CN105898230B - Stitching image luminance proportion method and device based on multiple input path - Google Patents
Stitching image luminance proportion method and device based on multiple input path Download PDFInfo
- Publication number
- CN105898230B CN105898230B CN201610340994.0A CN201610340994A CN105898230B CN 105898230 B CN105898230 B CN 105898230B CN 201610340994 A CN201610340994 A CN 201610340994A CN 105898230 B CN105898230 B CN 105898230B
- Authority
- CN
- China
- Prior art keywords
- image
- channel
- gain
- brightness
- data
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 241001269238 Data Species 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 11
- 239000000284 extract Substances 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000013144 data compression Methods 0.000 claims description 5
- 238000007906 compression Methods 0.000 claims description 4
- 230000006835 compression Effects 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000005303 weighing Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 9
- 230000007704 transition Effects 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 238000013459 approach Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000007323 disproportionation reaction Methods 0.000 description 3
- 238000013075 data extraction Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Processing (AREA)
Abstract
The stitching image luminance proportion method and device based on multiple input path that the present invention provides a kind of, wherein this method comprises the following steps: it is pre-processed in the setting period through each frame image that N number of channel inputs, obtain image data to be spliced, wherein, N >=2, and N is integer；The luminance mean value data of each channel image in the image brightness data of N frame image, and statistics current statistic period are restored from all image datas to be spliced；Judge whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold range, if the channel gain for then choosing setting brightness calculates the luminance gain of remaining channel according to reference channel gain as reference channel gain；Luminance proportion adjusting is carried out to each frame image inputted in next period through N number of channel according to luminance gain.Technical solution of the present invention can be improved the harmonious of stitching image brightness and solve false colour phenomenon, enhance the display effect of image.
Description
Technical field
The present invention relates to technical field of image processing, in particular to a kind of stitching image brightness based on multiple input path is equal
Weigh method and device.
Background technique
In safety defense monitoring system, generallys use panoramic camera and large scene is monitored.Panoramic camera uses more mirrors
Head input is simultaneously spliced the image of input by hardware in real time.However, being made since the hardware such as each camera lens, optical filter have differences
It is had differences at each image of acquisition, and then leads to the brightness irregularities of the panoramic picture of splicing.Common processing method be
Luminance proportion processing is carried out after being spliced into panoramic picture, the improvement that this method can play luminance proportion in certain degree is made
With, however this method cannot adaptive all scenes, such as in the biggish environment of luminance difference (such as indoor environment), bright dark transition
Region will appear colour cast phenomenon, in addition, brightness references channel choosing mistake can also often occur in above-mentioned method, and lead to panoramic picture
Still the problem of brightness disproportionation.
Summary of the invention
The technical problems to be solved by the present invention are: providing a kind of stitching image luminance proportion side based on multiple input path
Method can be improved the harmony of stitching image brightness, enhance the display effect of image.
In order to solve the above-mentioned technical problem, the technical solution adopted by the present invention are as follows: provide a kind of based on multiple input path
Stitching image luminance proportion method, includes the following steps:
It is pre-processed in the setting period through each frame image that N number of channel inputs, obtains image data to be spliced,
Wherein, N >=2, and N is integer；
The image brightness data of N frame image, and statistics current statistic are restored from all image datas to be spliced
The luminance mean value data of each channel image in period；
Judge whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold range, if then selecting
It takes the channel gain of setting brightness as reference channel gain, is increased according to the brightness that reference channel gain calculates remaining channel
Benefit；
Luminance proportion adjusting is carried out to each frame image inputted in next period through N number of channel according to luminance gain.
In order to solve the above-mentioned technical problem, another technical solution used in the present invention are as follows: provide a kind of based on multi input
The stitching image luminance proportion device in channel, comprising:
Preprocessing module, for pre-processing in the setting period through each frame image that N number of channel inputs, obtain to
The image data of splicing, wherein N >=2, and N is integer；
Extraction module, for restoring the image brightness data of N frame image from all image datas to be spliced, and
Count the luminance mean value data of each channel image in the current statistic period；
Judgment module, for judging whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold
Range, if the channel gain for then choosing setting brightness calculates residue according to reference channel gain as reference channel gain
The channel gain in channel；
Adjustment module, it is bright for being carried out according to channel gain to each frame image inputted in next period through N number of channel
Spend well-balanced adjustment.
The beneficial effects of the present invention are: middle monitor camera is different from the prior art and exists when being spliced into panoramic picture
The problem of brightness disproportionation and bright dark transition region colour cast, the stitching image brightness based on multiple input path that the present invention provides a kind of
Equalization methods, by first to each frame image of the multiple input path in the current statistic period carry out pretreatment obtain it is to be output
Then stitching image extracts image luminance information data, by carrying out to image brightness data from stitching image to be output
The statistical disposition of histogram and the luminance gain for calculating each channel, can be according to statistical result, and the gain that will be calculated
Coefficient feedback is adjusted to the luminance proportion that respective data input pin participates in each image of next period and is realized, entire calculation processing
Process while the false colour phenomenon for correcting bright dark transition zone that can be adaptive.It is right before image output in compared with the prior art
The adjusting of brightness of image, so that the panoramic picture overall brightness that each channel image data is spliced after processing is more evenly, face
Color more meets actual scene, reduces the various errors of brightness regulation, can preferably adjust brightness of image harmony, thus increases
The display effect of strong image.
Detailed description of the invention
Specific structure of the invention is described in detail with reference to the accompanying drawing
Fig. 1 is that the present invention is based on the flow charts of one embodiment of stitching image luminance proportion method of multiple input path；
Fig. 2 is the specific flow chart of step S10 in Fig. 1；
Fig. 3 is the specific flow chart of step S20 in Fig. 1；
Fig. 4 is that the present invention is based on the block diagrams of one embodiment of stitching image luminance proportion device of multiple input path；
Fig. 5 is the block diagram of preprocessing module in Fig. 4；
Fig. 6 is the block diagram of extraction module in Fig. 4.
Label declaration book:
10, preprocessing module: 11, gain unit；12, Date Conversion Unit；13, data compression unit；
20, extraction module: 21, reduction unit；22, extraction unit；23, statistic unit；
30, judgment module；
40, adjustment module.
Specific embodiment
In order to describe the technical content, the structural feature, the achieved object and the effect of this invention in detail, below in conjunction with embodiment
And attached drawing is cooperated to be explained in detail.
The most critical design of the present invention is: this programme is by showing that extraction image is bright in image to be spliced in pretreatment
Degree evidence, and each channel luminance gain coefficient to be feedback is calculated by statistics, and then increase according to the brightness of image of feedback
Beneficial coefficient carries out brightness regulation to input picture, can reduce the error of brightness postposition adjusting, pass through the feedback tune of brightness of image
Energy conservation is enough preferably to be adjusted brightness of image harmony and solves the bright dark transition zone colour cast phenomenon being likely to occur, thus enhances figure
The display effect of picture.
Referring to Fig. 1, the embodiment provides a kind of stitching image luminance proportion side based on multiple input path
Method includes the following steps:
S10, it is pre-processed in the setting period through each frame image that N number of channel inputs, obtains image to be spliced
Data, wherein N >=2, and N is integer.Each channel all corresponds to the image input of each channel camera lens, the image in each channel
Image data to be output is capable of forming after splicing.
S20, the image brightness data that N frame image is restored from all image datas to be spliced, and statistics are current
The luminance mean value data of each channel image in measurement period.In above-mentioned steps, all can through each frame image that N number of channel inputs
Image data to be spliced is formed by processing appropriate, which obtains panoramic picture number for the splicing in later period
According to using from image data extraction luminance information data to be spliced, being calculated to count to improve the splicing effect of image
Gain information to be feedback out, specifically, the feedback information is the brightness of image gain coefficient of calculated N frame image.In addition,
Measurement period was in the time in setting period.
S30, judge whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold range, if
The channel gain of setting brightness is then chosen as reference channel luminance gain, remaining channel is calculated according to reference channel gain
Luminance gain.In this step, judged by the brightness of image to adjacency channel, it is super in the brightness of image difference of adjacency channel
It out when preset brightness of image discrepancy threshold range, needs to be adjusted processing, is in the brightness of image difference of adjacency channel
When within the scope of preset brightness of image discrepancy threshold, illustrates that the brightness of image variation of adjacency channel is small, be not necessarily to well-balanced adjustment.Setting
The channel luminance gain of brightness can be set according to the user's choice, it is common there are three types of: most bright channel gain, most dark
Gain and average brightness gain.In order to allow the overall brightness of spliced panorama sketch to be unlikely to partially dark, this programme is preferentially chosen
Most bright channel is reference channel gain.Specifically, selecting channel corresponding to minimum brightness gain gain_min (≤1), this is weight
That most bright channel of folded area is set as chref, by the luminance mean value chref_light_average1 of the channel overlay region and
Chref_light_average2 and overlay region pixel sum chref_sum, with the luminance histogram statistics in the channel
As a result: the bright mean value lh_normal_light_average of normal areas, the bright mean value lh_high_light_ of highlight bar
Average, overall brightness mean value lh_sum_h are compared:
If 1, meeting following conditions,
Then illustrate the bright area of current channel just in overlay region, and the channel is not necessarily exactly most bright channel at this time, because
This cannot directly use the channel as reference channel.If lh_norm_percent maximum value≤20% in each channel, just use
The maximum channel lh_light_average is as reference channel, and corresponding gain value is as reference gain gain_ref；Otherwise
Just judged to select reference channel gain with lh_normal_light_average, lh_normal_light_average is maximum
Channel be selected as reference channel, corresponding luminance gain gain is set as reference gain gain_ref；Terminate to refer to reference channel
Selection.Otherwise enter next step.
If 2, meeting following conditions,
|(chref_light_average1+chref_light_average2)/2-lh_high_light_average|
≤ 10, lh_high_percent > 40%,
Current gain_min is then directly chosen as reference gain gain_ref.Terminate to select with reference to reference channel.It is no
Then enter next step.
If 3, meeting following conditions,
|(chref_light_average1+chref_light_average2)/2-lh_high_light_average|
> 10,
The brightness for then choosing two regions left after removing dark space in each channel luminance statistics with histogram result is equal
The maximum channel value lum_ave is as reference channel, and the gain value in the channel is as reference channel gain gain_ref.At this point,
Lum_ave=(lh_high_light_average*lh_sum_h+lh_normal_light_average*l h_
sum_n)/(lh_sum_h+lh_sum_n)。
S40, luminance proportion tune is carried out to each frame image inputted in next period through N number of channel according to channel gain
Section.This step participates in brightness of image equilibrium calculation by calculated each channel luminance gain, can preferably adjust image
Brightness.
The beneficial effects of the present invention are: middle monitor camera is different from the prior art and exists when being spliced into panoramic picture
The problem of brightness disproportionation and bright dark transition region colour cast, the stitching image brightness based on multiple input path that the present invention provides a kind of
Equalization methods, by first to each frame image of the multiple input path in the current statistic period carry out pretreatment obtain it is to be output
Then stitching image extracts image luminance information data, by carrying out to image brightness data from stitching image to be output
The statistical disposition of histogram and the channel luminance gain for calculating each channel, can be according to statistical result, and will be calculated
Gain coefficient feedback is adjusted to the luminance proportion that respective data input pin participates in each image of next period and is realized, entire to calculate
Treatment process while the false colour phenomenon for correcting bright dark transition zone that can be adaptive., compared with the prior art in it is defeated in image
Adjusting before out to brightness of image, so that the panoramic picture overall brightness that each channel image data is spliced after processing is more equal
Even, color more meets actual scene, reduces the various errors of brightness regulation, can preferably adjust brightness of image harmony,
Thus enhance the display effect of image.
Referring to figure 2., in a specific embodiment, each frame figure inputted in described pair of setting period through N number of channel
Picture is pre-processed, and is obtained the step S10 of image data to be spliced, is specifically included:
S11, gain process is carried out to the image data inputted in the setting period through N number of channel, obtains the figure of preliminary treatment
As data.The gain process namely luminance proportion processing, specifically: by the image data rawdata_i of input multiplied by corresponding
Luminance proportion coefficient gain obtains the image data rawdata_o of preliminary treatment.In addition, measurement period is 20s in the present embodiment
Interior luminance mean value, to reduce the influence that brightness abrupt transients correctly judge brightness change.
S12, the image data progress space of preliminary treatment is converted to RGB data corresponding with N number of channel；The RGB
The digit of data is 12, for the ease of realizing on FPGA, it is also necessary to which 12 RGB datas are compressed into 10 RGB numbers
According to.
S13, image data to be spliced is formed to RGB data progress compression processing according to tone-mapping algorithm.Data pressure
When contracting, image color cast may be caused the region of false colour, especially dark images occur to a certain extent, in order to reduce data letter
The loss of breath, preferably introducing tone-mapping algorithm tonemapping during data space is converted can be by 12 RGB
Data conversion realizes that data bit is compressed at 10 RGB by way of tabling look-up.
Referring to figure 3., in a specific embodiment, described that N frame figure is restored from all image datas to be spliced
Step S20 in the image luminance information data of picture, and statistics current period with the luminance mean value data of each image, specifically
Include:
S21, the RGB data that input picture corresponding with pretreatment top n channel is restored from image data to be spliced.
S22, each RGB data is converted into yuv data, and extracts brightness data Y as each frame figure from yuv data
The image brightness data of picture；
S23, according to the different brightness of image brightness data in image, divide the image into low clear zone lh_low light, in
The region that image is overlapped in clear zone lh_norm light, tri- regions highlight bar lh_high light and adjacency channel, and
Count the luminance mean value of corresponding region in the current statistic period.
The accounting lh_low_percent in low clear zone, middle clear zone accounting lh_norm_percent, the accounting of highlight bar
Lh_high_percent meets:
Lh_low_percent+lh_norm_percent+lh_high_percent=1；
Sum_all=lh_sum_l+lh_sum_n+lh_sum_h；
Wherein, sum_all indicates total pixel number in single channel, and lh_sum_l indicates low brightness area pixel
It counts, bright area pixel number in lh_sum_n expression, lh_sum_h highlight regions pixel number.
Low bright area lh_low_light_average, the middle bright area lh_normal_light_average, height of statistics
Luminance mean value lh_high_light_average, the overall brightness mean value lh_light_average of bright area and channel
The luminance mean value chx_light_average1, chx_light_average2 of image overlay region.
In one embodiment, the channel gain for choosing setting brightness is as reference channel gain, according to reference channel
Gain calculates the step of channel gain of remaining channel, specifically includes:
The setting highest channel gain of brightness is chosen as reference channel gain；
Reference channel gain is normalized, and calculates the luminance gain of remaining channel.
In order to make image overlay region brightness change seamlessly transit, need to meet condition:
Chx_light_average2*gainx ≈ ch (x+1) _ light_average1*gain (x+1),
In this way, the gain of remaining channel can be found out according to above-mentioned formula after the gain for setting a channel.Further
, a sequence can also be done to the gain of striving calculated, find maximum gain gain_max and least gain gain_min,
And average gain value gain_ave, wherein
Gain_max=MAX (gain0, gain1 ... ..gainN)；
Gain_min=MIN (gain0, gain1 ... gainN)；
Gain_ave=(gain0+gain1+...+gainN)/(N+1).
In order to avoid brightness change is violent, the gradually smooth transition of luminance proportion, in one embodiment, the basis are realized
The step of channel gain carries out luminance proportion adjusting to each frame image inputted in next period through N number of channel, specifically includes:
Gain section between channel gain and the channel current gain is divided into several equal portions, obtains the brightness of each equal portions
Regulated value；
Expected channel gain is gradually adjusted according to approximate algorithm, so as to inputted in next period through N number of channel
The luminance proportion of each frame image.
In this step, by the difference between calculated channel gain gainx_new and channel current gain gainx_cur
Value is divided into several equal portions (i.e. each to be divided into a step-length pace), to the brightness gain values that calculate by gainx_cur by
Step smoothly approaches gainx_new, and the length for designing step-length can design acording to the requirement of user and (indicate step number with num), a frame
It is set as a step size computation time, only need to guarantee to complete to approach within the 20s time.The calculation method of step-length are as follows:
Pace=| gainx_new-gainx_cur |/num, (num < 500)
Approximation Operator are as follows:
Gainx_cur=gainx_cur+pace (works as gainx_new > gainx_cur)
Or, gainx_cur=gainx_cur-pace (as gainx_new < gainx_cur).
Referring to figure 4., the present embodiment additionally provides a kind of stitching image luminance proportion device based on multiple input path, packet
Include preprocessing module 10, extraction module 20, judgment module 30 and adjustment module 40.
Preprocessing module 10 is obtained for pre-processing in the setting period through each frame image that N number of channel inputs
Image data to be spliced, wherein N >=2, and N is integer.Each channel all corresponds to the image input of each camera lens acquisition,
The image in each channel is capable of forming image data to be spliced after pretreatment.
Extraction module 20, for restoring the image brightness data of N frame image from all image datas to be spliced, with
And count the luminance mean value data of each channel image in the current statistic period.The each frame image inputted through N number of channel all can be through
It crosses processing appropriate and forms image data to be spliced, which obtains panoramic picture number for the splicing in later period
According to, in order to improve the splicing effect of image, using from image data extraction luminance information data to be spliced to calculate to anti-
Feedforward information, specifically, the feedback information is the luminance gain of the N frame image calculated.
Judgment module 30, for judging whether the brightness of image difference of adjacency channel exceeds preset brightness of image difference threshold
It is worth range, if the channel gain for then choosing setting brightness calculates surplus as reference channel gain according to reference channel gain
The channel gain in remaining channel.The judgment module 30 is judged by the brightness of image to adjacency channel, in the figure of adjacency channel
When image brightness difference exceeds preset brightness of image discrepancy threshold range, need to be adjusted processing, in the image of adjacency channel
When luminance difference is within the scope of preset brightness of image discrepancy threshold, illustrates that the brightness of image variation of adjacency channel is small, be not necessarily to
It adjusts.The channel gain of setting brightness can be set according to the user's choice, it is common there are three types of: most bright channel gain, most
Dark gain and average brightness gain.
Adjustment module 40, for being carried out according to channel gain to each frame image inputted in next period through N number of channel
Luminance proportion is adjusted.This adjustment module 40 participates in brightness of image equilibrium calculation by calculated channel gain, can be preferably
Adjust the brightness of image.
Referring to figure 5., in a specific embodiment, the preprocessing module 10 specifically includes gain unit 11, data
Converting unit 12 and data compression unit 13.
Gain unit 11 is obtained for carrying out luminance proportion processing to the image data inputted in the setting period through N number of channel
To the image data of preliminary treatment；The gain unit 11 is by the image data rawdata_i of input multiplied by corresponding luminance proportion
Coefficient gain obtains the image data rawdata_o of preliminary treatment.In addition, measurement period is the brightness in 20s in the present embodiment
Mean value, to reduce the influence that brightness abrupt transients correctly judge brightness change.
Date Conversion Unit 12, it is corresponding with N number of channel for being converted in the image data progress space of preliminary treatment
RGB data；It is handled through Date Conversion Unit 12 and obtains 12 RGB datas, for the ease of being realized on FPGA, it is also necessary to right
12 RGB datas are compressed into 10 RGB datas.
Data compression unit 13, it is to be spliced for being formed according to tone-mapping algorithm to RGB data progress compression processing
Image data.When data compression, image color cast may be caused the area of false colour, especially dark images occur to a certain extent
Domain preferably introduces during data space is converted according to tone-mapping algorithm to reduce the loss of data information
12 RGB datas can be converted into 10 RGB by tonemapping, realize that data bit is compressed by way of tabling look-up.
Fig. 6 is please referred to, in a specific embodiment, the extraction module 20 specifically includes reduction unit 21, extracts list
Member 22 and statistic unit 23.
Reduction unit 21, for input picture corresponding with N number of channel before restoring pretreatment from image data to be spliced
RGB data；
Extraction unit 22 for each RGB data to be converted into yuv data, and extracts brightness data Y from yuv data
Image brightness data as each frame image；
Statistic unit 23, for the different brightness according to image brightness data in image, divide the image into low clear zone, in it is bright
The region that image is overlapped in area, three regions of highlight bar and adjacency channel, and count corresponding region in the current statistic period
Luminance mean value.Low bright area lh_low_light_average, the middle bright area lh_ that can be counted by the statistic unit 23
Normal_light_average, the luminance mean value lh_high_light_average of highlight regions, overall brightness mean value lh_
The luminance mean value chx_light_average1, chx_light_ of the image overlay region in light_average and channel
average2。
In a specific embodiment, the judgment module 30 includes selection unit, computing unit and judging unit.
The judging unit, for judging whether the brightness of image difference of adjacency channel exceeds preset brightness of image difference threshold
It is worth range.
Selection unit, for choosing the setting highest channel gain of brightness as reference channel gain；
Computing unit for reference channel gain to be normalized, and calculates the channel gain of remaining channel.
In order to make image overlay region brightness change seamlessly transit, need to meet condition:
Chx_light_average2*gainx ≈ ch (x+1) _ light_average1*gain (x+1),
In this way, the gain of remaining channel can be found out according to above-mentioned formula after the gain for setting a channel.Further
, a sequence can also be done to the gain of striving calculated, find maximum gain gain_max and least gain and gain_
Min and average gain value gain_ave, wherein
Gain_max=MAX (gain0, gain1 ... ..gainN)；
Gain_min=MIN (gain0, gain1 ... gainN)；
Gain_ave=(gain0+gain1+...+gainN)/(N+1).
Preferably, the adjustment module 40 specifically includes division unit and adjusts unit,
Division unit obtains every for gain section between channel gain and channel current gain to be divided into several equal portions
The brightness regulation value of one equal portions；
Unit is adjusted, for gradually being adjusted according to approximate algorithm to channel current gain, so that through N in next period
The luminance proportion of each frame image of a channel input.
The division unit is by the difference between calculated channel gain gainx_new and channel current gain gainx_cur
Value is divided into several equal portions (i.e. each to be divided into a step-length pace), and calculates brightness by approximate algorithm by computing unit
Yield value gradually smoothly approaches gainx_new by gainx_cur, and the length for designing step-length can design acording to the requirement of user
(indicating step number with num), a frame are set as a step size computation time, only need to guarantee to complete to approach within the 20s time.
In conclusion the stitching image luminance proportion device provided by the invention based on multiple input path, by using with
The luminance proportion method of upper closed-loop system can carry out the false colour that each channel luminance of image and bright dark transition zone are likely to occur real
Shi Jiaozheng, it is thus possible to which preferably adjusting brightness of image harmony avoids false colour phenomenon from generating, thus enhances the display effect of image
Fruit.
The above description is only an embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (6)
1. a kind of stitching image luminance proportion method based on multiple input path, which comprises the steps of:
It is pre-processed in the setting period through each frame image that N number of channel inputs, obtains image data to be spliced,
In, N >=2, and N is integer；
The image brightness data of N frame image, and statistics current statistic period are restored from all image datas to be spliced
The luminance mean value data of interior each channel image；
Judge whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold range, is set if then choosing
The channel gain of brightness is determined as reference channel gain, the luminance gain of remaining channel is calculated according to reference channel gain, is had
Body includes choosing the setting highest channel gain of brightness as reference channel gain；Place is normalized to reference channel gain
Reason, and the luminance gain of remaining channel is calculated, the gain of the remaining channel is calculated according to the following equation,
Chx_light_average2*gainx ≈ ch (x+1) _ light_average1*gain (x+1),
Wherein, chx_light_average2 is the luminance mean value of x-th of image overlay region, and gainx is x-th of image overlay region
Luminance gain；
Luminance proportion adjusting is carried out to each frame image inputted in next period through N number of channel according to luminance gain, it is specific to wrap
It includes: gain section between channel gain and the channel current gain being divided into several equal portions, obtains the brightness regulation of each equal portions
Value；Expected channel gain is gradually adjusted according to approximate algorithm, so that each frame inputted in next period through N number of channel
The luminance proportion of image.
2. the stitching image luminance proportion method based on multiple input path as described in claim 1, which is characterized in that described right
The step of being pre-processed through each frame image that N number of channel inputs in the setting period, obtain image data to be spliced, specifically
Include:
Gain process is carried out to the image data inputted in the setting period through N number of channel, obtains the image data of preliminary treatment；
The image data of preliminary treatment is subjected to space and is converted to RGB data corresponding with N number of channel；
Compression processing is carried out to RGB data according to tone-mapping algorithm and forms image data to be spliced.
3. the stitching image luminance proportion method based on multiple input path as claimed in claim 2, which is characterized in that it is described from
The image brightness data of N frame image, and statistics current statistic period Nei Getong are restored in all image datas to be spliced
It the step of luminance mean value data of road image, specifically includes:
The RGB data of input picture corresponding with N number of channel before pre-processing is restored from image data to be spliced；
Each RGB data is converted into yuv data, and extracts image of the brightness data Y as each frame image from yuv data
Brightness data；
According to the different brightness of image brightness data in image, divide the image into low clear zone, middle clear zone, three regions of highlight bar,
And the region that image is overlapped in adjacency channel, and count the luminance mean value of corresponding region in the current statistic period.
4. a kind of stitching image luminance proportion device based on multiple input path characterized by comprising
Preprocessing module obtains to be spliced for pre-processing in the setting period through each frame image that N number of channel inputs
Image data, wherein N >=2, and N be integer；
Extraction module, for restoring the image brightness data of N frame image, and statistics from all image datas to be spliced
The luminance mean value data of each channel image in the current statistic period；
Judgment module, for judging whether the brightness of image difference of adjacency channel exceeds preset brightness of image discrepancy threshold model
It encloses, if the channel gain for then choosing setting brightness calculates remaining logical as reference channel gain according to reference channel gain
The channel gain in road, the judgment module specifically include,
Selection unit, for choosing the setting highest channel gain of brightness as reference channel gain；
Computing unit for reference channel gain to be normalized, and calculates the channel gain of remaining channel；
Adjustment module, it is equal for carrying out brightness to each frame image inputted in next period through N number of channel according to channel gain
Weighing apparatus is adjusted, and the adjustment module includes,
Division unit obtains each for the gain section between channel gain and channel current gain to be divided into several equal portions
The brightness regulation value of equal portions；
Unit is adjusted, for gradually being adjusted according to approximate algorithm to channel gain, so that defeated through N number of channel in next period
The luminance proportion of each frame image entered.
5. the stitching image luminance proportion device based on multiple input path as claimed in claim 4, which is characterized in that described pre-
Processing module specifically includes:
Gain unit obtains preliminary treatment for carrying out gain process to the image data inputted in the setting period through N number of channel
Image data；
Date Conversion Unit is converted to RGB number corresponding with N number of channel for the image data of preliminary treatment to be carried out space
According to；
Data compression unit, for carrying out compression processing to RGB data according to tone-mapping algorithm to form image to be spliced
Data.
6. the stitching image luminance proportion device based on multiple input path as claimed in claim 5, which is characterized in that described to mention
Modulus block specifically includes:
Reduction unit, for restoring the RGB number of input picture corresponding with pretreatment top n channel from image data to be spliced
According to；
Extraction unit extracts brightness data Y as every for each RGB data to be converted into yuv data, and from yuv data
The image brightness data of one frame image；
Statistic unit divides the image into low clear zone, middle clear zone, height for the different brightness according to image brightness data in image
The region that image is overlapped in three, clear zone region and adjacency channel, and count the brightness of corresponding region in the current statistic period
Mean value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610340994.0A CN105898230B (en) | 2016-05-20 | 2016-05-20 | Stitching image luminance proportion method and device based on multiple input path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610340994.0A CN105898230B (en) | 2016-05-20 | 2016-05-20 | Stitching image luminance proportion method and device based on multiple input path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105898230A CN105898230A (en) | 2016-08-24 |
| CN105898230B true CN105898230B (en) | 2019-02-05 |
Family
ID=56716660
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610340994.0A Active CN105898230B (en) | 2016-05-20 | 2016-05-20 | Stitching image luminance proportion method and device based on multiple input path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105898230B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2554667B (en) * | 2016-09-30 | 2021-10-20 | Apical Ltd | Image processing |
| CN108259783B (en) * | 2016-12-29 | 2020-07-24 | 杭州海康威视数字技术股份有限公司 | Digital matrix synchronous output control method and device and electronic equipment |
| CN107230193B (en) * | 2017-06-09 | 2020-11-13 | 西安煤航遥感信息有限公司 | Image brightness equalization method for aerial digital camera |
| CN107330872A (en) * | 2017-06-29 | 2017-11-07 | 无锡维森智能传感技术有限公司 | Luminance proportion method and apparatus for vehicle-mounted viewing system |
| CN107820067B (en) * | 2017-10-29 | 2019-09-20 | 苏州佳世达光电有限公司 | The joining method and splicing apparatus of more projected pictures |
| CN109361855B (en) * | 2018-10-24 | 2020-12-11 | 深圳六滴科技有限公司 | Panoramic image pixel brightness correction method and device, panoramic camera and storage medium |
| CN109461639B (en) * | 2018-11-23 | 2020-11-24 | 北方夜视技术股份有限公司 | Optical adjusting method and device |
| CN110060213B (en) * | 2019-04-09 | 2021-06-15 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN112312108B (en) * | 2019-08-02 | 2022-09-13 | 浙江宇视科技有限公司 | White balance abnormity determining method and device, storage medium and electronic equipment |
| CN114723637B (en) * | 2022-04-27 | 2024-06-18 | 上海复瞰科技有限公司 | Color difference adjusting method and system |
| CN115809960A (en) * | 2022-06-08 | 2023-03-17 | 北京爱芯科技有限公司 | Image splicing device, chip and image splicing method |
| CN116506562B (en) * | 2023-06-27 | 2023-09-05 | 深圳市门钥匙科技有限公司 | Video display method and system based on multiple channels |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103942762A (en) * | 2014-04-24 | 2014-07-23 | 胡建国 | Two-dimension code preprocessing method and device |
| CN105005963A (en) * | 2015-06-30 | 2015-10-28 | 重庆市勘测院 | Multi-camera images stitching and color homogenizing method |
| CN105321151A (en) * | 2015-10-27 | 2016-02-10 | Tcl集团股份有限公司 | Panorama stitching brightness equalization method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6020036B2 (en) * | 2012-10-24 | 2016-11-02 | 富士ゼロックス株式会社 | Information processing apparatus, information processing system, and program |
-
2016
- 2016-05-20 CN CN201610340994.0A patent/CN105898230B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103942762A (en) * | 2014-04-24 | 2014-07-23 | 胡建国 | Two-dimension code preprocessing method and device |
| CN105005963A (en) * | 2015-06-30 | 2015-10-28 | 重庆市勘测院 | Multi-camera images stitching and color homogenizing method |
| CN105321151A (en) * | 2015-10-27 | 2016-02-10 | Tcl集团股份有限公司 | Panorama stitching brightness equalization method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105898230A (en) | 2016-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105898230B (en) | Stitching image luminance proportion method and device based on multiple input path | |
| US10134359B2 (en) | Device or method for displaying image | |
| US7656375B2 (en) | Image-processing device and method for enhancing the luminance and the image quality of display panels | |
| CA3114448C (en) | Device and method of improving the perceptual luminance nonlinearity - based image data exchange across different display capabilities | |
| JP4313032B2 (en) | Image brightness control apparatus and method | |
| US9646397B2 (en) | Image processing apparatus and image processing method | |
| US7352410B2 (en) | Method and system for automatic brightness and contrast adjustment of a video source | |
| CN103491357B (en) | A kind of auto white balance treatment method of image sensor | |
| EP2076013A2 (en) | Method of high dynamic range compression | |
| US20090317017A1 (en) | Image characteristic oriented tone mapping for high dynamic range images | |
| WO2007122966A1 (en) | Visual processing device, visual processing method, program, display device, and integrated circuit | |
| TWI519151B (en) | Image processing method and image processing apparatus | |
| KR20090067911A (en) | Apparatus and method for removing color noise of image signal | |
| US8860806B2 (en) | Method, device, and system for performing color enhancement on whiteboard color image | |
| WO2019169589A1 (en) | Method for high-quality panorama generation with color, luminance, and sharpness balancing | |
| KR101113483B1 (en) | Apparatus for enhancing visibility of color image | |
| CN105898252A (en) | Television color adjustment method and device | |
| WO2005055588A1 (en) | Image processing device for controlling intensity of noise removal in a screen, image processing program, image processing method, and electronic camera | |
| CN112785984B (en) | LCD high-efficiency self-adaptive global backlight adjusting method for image gray level perception | |
| CN106327444A (en) | Method and device for color combination in high-dynamic-range image combination | |
| JP2007013666A (en) | Image processor | |
| KR20130117564A (en) | Device for correcting image obtained from camera for vehicle and method for correcting image using the same | |
| US8026926B2 (en) | Image display device and image display method | |
| KR101514152B1 (en) | Method and apparatus for improving image quality using singular value decomposition | |
| US10497149B2 (en) | Image processing apparatus and image processing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |