CN106780384A - A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method - Google Patents
A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method Download PDFInfo
- Publication number
- CN106780384A CN106780384A CN201611163880.XA CN201611163880A CN106780384A CN 106780384 A CN106780384 A CN 106780384A CN 201611163880 A CN201611163880 A CN 201611163880A CN 106780384 A CN106780384 A CN 106780384A
- Authority
- CN
- China
- Prior art keywords
- image
- mist
- dark channel
- channel diagram
- defogging
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 210000000683 abdominal cavity Anatomy 0.000 title claims abstract description 17
- 230000006978 adaptation Effects 0.000 title claims abstract description 5
- 239000003595 mist Substances 0.000 claims abstract description 76
- 238000010586 diagram Methods 0.000 claims abstract description 45
- 238000005286 illumination Methods 0.000 claims abstract description 24
- 230000008030 elimination Effects 0.000 claims abstract description 23
- 238000003379 elimination reaction Methods 0.000 claims abstract description 23
- 238000002834 transmittance Methods 0.000 claims abstract description 17
- 238000003384 imaging method Methods 0.000 claims abstract description 10
- 230000004438 eyesight Effects 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 8
- 239000002245 particle Substances 0.000 description 7
- 238000002357 laparoscopic surgery Methods 0.000 description 6
- 235000019504 cigarettes Nutrition 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000001228 spectrum Methods 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 4
- 239000003086 colorant Substances 0.000 description 3
- 238000007912 intraperitoneal administration Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 239000000779 smoke Substances 0.000 description 3
- 210000001015 abdomen Anatomy 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000009738 saturating Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method, including step：The frame per second of abdominal cavity image vision signal is obtained, the sample frame of video image is selected according to frame per second；Obtain the original image containing mist of sample frame；Former image containing mist is obtained into dark channel diagram after dark channel prior defogging principle is calculated, the transmittance figure of dark channel diagram is calculated, the air illumination tensor of dark channel diagram is calculated using the method for autoregressive moving average；Using former image containing mist, dark channel diagram, air illumination tensor and transmittance figure, mist figure imaging equation is solved, obtain mist elimination image；Mist elimination image is exported to form defogging video in real time.Due to seeking its transmissivity using down-sampled for high-definition image, drastically reduce the area needs data volume to be processed, improves the real-time of smog；And calculated air illumination tensor and be transformed into defogging figure HSV space with autoregressive moving average method and carry out image enhaucament, eliminate the dim side effect of image after defogging.
Description
Technical field
The present invention relates to the digital image processing techniques field of medical image, and in particular to a kind of applicable cold light source abdominal cavity shadow
Smog method is gone as parameter adaptive in real time.
Background technology
With the continuous progress of medical level, laparoscope is widely used to various surgical operations.In some diseases, abdomen
Endoscope-assistant surgery has turned into the preferred manner for the treatment of.However, laparoscope eyeglass hazes (Laparoscopic lens in operation
Fogging, LLF) situation usually make operation need pause, lose eyesight suddenly equivalent to surgeon, leverage hand
The fluency of art, or even cause unnecessary the medical treatment safe problem sometimes.LLF is to keep the clear visual field in laparoscopic procedures
A major obstacle.It is caused by slow accumulation, to often lead to operating time increase on camera lens by particle, blood and smog.
The reason for lens of laparoscope hazes be due to it is intraperitoneal it is outer there is the temperature difference, if the temperature of lens of laparoscope be less than intraperitoneal temperature
Degree, after camera lens enters abdominal cavity, intraperitoneal hot gas runs into cold camera lens, and part vapor is very easy to cohesion, so as to be formed
One layer of mist, influence visual field definition on a monitor.
Current research shows that the method for improving laparoscope degree of getting a clear view is broadly divided into four classes：A) physical method heating
Laparoscope；B) physics of anti-fog solutions is wiped；C) the material technology innovation of laparoscopic device；D) tissue is wiped.Above method base
Originally belong to physics defogging method, substantial amounts of time consumption keep surgical field of view it is clear on, operation is become tediously long, and
Effect is limited, and often time-consuming, high cost, and without enough data come support its improve effect.In order to more have
The solution eyeglass of effect hazes this problem, it is contemplated that carrying out Digital Image Processing using computer processing technology.Noise
(noise) CCD/CMOS is primarily referred to as to receive and thick in produced image during exporting light as signal is received
Rough part, also refers to the external pixel that should not occur in image, it appears that just as image is covered with some tiny rough points.Fog and cigarette
Dirt is located at namely noise among image, and the noise reduction technology of single-frame images is highly developed.Video image is multiple single frames
Adding up for image, if noise reduction technology is used for into Computer Vision, can haze fulsome in laparoscopic surgery
Cigarette is removed.
Digitized image defogging has many methods, and the scope of application is not quite similar.More reliable defogging method is at present
Bright et al. " the image defogging methods based on dark channel prior " for proposing of He Kai, for single-frame images, dark channel prior defogging is
The algorithms most in use of digital picture defogging.It estimates the transmission of artwork by asking for dark matrix, the atmosphere light composition of original image
Rate matrix, is finally calculated by mist figure imaging model and tries to achieve defogging figure.This method is applied widely, can be good at for abdomen
Close the features such as narrow and light source feature and carry out defogging treatment in the hysteroscope visual field.In order to obtain finer transmission rate matrix, lead to
Frequently with the method for Steerable filter.But many floating-point operations are included during this, so the speed for the treatment of image is slow,
It is difficult to meet the requirement of the real-time defogging of laparoscopic surgery high-resolution video in recent years.Need to carry out technological improvement that reality could be met
The requirement of when property, and also have the space for continuing to optimize.
Due to the upgrading of hardware device, the resolution ratio of laparoscopic surgery video was commonly the high definition of 1920*1080 in recent years
Signal, so optimization dark channel prior defogging algorithm, realizes the real-time defogging of high-resolution video, builds a set of digitlization defogging
System, the picture rich in detail after treatment is shown in real time, can be brought great convenience to the laparoscopic surgery of doctor, reduces operation
Risk.
The content of the invention
Go that the computing of smog method is complicated, processing speed is slow, poor real problem, this Shen for cold light source abdominal cavity image
A kind of the real-time of cold light source abdominal cavity image self-adaptive that be applicable please be provided and go smog method, including step：
The frame per second of abdominal cavity image vision signal is obtained, the sample frame of video image is selected according to frame per second；
The video image is sampled according to sample frame, to obtain the original image containing mist of sample frame；
Former image containing mist is obtained into dark channel diagram after dark channel prior defogging principle is calculated, the saturating of dark channel diagram is calculated
Rate figure is penetrated, the air illumination tensor of dark channel diagram is calculated using the method for autoregressive moving average；
Using former image containing mist, dark channel diagram, air illumination tensor and transmittance figure, mist figure imaging equation is solved, obtained
Mist elimination image；
Mist elimination image is exported to form defogging video in real time.
In a kind of embodiment, the air illumination tensor of dark channel diagram is calculated using the method for autoregressive moving average, including
Step：
Influence factor corresponding to the air illumination tensor of the dark channel diagram for setting each frame；
The error term of air illumination tensor is estimated using moving average model；
The model of autoregressive moving average is set up according to influence factor and error term；
Air illumination tensor is gradually approached and obtained using the model of autoregressive moving average.
In a kind of embodiment, the sample frame of video image, including step are selected according to frame per second：
Whether frame per second is judged more than 25FPS, if being more than, sample frame is that otherwise, sample frame is to sample frame by frame every frame sampling.
In a kind of embodiment, accorded with by pointer operation and access the original image containing mist, the dark channel diagram that are stored in memory headroom
Picture and transmittance figure.
In a kind of embodiment, the transmittance figure of dark channel diagram, including step are calculated：
Judge whether dark channel diagram is high-definition image, if it is not, transmissivity is calculated according to dark channel diagram, if so, to helping secretly
Road figure carries out down-sampled, and calculates transmissivity according to the dark channel diagram after down-sampled.
In a kind of embodiment, by former image containing mist during obtaining dark channel diagram after dark channel prior defogging is calculated
The step of including being processed dark channel diagram using medium filtering.
In a kind of embodiment, using multiple median filters to dark gray level image parallel processing.
In a kind of embodiment, the size of the filter window of median filter is in just with the filter radius of the median filter
It is related.
In a kind of embodiment, mist figure imaging equation is solved, obtain mist elimination image, including step：
Judge whether former image containing mist is high-definition image, if so, original image containing mist is carried out it is down-sampled, if it is not, then straight
Connect the former image containing mist of extraction；
Original image containing mist after will be down-sampled or the original image containing mist for directly extracting, dark channel diagram, air illumination tensor and
Transmittance figure is solved by mist figure imaging model, obtains mist elimination image.
In a kind of embodiment, mist elimination image is exported before forming defogging video also including mist elimination image is transformed into HSV
The step of color space carries out image enhaucament.
Smog method is gone in real time according to above-described embodiment, because the method for autoregressive moving average calculates atmosphere light according to system
Number, reduces the error between estimate and true coefficient, and mist elimination image is transformed into HSV color spaces carries out image increasing
By force, effectively overcome the cross-color occurred during dark defogging, make mist elimination image more life-like.
In addition, during dark channel prior, being accorded with by pointer operation and accessing data, dark channel prior is greatly reduced
Run time, and, take down-sampled method to ask its transmittance figure and defogging figure, drastically reduce the area needs number to be processed
According to amount.
In addition, during dark channel prior, by median filter process dark channel image, can preferably preserve figure
As edge, moreover it is possible to smooth flat region, moreover it is possible to by adjusting the size of filter window, to adapt to going for different size of smoke particle
Remove, can reach and go cigarette effect well.
Brief description of the drawings
Fig. 1 is to be applicable the real-time of cold light source abdominal cavity image self-adaptive to remove smog method flow diagram；
Fig. 2 is to be applicable the real-time of cold light source abdominal cavity image self-adaptive to remove smog method details flow chart.
Specific embodiment
The present invention is described in further detail below by specific embodiment combination accompanying drawing.
This example provides a kind of the real-time of cold light source abdominal cavity image self-adaptive that be applicable and goes smog method, its flow chart such as Fig. 1 institutes
Show, its particular flow sheet is as shown in Fig. 2 comprise the following specific steps that.
S1：The frame per second of abdominal cavity image vision signal is obtained, the sample frame of video image is selected according to frame per second.
Using the original abdominal cavity image vision signal for having mist as input, the frame per second of abdominal cavity image vision signal is obtained, according to
Frame per second selects the sample frame of video image, i.e., take different interframe sample modes, current laparoscope respectively according to different frame per second
The frame per second that photographic equipment is used mainly has 25FPS and two kinds of 50FPS, specifically, whether frame per second is judged more than 25FPS, if being more than,
Sample frame is that otherwise, sample frame is to sample frame by frame every frame sampling.
Our step unification is sampled according to 25FPS, and the frame per second of image output video flowing is also set to 25FPS so that
The time-consuming reduction that frame rate signal high is processed in defogging, while also can guarantee that video real-time processing without interim card, delay.
S2：Video image is sampled according to sample frame, to obtain the original image containing mist of the sample frame.
S3：Dark channel prior defogging principle is calculated, and obtains dark channel diagram, air illumination tensor and transmittance figure.
Statistics law discovery, in most of regional areas of non smoke image, some of which pixel can be at certain
Color Channel is contained within low-down pixel value, i.e., the minimum value very little of each passage of all pixels in certain region.One
As think, this phenomenon is caused by the shade in image scene, color body or black object.This example presses block (Local map
Picture) define image dark channel image be：
J in formulacRepresent each passage of coloured image, Ω (x) represents a window centered on pixel x, c represent r,
G, b triple channel.
The theory of dark channel prior is pointed out：
Jdark→0 (2)
This conclusion is the statistics based on a large amount of natural image dark channel diagrams and the priori conclusion that draws.Made in real life
Low channel value mainly has three factors into three primary colors：A) automobile, building and in city glass window shade, or tree
The projection of the natural landscapes such as leaf, tree and rock；B) bright object or surface, some passages in three passages of RGB
Value it is very low (flower, the leaf of such as green meadow, tree, plant, red or yellow, or blueness the water surface)；C) color
Dark object or surface, such as ash dark-coloured trunk and stone.In a word, it is shade or coloured silk in natural scene everywhere
Color, the low channel value of three primary colors of the image of these scenery is always very low.Dark theory is applied to one in laparoscopic surgery
Sample is applicable, and it is then human organ, tissue, blood vessel etc. that the low channel value of three primary colors is very low in lens of laparoscope.
This step carries out dark primary priori treatment to former mist elimination image using above-mentioned dark formula, this step it
Before, in addition it is also necessary to median filter process is carried out to former mist elimination image, denoising is carried out to former mist elimination image, effectively weaken picture noise,
Be conducive to subsequently setting up accurate mist formation model, step S3 specifically includes following steps：
S31：Accorded with by pointer operation and access corresponding original image containing mist.
In this step and following steps, when being related to the access of image, read using pointer operation symbol traversal,
Accord with accessing by pointer operation and be stored in original image containing mist in memory headroom, dark channel image and saturating
Rate figure is penetrated, pointer operation symbol can carry out high-speed read-write, substantially reduce program runtime to data,
S32：Former image containing mist is obtained into dark channel diagram after dark channel prior defogging is calculated.
Also include processing dark channel diagram using medium filtering in this step, median filtering algorithm can be protected preferably
Deposit dark channel diagram edge, moreover it is possible to smooth flat region, meanwhile, median filtering algorithm is more efficient than mini-value filtering；Further,
This step using multiple median filters to dark gray level image parallel processing, filter under the conditions of multinuclear is called in support by multiple intermediate values
Ripple device carries out concurrent operation, accelerates the computing of this step.
The size of the filter window of the median filter of this example is proportionate with the filter radius of median filter, typically has
The size of WindowSize=Radius*2+1, i.e. window is determined by the size of filter radius, and the size value of filter radius
The size of the cigarette in view of solid particle shape is needed, so, medium filtering has very to the solid particle similar to highlighted noise
Good removal effect, also, by adjusting the size of filter window, it is adapted to the removal of different size of particle.
S33：The air illumination tensor of dark channel diagram is calculated using the method for autoregressive moving average.
The method of autoregressive moving average is：Regard the data sequence that prediction index is formed over time as one
Random sequence, the dependence that this group of stochastic variable has embodies initial data continuity in time.On the one hand, shadow
The influence of the factor of sound, on the other hand, there is itself Fluctuation again, first sets influence factor as x1, x2 ..., xk, its regression analysis
It is formula (3)；
Y=β0+β1x1+β2x2+...+βkxk+e (3)；
Wherein Y is the observation for predicting object, and e is error, is influenceed by Self-variation as prediction object Yt, its rule
Rule can be embodied by formula (4), and wherein β is the weight parameter in autoregression model, and what is represented in formula (4) is p rank autoregressions
Model；
Yt=β0+β1xt-1+β2xt-2+...+βpxt-p+et(4)；
The random degree of Changing Pattern of error term (including all types of signal noises) is larger, can use moving average model
Estimated, i.e., the error term of air illumination tensor is estimated using moving average model, and it represents that wherein α is sliding by formula (5)
Weight parameter in dynamic averaging model, μtIt is the desired value of error term, the middle representative of formula (5) is q rank moving average models；
et=α0+ααet-1+α2et-2+...+αqet-q+μt(5)；
The autoregressive moving-average model expression formula set up by formula (4) and formula (5) is formula (6)；
Yt=β0+β1xt1+β2xt-2+...+βpxt-q+α0+α1et-1α2et-2+...+αqet-q+μt(6)；
The atmosphere light of air illumination tensor A1, A2 ..., Ak, Y the correspondence prediction of each frame of wherein x1, x2 ..., xk correspondence
According to the actual value of coefficient A, using autoregressive moving-average model can gradual approaching to reality air illumination tensor.
S34：Calculate the transmissivity of dark channel diagram.
Judge whether dark channel diagram is high-definition image, if it is not, transmissivity is calculated according to dark channel diagram, if so, to helping secretly
Road figure carries out down-sampled, and calculates transmissivity according to the dark gray level image after down-sampled so that, computation burden is small, efficiency
It is high.
S4：Using former image containing mist, dark channel diagram, air illumination tensor and transmittance figure, mist figure imaging equation is solved, obtained
Obtain mist elimination image.
In computer vision and Digital Image Processing, the mist figure described by following equations is widely used into model：I
(x)=J (x) t (x)+A (1-t (x)) (7)；
Wherein, I (x) is the image that has had, this example middle finger original mist elimination image now, and J (x) is the fogless figure to be recovered
Picture, this example middle finger mist elimination image, A is global atmosphere light image, and t (x) is transmittance figure picture.
Formula (7) is deformed into following formula：
Wherein subscript c represents r, g, b triple channel.
First, it is assumed that being constant in each window internal transmission factor t (x), it is defined asAnd A values have given, so
Minimum operation twice is asked to formula (8) both sides afterwards, following formula is obtained：
Wherein J is fogless image to be asked, and is had according to foregoing dark primary priori theoretical：
Therefore have：
Obtained during formula (11) is substituted into formula (9)：
In above formulaThe as estimate of transmissivity.
In actual life, even fine day white clouds, some particles are there is also in air, therefore, see the object of distant place
Or the influence of mist can be felt, in addition, the presence of mist allows the mankind to feel the presence of the depth of field, therefore, it is necessary to defogging when
Wait and retain a certain degree of mist, this can be by introducing a weight factor ω in formula (12), and formula (12) is modified to：
In this example, ω=0.75 is taken.Soft matting correction algorithms be used to eliminate the image saw that section technique is produced
Tooth very edge blurry, but because this efficiency of algorithm is very low, algorithm is significantly increased and takes, there is serious unfavorable shadow to real-time processing
Ring, and lifting in defog effect is not obvious, so the step for this example has been given up.
Based on the known hypothesis of A values in above-mentioned inference.In practical operation, comprising the following steps that for A values is determined：(1) from
Preceding 0.1% pixel is taken according to the size of brightness in dark channel diagram；(2) in these positions, found in original foggy image I
The value of the corresponding point with maximum brightness, as A values.
When the value very little of projection ratio t, the value of J can be caused bigger than normal, so that image is overall excessive to white field, therefore one
As a threshold value T is set0=0.1, t ＜ T0When take t=0.1.
Final recovery formula is as follows：
, it is necessary to first judge whether former image containing mist is high-definition image before being calculated using formula (14), if so, containing to original
Mist image carry out it is down-sampled, if it is not, then directly extract former image containing mist, to the original image drop sampling containing mist of high definition after, can be with
Reducing needs data volume to be processed.
Then, then will be down-sampled after original image containing mist or directly extract original image containing mist, dark channel diagram, atmosphere light photograph
Coefficient and transmittance figure are solved by mist figure imaging model (14), finally, obtain mist elimination image.
S5：Mist elimination image is exported to form defogging video in real time.
Due to employing down-sampled treatment in step S3 and step S4, so, before video is formed, in addition it is also necessary to going
Mist image is transformed into HSV color spaces, carries out image enhaucament.
HSV (Hue, Saturation, Value) was created in 1978 years by A.R.Smith according to the intuitive nature of color
A kind of color space, also referred to as hexagonal pyramid model (Hexcone Model).The parameter of color is respectively in this model：Color
Adjust (H), saturation degree (S), lightness (V).
Tone H：Measured with angle, span is 0 °~360 °, is calculated counterclockwise since red, and red is
0 °, green is 120 °, and blueness is 240 °.Their complementary color is：Yellow is 60 °, and cyan is 180 °, and magenta is 300 °；
Saturation degree S：Saturation degree S represents color close to the degree of spectrum colour.A kind of color, can regard certain spectrum colour as
The result mixed with white.Ratio wherein shared by spectrum colour is bigger, and color is just higher close to the degree of spectrum colour, color it is full
It is also just higher with spending.Saturation degree is high, and color is then deep and gorgeous.The white light composition of spectrum colour is 0, and saturation degree reaches highest.Generally take
Value scope is 0%~100%, and value is bigger, and color gets over saturation.
Lightness V：Lightness represents bright degree, and for light source colour, brightness value is relevant with the brightness of illuminator；It is right
In object color, this value is relevant with the transmittance or reflectivity of object.Usual span is 0% (black) to 100% (white).
RGB and CMY color model is all that, towards hardware, and HSV (Hue Saturation Value) color model is
Manward's vision, compensate just for color saturation (Saturation) to carry out image increasing in this example all the time
By force.
Specifically, the image after defogging is first transformed into HSV space, the color saturation of image is then reduced to original
80%, enhanced image is more more life-like than the color of image after defogging.
By step S1~S5, this example is by two sets of frame per second respectively 25FPS and 50FPS, resolution ratio providing hospital
The laparoscopic surgery video (wherein 25FPS takes 6 videos, and 50FPS takes 1 video) for being 1920*1080 is tested,
CPU is i7 4700, translation and compiling environment under conditions of Visual Studio 2010, output defogging video is relative to former video
Average lag-time is 30.6ms, and maximum lag time is 35ms, realizes real-time defogging, and have at self adaptation for frame per second
Reason ability.Specific experiment lag time statistics is shown in Table 1.
The defogging of table 1. processes lag time statistical form
Because by adaptively selected sample frame, the sample frame according to selection is adopted to video frame rate in step sl
Sample, and in step s3, accorded with by pointer operation and access data, the operation time of dark channel prior is greatly reduced, and, adopt
Air illumination tensor is calculated with the method for autoregressive moving average, its estimate is more approached actual value, for helping secretly for high definition
Road gray level image seeks its transmittance figure by down-sampled, and drastically reduce the area needs data volume to be processed, and number is greatly improved
According to treatment effeciency, further, the real-time of smog is improved, and by median filter process dark gray level image, can
Image border is preferably preserved, moreover it is possible to smooth flat region, moreover it is possible to by adjusting the size of filter window, to adapt to different size
Smoke particle removal, can reach and go cigarette effect well.
Use above specific case is illustrated to the present invention, is only intended to help and understands the present invention, is not used to limit
The system present invention.For those skilled in the art, according to thought of the invention, can also make some simple
Deduce, deform or replace.
Claims (10)
1. a kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method, it is characterised in that including step：
The frame per second of abdominal cavity image vision signal is obtained, the sample frame of video image is selected according to the frame per second；
The video image is sampled according to the sample frame, to obtain the original image containing mist of the sample frame；
The original image containing mist is obtained into dark channel diagram after dark channel prior defogging principle is calculated, the dark channel diagram is calculated
Transmittance figure, the air illumination tensor of the dark channel diagram is calculated using the method for autoregressive moving average；
Using the original image containing mist, dark channel diagram, air illumination tensor and transmittance figure, mist figure imaging equation is solved, obtained
Mist elimination image；
By the mist elimination image, output forms defogging video in real time.
2. smog method is gone in real time as claimed in claim 1, it is characterised in that the method for the use autoregressive moving average
Calculate the air illumination tensor of the dark channel diagram, including step：
Influence factor corresponding to the air illumination tensor of the dark channel diagram for setting each frame；
The error term of air illumination tensor is estimated using moving average model；
The model of the autoregressive moving average is set up according to the influence factor and error term；
The air illumination tensor is gradually approached and obtained using the model of the autoregressive moving average.
3. smog method is gone in real time as claimed in claim 2, it is characterised in that described that adopting for video image is selected according to frame per second
Sample frame, including step：
Whether frame per second is judged more than 25FPS, if being more than, the sample frame is that otherwise, sample frame is to sample frame by frame every frame sampling.
4. smog method is gone in real time as claimed in claim 3, it is characterised in that access is accorded with by pointer operation and is stored in internal memory
Original image containing mist, dark channel image and transmittance figure in space.
5. smog method is gone in real time as claimed in claim 4, it is characterised in that the transmittance figure of the calculating dark channel diagram,
Including step：
Judge whether the dark channel diagram is high-definition image, if it is not, transmissivity is calculated according to the dark channel diagram, if so, right
The dark channel diagram carries out down-sampled, and calculates transmissivity according to the dark channel diagram after down-sampled.
6. smog method is gone in real time as claimed in claim 5, it is characterised in that described that former image containing mist is passed through into dark elder generation
Test and include the dark channel diagram is processed using medium filtering during dark channel diagram is obtained after defogging principle is calculated
Step.
7. smog method is gone in real time as claimed in claim 6, it is characterised in that helped secretly to described using multiple median filters
Road gray level image parallel processing.
8. go smog method in real time as claimed in claim 7, it is characterised in that the filter window of the median filter it is big
The small filter radius with the median filter are proportionate.
9. smog method is gone in real time as claimed in claim 8, it is characterised in that the solution mist figure imaging equation, gone
Mist image, including step：
Judge whether former image containing mist is high-definition image, if so, the former image containing mist is carried out it is down-sampled, if it is not, then straight
Connect the former image containing mist of extraction；
Original image containing mist or the directly original image containing mist of extraction, dark channel diagram, air illumination tensor and transmission after will be down-sampled
Rate figure is solved by mist figure imaging model, obtains mist elimination image.
10. smog method is gone in real time as claimed in claim 9, it is characterised in that export to form defogging by the mist elimination image
The step of also including that the mist elimination image is transformed into HSV color spaces carries out image enhaucament before video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611163880.XA CN106780384A (en) | 2016-12-15 | 2016-12-15 | A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611163880.XA CN106780384A (en) | 2016-12-15 | 2016-12-15 | A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106780384A true CN106780384A (en) | 2017-05-31 |
Family
ID=58892802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611163880.XA Pending CN106780384A (en) | 2016-12-15 | 2016-12-15 | A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780384A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108093175A (en) * | 2017-12-25 | 2018-05-29 | 北京航空航天大学 | A kind of adaptive defogging method of real-time high-definition video and device |
| CN109408953A (en) * | 2018-10-23 | 2019-03-01 | 上海奕瑞光电子科技股份有限公司 | A kind of background values compensation method suitable for continuous variable frame per second |
| CN109509155A (en) * | 2018-12-17 | 2019-03-22 | 浙江工业大学 | Video defogging method based on rolling time horizon particle group optimizing |
| CN110547752A (en) * | 2019-09-16 | 2019-12-10 | 北京数字精准医疗科技有限公司 | Endoscope system, mixed light source, video acquisition device and image processor |
| CN111539891A (en) * | 2020-04-27 | 2020-08-14 | 高小翎 | Wave band self-adaptive demisting optimization processing method for single remote sensing image |
| CN114638763A (en) * | 2022-03-24 | 2022-06-17 | 华南理工大学 | Image defogging method, system, computer device and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023118A (en) * | 2016-06-13 | 2016-10-12 | 凌云光技术集团有限责任公司 | Image defogging method and realization method on FPGA |
| EP1981390B1 (en) * | 2006-01-30 | 2016-11-09 | Covidien LP | Device for white balancing and applying an anti-fog agent to medical videoscopes prior to medical procedures |
-
2016
- 2016-12-15 CN CN201611163880.XA patent/CN106780384A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1981390B1 (en) * | 2006-01-30 | 2016-11-09 | Covidien LP | Device for white balancing and applying an anti-fog agent to medical videoscopes prior to medical procedures |
| CN106023118A (en) * | 2016-06-13 | 2016-10-12 | 凌云光技术集团有限责任公司 | Image defogging method and realization method on FPGA |
Non-Patent Citations (5)
| Title |
|---|
| LEI GU等: "Virtual Digital Defogging Technology Improves Laparoscopic Imaging Quality", 《SURGICAL INNOVATION》 * |
| 杨燕等: "基于暗通道先验的补偿快速去雾算法", 《计算机工程》 * |
| 王教余: "视频压缩感知和ARMA融合研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 通信学会CIC: "ARMA模型", 《HTTPS://BAIKE.BAIDU.COM/HISTORY/ARMA%E6%A8%A1%E5%9E%8B/8048415/73142415》 * |
| 高健: "基于Zynq7000平台的去雾算法研究及实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108093175A (en) * | 2017-12-25 | 2018-05-29 | 北京航空航天大学 | A kind of adaptive defogging method of real-time high-definition video and device |
| CN109408953A (en) * | 2018-10-23 | 2019-03-01 | 上海奕瑞光电子科技股份有限公司 | A kind of background values compensation method suitable for continuous variable frame per second |
| CN109509155A (en) * | 2018-12-17 | 2019-03-22 | 浙江工业大学 | Video defogging method based on rolling time horizon particle group optimizing |
| CN109509155B (en) * | 2018-12-17 | 2021-10-15 | 浙江工业大学 | Video defogging method based on rolling time domain particle swarm optimization |
| CN110547752A (en) * | 2019-09-16 | 2019-12-10 | 北京数字精准医疗科技有限公司 | Endoscope system, mixed light source, video acquisition device and image processor |
| CN111539891A (en) * | 2020-04-27 | 2020-08-14 | 高小翎 | Wave band self-adaptive demisting optimization processing method for single remote sensing image |
| CN114638763A (en) * | 2022-03-24 | 2022-06-17 | 华南理工大学 | Image defogging method, system, computer device and storage medium |
| CN114638763B (en) * | 2022-03-24 | 2024-05-24 | 华南理工大学 | Image defogging method, system, computer device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106780384A (en) | A kind of the real-time of cold light source abdominal cavity image parameters self adaptation that be applicable goes smog method | |
| CN107527332B (en) | Low-illumination image color retention enhancement method based on improved Retinex | |
| Gao et al. | Sand-dust image restoration based on reversing the blue channel prior | |
| CN102170574B (en) | Real-time video defogging system | |
| CN104794697B (en) | A kind of image defogging method based on dark primary priori | |
| CN106846259A (en) | A kind of the real-time of laparoscopic surgery video frame rate self adaptation goes smog method | |
| CN110570360B (en) | Retinex-based robust and comprehensive low-quality illumination image enhancement method | |
| CN108734670A (en) | The restoration algorithm of single width night weak illumination haze image | |
| CN105046663A (en) | Human visual perception simulation-based self-adaptive low-illumination image enhancement method | |
| CN112419162A (en) | Image defogging method and device, electronic equipment and readable storage medium | |
| WO2024060576A1 (en) | Image dehazing method based on dark channel prior | |
| CN111161167B (en) | Single image defogging method based on middle channel compensation and self-adaptive atmospheric light estimation | |
| CN105447825B (en) | Image defogging method and its system | |
| CN105205794A (en) | Synchronous enhancement de-noising method of low-illumination image | |
| CN109087254A (en) | Unmanned plane image haze sky and white area adaptive processing method | |
| CN108257094A (en) | The quick minimizing technology of remote sensing image mist based on dark | |
| CN107203980B (en) | Underwater target detection image enhancement method of self-adaptive multi-scale dark channel prior | |
| CN110223240A (en) | Image defogging method, system and storage medium based on color decaying priori | |
| Lei et al. | A novel intelligent underwater image enhancement method via color correction and contrast stretching✰ | |
| CN108629750A (en) | A kind of night defogging method, terminal device and storage medium | |
| Liang et al. | Learning to remove sandstorm for image enhancement | |
| CN106504216B (en) | Single image to the fog method based on Variation Model | |
| CN112488926A (en) | System and method for neural network based color restoration | |
| CN108711160A (en) | A kind of Target Segmentation method based on HSI enhancement models | |
| CN109859138B (en) | Infrared image enhancement method based on human visual characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170531 |