dark space data from scene analysis device 30 are sent to storage area 48, this information 49 is used to handle and for example adapts to the initial data that noise reduces or the like. and for flaw control, need the information on the flaw location and be stored in the storage area 50.This information 51 is used as the flaw compensation at nonlinear filtering square 16.Control signal 52 from scene analysis device 30 is to be sent to the lens module, in order to the control iris diaphgram and to focus lens and shutter speed.
Control and decision means 27 are implemented with efficient processor. and processor can be incorporated on the wafer identical with data path.Another is chosen as these heavy processing is moved to more powerful processor, for example is mainly used in the back end processor of application.When these controls were moved to back end processor, the front end wafer was to carry out signal processing at a high speed with hardware. this makes design hardware data path keep elasticity easily and easily.In addition, because the enforcement of the hardware of high speed data path, so compared to the pure software implementer, power consumption is also less.This data path does not need complicated control and arbitration functions yet, and this also makes data path be integrated into the video sensing wafer easily.
Example 2
In the 3rd figure, show another embodiment of the present invention. be different from image processing stage early, color component is changed into RGB by initial data, this example is the final stage that is placed on image processing to the RGB conversion with original. geometric transformation and filtering are carried out at initial data, rather than carry out at the RGB composition. geometric transformation square 20 is directly delivered in the output 17 from first order nonlinear filtering 16, afterbody original to RGB output 55 be for coming the output of data path since then.All these operations are finished at initial data, and this is corresponding to the output from image sensor itself.Number is almost identical with the 2nd figure person by other squares of 15 to 51 with connection, and except conversion parameter and control signal, these are to be placed between scene analysis device 30 and the geometry square 20, and filtering square 25 clearly is not shown among the 3rd figure.Usually this conversion is finished by the secondary formula of other softwares in square.
When initial data was transferred to the RGB data, data volume will expand, and for example three times greater than original raw data.With the processing of this initial data, the data of being handled are reduced tempestuously, and therefore, this allows to reduce processing power.
The invention provides the intelligent framework that is used for camera system, provide elasticity, density data path, it is incorporated into a square with co-operate, and have cut apart control and decision means to processor.Elasticity allows system designer to upgrade and the change deduction after setting up this system.Data path can be suitable for to various operations, and in addition, outside processor also can be realized various controls and judge deduction.With powerful processor, complicated scene analysis and sorted deduction method can be realized according to system requirements.
In addition, when the data path with high-frequency operation was realized with hardware blocks, the power consumption of data path can be lower than pure software and implement method.Yet system resilience is identical with pure software enforcement method.
The invention provides and be used for future version application, for example intelligent camera system of robot version.
The main element symbol description
1 group
2 groups
3 groups
4 groups
5 geometric transformation
6 filtering
7 gains change
8 inputs
9 filtering squares
10 change square
11 obtain and interpolation
12 last levels
13 outputs
15 initial data
16 nonlinear filterings
18 conversion squares
20 geometric transformation squares
21 address generators
23 regional memory bodys
24 outputs
25 filtering squares
26 outputs
27 times parts
The input of 28 initial data
29 outputs
30 major functions
31 parameters
32 parameters
33 conversion construction devices
34 distribute square
35 circuits
36 control signals
38 faders
40 filtering combiners
41 outputs
42 outputs
43 filter structure devices
44 operating results
45 scene information
46 information squares
48 storage areas
49 information
50 storage areas
51 information
52 control signals
55 outputs
60 inputs
61 flaws compensation square
62 position datas
63 shadow compensation squares
64 shade table data
65 conversion squares
66 noises reduce square
67 edge strengthening squares
68 automatic exposure squares
69 lines
70 exposures measure square
71 control squares
72 lines
73 exposure control squares
74 Automatic white balance squares
75 outputs
76 grasp square
77 control squares
78 control signals
79 gammas compensation square
80 gamma tables
81 special effects squares
82 gain compensation squares
83 results
84 system controllers
Claims (4)
1. video signal processing circuit comprises:
Be used to handle the data path mechanism of signal of video signal, merge concentrated have nonlinear filtering mechanism, color space transformation mechanism, geometric transformation mechanism, linear filtering mechanism and gain adjusting mechanism, to handle an original image signal, to export a processed signal; And
Control and decision mechanism, in order to receive this original image signal and the processed signal of output, produce most nonlinear filtering mechanism, color space transformation mechanism, geometric transformation mechanism, linear filtering mechanism and gain controlling mechanisms that signal is given this data path mechanism respectively that adjust from data path mechanism;
This nonlinear filtering mechanism is in order to realize that noise reduces and the function of flaw compensation;
This color space transformation mechanism in order to will be received from the signal of aforementioned nonlinear filtering mechanism, is converted to color space data by the original image signal data;
This geometric transformation mechanism is in order to carry out distortion compensation and blurring function;
This linear filtering mechanism is in order to carry out the edge strengthening function; And
This adjusting mechanism that gains is in order to carry out automatic exposure control, Automatic white balance control and gamma compensation;
This control and decision mechanism comprise a control routine, and in order to carry out scene analysis and data path construction, this control routine comprises:
Scene analysis mechanism in order to taking out the scene feature and the parameter of original image signal and processed signal respectively, and exports a control signal and gives a lens module, with the control iris diaphgram and to focus lens and shutter speed;
Data path construction mechanism is used for the parameter and the signal of construction data path in order to calculating;
Geometric transformation construction mechanism produces one and adjusts signal, gives this geometric transformation mechanism, with the geometric transformation mechanism of framework in data path;
Gain adjusting mechanism and local filtering mechanism in order to be connected to a filter structure mechanism, adjust signal to produce one, give this linear filtering mechanism and this gain controlling mechanism.
Dark space information stocking mechanism is adjusted signal to this nonlinear filtering mechanism in order to a dark space information to be provided; And
The flaw location stocking mechanism is adjusted signal to this nonlinear filtering mechanism in order to a flaw location to be provided.
2. as 1 described video signal processing circuit of claim the, wherein these color space data are red, green, blue (RGB) color space data.
3. as 1 described video signal processing circuit of claim the, wherein these color space data are YCrCb color space data.
4. as 1 described video signal processing circuit of claim the, wherein this data path mechanism and this control and decision mechanism are to be incorporated into on the semiconductor wafer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101185519A CN100464568C (en) | 2005-10-31 | 2005-10-31 | Image processing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101185519A CN100464568C (en) | 2005-10-31 | 2005-10-31 | Image processing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1960437A CN1960437A (en) | 2007-05-09 |
| CN100464568C true CN100464568C (en) | 2009-02-25 |
Family
ID=38071911
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101185519A Expired - Fee Related CN100464568C (en) | 2005-10-31 | 2005-10-31 | Image processing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100464568C (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8797429B2 (en) * | 2012-03-05 | 2014-08-05 | Apple Inc. | Camera blemish defects detection |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0832827A (en) * | 1994-07-13 | 1996-02-02 | Toppan Printing Co Ltd | Gradation correction divice for digital image |
| JPH0965158A (en) * | 1995-08-25 | 1997-03-07 | Sanyo Electric Co Ltd | Picture processor |
| JP2003260027A (en) * | 2003-02-06 | 2003-09-16 | Olympus Optical Co Ltd | Image processing unit for endoscope |
| US20050134705A1 (en) * | 2003-12-22 | 2005-06-23 | Moon-Cheol Kim | Digital image processing apparatus and method thereof |
| US20050152597A1 (en) * | 2004-01-14 | 2005-07-14 | Eastman Kodak Company | Constructing extended color gamut digital images from limited color gamut digital images |
| CN1643894A (en) * | 2002-08-30 | 2005-07-20 | 佳能株式会社 | Image processing apparatus and method |
| US20050168595A1 (en) * | 2004-02-04 | 2005-08-04 | White Michael F. | System and method to enhance the quality of digital images |
-
2005
- 2005-10-31 CN CNB2005101185519A patent/CN100464568C/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0832827A (en) * | 1994-07-13 | 1996-02-02 | Toppan Printing Co Ltd | Gradation correction divice for digital image |
| JPH0965158A (en) * | 1995-08-25 | 1997-03-07 | Sanyo Electric Co Ltd | Picture processor |
| CN1643894A (en) * | 2002-08-30 | 2005-07-20 | 佳能株式会社 | Image processing apparatus and method |
| JP2003260027A (en) * | 2003-02-06 | 2003-09-16 | Olympus Optical Co Ltd | Image processing unit for endoscope |
| US20050134705A1 (en) * | 2003-12-22 | 2005-06-23 | Moon-Cheol Kim | Digital image processing apparatus and method thereof |
| US20050152597A1 (en) * | 2004-01-14 | 2005-07-14 | Eastman Kodak Company | Constructing extended color gamut digital images from limited color gamut digital images |
| US20050168595A1 (en) * | 2004-02-04 | 2005-08-04 | White Michael F. | System and method to enhance the quality of digital images |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1960437A (en) | 2007-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111698434B (en) | Image processing apparatus, control method thereof, and computer-readable storage medium | |
| CN101309360B (en) | Video input processor, imaging signal-processing circuit, and method for reducing noises in imaging signals | |
| CN103327342B (en) | There is the imaging system of opaque filter pixel | |
| US8670043B2 (en) | Image processing apparatus for detecting a color flicker in an image | |
| CN101283604B (en) | Image processing device with automatic white balance | |
| US7965319B2 (en) | Image signal processing system | |
| US8169500B2 (en) | Dynamic range compression apparatus, dynamic range compression method, computer-readable recording medium, integrated circuit, and imaging apparatus | |
| US8629919B2 (en) | Image capture with identification of illuminant | |
| US7880783B2 (en) | Image pickup apparatus and method | |
| US20070024721A1 (en) | Compensating for improperly exposed areas in digital images | |
| WO2008056565A1 (en) | Image picking-up system and image processing program | |
| US9307213B2 (en) | Robust selection and weighting for gray patch automatic white balancing | |
| CN106358030A (en) | Image processing apparatus and image processing method | |
| CN101895692B (en) | Image processor and image processing method | |
| CN110213502A (en) | Image processing method, device, storage medium and electronic equipment | |
| US7995120B2 (en) | Image brightness compensating apparatus and method, recorded medium recorded the program performing it | |
| CN101409850A (en) | Method and apparatus for processing color components of pixel | |
| JPH11102430A (en) | Image processor | |
| JP2022115944A (en) | Imaging apparatus | |
| JP2003333616A (en) | White balance adjustment method, image pickup device, program, and storage medium | |
| CN100464568C (en) | Image processing system | |
| JP2000165896A (en) | White balance control method and its system | |
| CN110876042B (en) | Circuit for controlling image capture device and related control method | |
| CN114531521A (en) | Image processing method, image processing device, storage medium and electronic equipment | |
| US7719593B2 (en) | Variable frame rate imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090225 Termination date: 20151031 |
|
| EXPY | Termination of patent right or utility model |