CN108198153A - A kind of defogging method based on floor heating capsule moving image - Google Patents
A kind of defogging method based on floor heating capsule moving image Download PDFInfo
- Publication number
- CN108198153A CN108198153A CN201810182947.7A CN201810182947A CN108198153A CN 108198153 A CN108198153 A CN 108198153A CN 201810182947 A CN201810182947 A CN 201810182947A CN 108198153 A CN108198153 A CN 108198153A
- Authority
- CN
- China
- Prior art keywords
- capsule
- image
- unit
- value
- floor heating
- 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
- 239000002775 capsule Substances 0.000 title claims abstract description 133
- 238000010438 heat treatment Methods 0.000 title claims abstract description 130
- 238000000034 method Methods 0.000 title claims abstract description 29
- 230000000694 effects Effects 0.000 claims abstract description 9
- 239000003595 mist Substances 0.000 claims abstract description 8
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 238000002834 transmittance Methods 0.000 claims abstract description 4
- 238000012545 processing Methods 0.000 claims description 24
- 230000008030 elimination Effects 0.000 claims description 23
- 238000003379 elimination reaction Methods 0.000 claims description 23
- 230000007246 mechanism Effects 0.000 claims description 22
- 239000007921 spray Substances 0.000 claims description 15
- 238000005507 spraying Methods 0.000 claims description 14
- 239000004831 Hot glue Substances 0.000 claims description 13
- 238000003860 storage Methods 0.000 claims description 11
- 230000033001 locomotion Effects 0.000 claims description 10
- 230000015556 catabolic process Effects 0.000 claims description 8
- 238000006731 degradation reaction Methods 0.000 claims description 8
- 239000000126 substance Substances 0.000 claims description 8
- 238000010586 diagram Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000009826 distribution Methods 0.000 claims description 6
- 239000003814 drug Substances 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 5
- 206010037660 Pyrexia Diseases 0.000 claims description 4
- 239000007788 liquid Substances 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000013459 approach Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005303 weighing Methods 0.000 claims description 3
- 229940079593 drug Drugs 0.000 claims description 2
- 238000013016 damping Methods 0.000 claims 1
- 230000005389 magnetism Effects 0.000 claims 1
- -1 nozzle Substances 0.000 claims 1
- 238000009434 installation Methods 0.000 abstract description 2
- 230000000007 visual effect Effects 0.000 abstract 2
- 230000004456 color vision Effects 0.000 abstract 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 230000005611 electricity Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 229910021389 graphene Inorganic materials 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000004659 sterilization and disinfection Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 230000003137 locomotive effect Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000009987 spinning Methods 0.000 description 1
- 238000012546 transfer Methods 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
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24D—DOMESTIC- OR SPACE-HEATING SYSTEMS, e.g. CENTRAL HEATING SYSTEMS; DOMESTIC HOT-WATER SUPPLY SYSTEMS; ELEMENTS OR COMPONENTS THEREFOR
- F24D15/00—Other domestic- or space-heating systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Thermal Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
Abstract
A kind of defogging method based on floor heating capsule moving image is claimed in the present invention, includes the following steps：Floor heating capsule moving image is obtained, original mist figure is enhanced using the sub-block local histogram equalization that partly overlaps, and passes through DCP algorithms and obtains the transmissivity after optimization；Using MRF models to the binding feature of picture structure detailed information, transmissivity is modeled, further to refine transmissivity；By the notable feature of sky domain, air light value is asked for by partitioned searching method.The present invention can obtain more accurate transmittance figure, effectively keep image structure information, and the figure after defogging shows abundant details and more actually color vision effect.Present invention installation is simple, conveniently, and the visual modularization floor heating track of innovative invention, visual pattern.
Description
Technical field
The invention belongs to a kind of defogging methods of image in floor heating technical field more particularly to floor heating track.
Background technology
In existing ground warm device, hydroelectric floor heating, medium is water and electricity, by water and electrical heating, then passes through pipeline
Motionless with insulating layer, direct-burried backfills secondary heat dissipation, consumes a large amount of thermal resource, and initial temperature is slow, and operating cost is high, can not return
The unsafe problem of concealed work leak electric leakage is kept away, trouble and worry is can not to repair.Existing floor heating is all by water or electricity
It flows in the duct, the transmission of heat is realized by the flowing of water, thermal storage time is longer, and traditional idea is must to use water
Or electricity radiates, and the present invention it is innovative provide a kind of heat eliminating medium, by the way that capsule is set as medium, medium
In the move in tracks specially set, so that radiating surface is uniform, and heat transfer is fast, and by using pole to capsule heating unit
Property repel arrangement, realize around track run.However during the image motion of subsurface, due to the inside and outside temperature difference, the temperature of capsule
With the temperature difference in space in track, the camera lens that camera is shot is caused easily to shoot foggy image, lead to the monitoring effect of outside
It is less desirable, it is impossible to achieve the purpose that clear monitoring, it is therefore necessary to propose a kind of clear monitoring process method.
Invention content
Present invention seek to address that above problem of the prior art.Propose it is a kind of it is intelligent, carry high-definition be based on ground
The defogging method of warm capsule moving image.Technical scheme is as follows：
A kind of defogging method based on floor heating capsule moving image, includes the following steps：
The floor heating capsule track includes several capsule heating units, at least one hot glue capsule unit and at least of being not added with
One electronic capsule unit, the capsule heating unit include housing, the first noise elimination mechanism, the shell are provided with outside the housing
It is provided with fever oil for heating in vivo, the first magnetic devices and the second magnetic devices are additionally provided in the housing, if when
When a dry capsule heating unit is navigated in floor heating track, by being arranged using polar repulsion capsule heating unit, i.e., first
During the head arranged adjacent of the tail portion of capsule heating unit and the second capsule heating unit, the tail portion magnetic of the first capsule heating unit
The head magnetic pole of pole and the second capsule heating unit on the contrary, pushed by transmission device to first capsule heating unit,
First capsule heating unit moves forward, remaining capsule heating unit is then pushed to move in orbit successively；It is described to be not added with
Hot glue capsule unit includes hollow housing, the second noise elimination mechanism, camera and control chip, the second noise elimination mechanism, camera
It is arranged on outside hollow housing, movement foggy image D (x) is obtained by camera；The second noise elimination mechanism is rotatable opening
Structure, the control chip are set in hollow housing, and the control chip passes through in the floor heating track that shoots camera
Video carries out defogging processing, stores and carries out real-time display by radioing to external display device；The electronic capsule
Unit uses self-contained power control structure, by the electronic capsule unit) it controls, remaining capsule heating unit is pushed to exist
Move in tracks；The step of control chip carries out defogging processing specifically includes：1) it, obtains in floor heating capsule track first
Move foggy image D (x)；
2), original motion foggy image D (x) is enhanced using the sub-block local histogram equalization algorithm that partly overlaps,
Specifically include step：(1) for a width input picture D (x, y), size is M × N, and it is H (x, y) to define it and export image, is added up
Number is count；
(2) the sliding window f that a size is m × n is definedT, and set its horizontal moving step length with vertical direction and distinguish
For w and h, the initial position of sliding window is located at the image upper left corner；
(3) image currently covered to sliding window carries out histogram equalization, and result is added to correspondence after equilibrium
On pixel, i.e. HT=HT+T(fT), while accumulative frequency adds 1, i.e. countx,y=countx,y+1；
(4) sliding window is continued into the step-length w that moves right, it is no if window without departing from image boundary, goes to step (3)
It then goes in next step；
(5) sliding window is continued to move down step-length h, it is no if window without departing from image boundary, goes to step (3)
It then goes in next step；
(6) value of each pixel divided by corresponding operation times in image will be exported and obtain final output image, i.e.,
3), scheme enhanced image as input, transmissivity is asked for using dark channel prior principle
4), in Markov random field model, to transmissivityMRF model modelings are carried out, i.e., are built using MRF models
The restriction relation formula between the transmittance values of each pixel is found, so as to acquire the transmissivity t (x) after refinement；
5), air light value A is asked for using partitioned searching method；
6), in the above-mentioned transmissivity t (x) asked for and air light value A generations, are returned in foggy image degradation model, inverting obtains most
Whole defogging figure H (x).
Further, it is described control chip include video processing unit, alarm unit, wireless communication module, processor and
Storage unit, the video that the video processing unit is used in the floor heating track shot to camera carry out including denoising, gray scale
Pretreatment including change processing, and be transferred to processor and carry out processing analysis, obtain movement of the capsule unit in floor heating track
Image, the alarm unit are used to detect the heating work state of capsule heating unit, when detecting that capsule heating unit is less than
It is then alarmed by alarm unit during the efficiency value of setting and is handled to processor, the report that the processor sends alarm unit
Alert signal is analyzed and processed, and it is single that the executing agency's selection markers for being not added with hot glue capsule unit is controlled to go out the capsule heating to be detected
Member, takes out the capsule heating unit to be detected, the storage unit treated for storing processor video data.
Further, scheme in the step 3 using enhanced image as input, in conjunction with dark channel prior principle, ask
The transmissivity after initial optimization is taken to specifically include：For a width fog free images, dark channel diagram may be defined as：
In above formula, x be input image pixels point coordinate, Ω (x) be withxCentered on a certain regional area, Hc(x) it is
The intensity value of a certain pixel c channels, therefore dark channel prior principle is utilized, the formula for seeking transmissivity is：
Here for more true scene is restored, picture is made to show Deep Canvas, introduce corrected parameter ω, value
Range is [0,1], and value is smaller, and defog effect gets over unobvious.
Further, it is three notable spies according to sky domain that the step 5 asks for air light value A using partitioned searching method
For sign come what is asked for, there are following three notable features in sky areas：A, brightness value is higher；B, three channel pixel values approach；
C, position is on the upper side, and step is：
(1) atmosphere light region is determined first, chooses the region of image top 25% as the time for asking for air light value
Favored area；
(2) candidate region is divided into four grade blocks, and calculates the average brightness in each block region, then by average value most
Big part continues to be divided into quarter, repeats this process, until the number of pixel in gained block is less than preset threshold
Value；
(3) estimated value of the average brightness as atmosphere light of the sub-block is finally calculated.
Further, in the transmissivity t (x) asked in above-mentioned steps and air light value A generations, are returned to greasy weather figure by the step 6
As degradation model, inverting obtains final defogging figure H (x), specifically includes：Misty Image degradation model is：
D (x)=H (x) t (x)+A (1-t (x))
In above formula, D (x) represents foggy image, and H (x) represents the clear image after defogging, and A represents air luminous intensity, t (x)
Represent transmittivity, when air is uniform, transmissivity is represented by t (x)=e-rd(x), d (x) represent equipment receiving point away from
The distance of scene point, r represent atmospheric scattering coefficient；
Therefore, obtaining final mist elimination image is：
Further, Markov model is non-directed graph in the step 4, and the set Y of stochastic variable is by non-directed graph G institutes
The Markov random field of description, the joint probability distribution of stochastic variable Y is Gibbs distributions at this time, is expressed as：
In above formula,yFor the value of stochastic variable,gTo scheme the group of G,γFor model parameter,μFor normalization coefficient, VgFor a gesture
Function, since potential function is that just, definition group energy function is：
eg(yg| γ)=- logVg(yg|γ)
Then the energy function for obtaining Markov random field is：
By upper, the maximum a-posteriori estimation problem of stochastic variable Y is converted for energy function minimization problem, i.e.,：
Therefore, the cost function of MRF models is expressed as form by us：
Wherein,For single order data item, value is that node i value is yiWhen cost, for weighing stochastic variable yiIt takes
Difference degree between value and observation data；For the smooth item of second order, value is adjacent nodei,jValue is yi,yjWhen generation
Valency, for adjacent node in constraint space value as close possible to.
Further, the inside of the electronic capsule unit also sets up chemical spraying device, and the chemical spraying device includes accommodating medicine
The containers of product, nozzle, spray pipeline and spraying control device, the container are connected by the pipeline that sprays with nozzle, when needing to spray
Then nozzle valve is controlled to open and close by controller during medicine, realize the ejection of liquid.
It advantages of the present invention and has the beneficial effect that：
The present invention provides a kind of defogging method based on floor heating capsule moving image, by innovative by original water
Or electricity is substituted for capsule heating unit, including several capsule heating units, at least one hot glue capsule unit and extremely of being not added with
A few electronic capsule unit realizes the effect of ' locomotive engine ' promotion, and pass through battery and driving by the electronic capsule unit
Unit controls；The capsule heating unit includes housing, and the first noise elimination mechanism is provided with outside the housing, is eliminated the noise by setting
Mechanism can substantially reduce noise caused by due to being careful, and play the function of sound insulation.It is provided in the housing for heating
Fever is oily, and the first magnetic devices and the second magnetic devices are additionally provided in the housing, when several capsule heating units are advanced
When in floor heating track, by being arranged using polar repulsion capsule heating unit, it is preferred to use NN is arranged or SS arrangements, profit
The principle mutually repelled with the same sex, i.e. the head arranged adjacent of the tail portion of the first capsule heating unit and the second capsule heating unit
When, the head magnetic pole of the tail portion magnetic pole of the first capsule heating unit and the second capsule heating unit is on the contrary, pass through transmission device pair
First capsule heating unit is pushed, and first capsule heating unit moves forward, and then remaining capsule is pushed to add successively
Hot cell moves in orbit；The hot glue capsule unit that is not added with includes hollow housing, the second noise elimination mechanism, camera and control core
Piece, the second noise elimination mechanism, camera are arranged on outside hollow housing, and the second noise elimination mechanism is rotatable Unclosing structure,
Camera, control chip etc. can be thus placed, the control chip is set in hollow housing, and the control chip passes through
Video in the floor heating track of camera shooting is handled and analyzed, stored and is filled by radioing to external display
Put carry out real-time display.The capsule heating unit of front is pushed by capsule heating unit below, realizes loopy moving, institute
Noise elimination mechanism is stated using the wear-resisting quieter material including rubber or graphene, noise reduction abrasion resistant effect is good.The control core
Piece includes video processing unit, alarm unit, wireless communication module, processor and storage unit, and the video processing unit is used
The pretreatment including denoising, gray processing processing is carried out in the video in the floor heating track shot to camera, is passed through in this way
Pretreated picture is more clear, and solves the problems, such as that shooting photo is not known, and be transferred at processor in track
Reason analysis, obtains moving image of the capsule unit in floor heating track, and the alarm unit is used to detect capsule heating unit
Heating work state, when detect capsule heating unit less than setting efficiency value when then alarmed by alarm unit to processor
It is handled, the alarm signal that the processor sends alarm unit, which is analyzed and processed and controlled, is not added with hot glue capsule unit
Executing agency's selection markers go out the capsule heating unit to be detected, take out the capsule heating unit to be detected, this addresses the problem
The capsule heating unit working time in track is accurate and the problem of can not find damaged location long afterwards it is possible that damage
It is really positioned and passes through cycle and take out the doubtful damage heating unit, replaced or repaired.The storage unit is used for
Store processor treated video data.The capsule heating unit is provided by being set to external roller transmission mechanism
Power realizes that capsule heating unit pushes front capsule heating unit movement below, then realizes when realization joins end to end and follows
Ring drives.The inside of the electronic capsule unit also sets up chemical spraying device, and the chemical spraying device includes accommodating the container of drug, spray
Mouth, spray pipeline and spraying control device, the container be connected by the pipeline that sprays with nozzle, when need spray when then pass through control
Device control nozzle valve processed is opened and is closed, and realizes the ejection of liquid.Realize indoor disinfection.The spraying control device is also set
Timer is put, by the Time And Frequency of timer setting spray, realizes the uniform disinfection in time uniform space, and can be only
It is vertical to perform heating, perform sprinkling task.The present invention enhances original mist figure using POSHE algorithms, improves final defogging
The brightness of figure；Transmissivity is optimized using Markov random field model, can preferably the detailed information of restored image and holding be schemed
As marginal texture；Air light value is asked for by partitioned searching method, reduces the color error ratio of defogging figure.
Description of the drawings
Fig. 1 is the assembling schematic diagram that the present invention provides preferred embodiment capsule heating unit 21；
Fig. 2 is the schematic diagram of the electronic capsule unit of the preferred embodiment of the present invention；
Fig. 3 is preferred embodiment of the present invention capsule-type floor heating heating unit schematic diagram；
Fig. 4 is preferred embodiment of the present invention control chip circuit connection diagram；
Fig. 5 is defogging flow chart.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, detailed
Carefully describe.Described embodiment is only the part of the embodiment of the present invention.
The present invention solve above-mentioned technical problem technical solution be：
It is as Figure 1-Figure 2 a kind of capsule-type floor heating heating unit with control chip, applied in floor heating track,
It include several capsule heating units, it is at least one be not added with hot glue capsule unit and at least one electronic capsule unit 22, it is described
Capsule heating unit includes housing 14, is provided with the first noise elimination mechanism outside the housing 14, is provided with and is used in the housing 14
The fever oil 15 of heating is additionally provided with the first magnetic devices 14 and the second magnetic devices 16 in the housing 14, when several glue
When capsule heating unit is navigated in floor heating pipeline, by being arranged using polar repulsion capsule heating unit, i.e., the first capsule adds
During the head arranged adjacent of the tail portion of hot cell and the second capsule heating unit, the tail portion magnetic pole of the first capsule heating unit and
The head magnetic pole of two capsule heating units is on the contrary, push first capsule heating unit by transmission device, first
Capsule heating unit moves forward, remaining capsule heating unit is then pushed to move in orbit successively；It is as shown in Fig. 2, described
It is not added with hot glue capsule unit and includes hollow housing, the second noise elimination mechanism, camera and control chip, the second noise elimination mechanism is taken the photograph
As head is arranged on outside hollow housing, the second noise elimination mechanism is rotatable Unclosing structure, and the control chip is set to hollow
In housing, the control chip is handled and is analyzed by the video in the floor heating track that shoots camera, is stored and is led to
It crosses and radios to external display device progress real-time display；；The electronic capsule unit 22 is using self-contained power control knot
Structure by controlling the electronic capsule unit 22, pushes remaining capsule heating unit to move in orbit.
Preferably, as described in Figure 4 control chip include video processing unit 24, alarm unit 25, wireless communication module 26,
Processor 23 and storage unit 27, the video that the video processing unit 24 is used in the floor heating track shot to camera carry out
Pretreatment including denoising, gray processing processing, is more clear by pretreated picture, solves in track and clap in this way
The problem of photo is not known is taken the photograph, and is transferred to processor 23 and carries out processing analysis, obtains fortune of the capsule unit in floor heating track
Motion video, the alarm unit 25 are used to detect the heating work state of capsule heating unit, when detecting capsule heating unit
Less than setting efficiency value when then alarmed by alarm unit 25 and handled to processor 23, the processor 23 is single to alarm
The alarm signal that member 25 is sent is analyzed and processed, and it is to be checked that the executing agency's selection markers for being not added with hot glue capsule unit is controlled to go out this
Capsule heating unit is surveyed, takes out the capsule heating unit to be detected, this addresses the problem the capsule heating unit work in track
It is possible that damage after of long duration, and the problem of damaged location can not be found, accurately positioned and pass through cycle and taken out
The doubtful damage heating unit, is replaced or is repaired.The storage unit treated for storing processor video counts
According to.
Preferably, the first noise elimination mechanism and the second noise elimination mechanism are the round shape ring junction being set to outside housing 14
Structure.
Preferably, the first noise elimination mechanism and the second noise elimination mechanism are using wear-resisting including rubber or graphene
Quieter material is good using such material noise reduction abrasion resistant effect.
Preferably, it is convenient before the one of capsule heating unit for being not added with hot glue capsule unit and being arranged in track
Shooting.
Preferably, the first magnetic devices 12 of the housing 14 of the capsule heating unit are by the first connection fixing device
11 is fixed install, and second magnetic devices 16 are to carry out installation fixation by the second connection fixing device 17
's.
Preferably, the first connection fixing device 11 and second is connected fixing device 17 and is carried out using bolt arrangement
Connection is fixed, and spinning in and out are realized by the bolt arrangement.
Preferably, the capsule heating unit provides power by being set to external gear drive, after realization
Face capsule heating unit driving front capsule heating unit movement, mutual driving is then realized when realization joins end to end.Fig. 3
In, 1 the first revetment of expression, 8 represent tracks, the second revetment 3, and the track 8 includes straight trip track and track of turning, described first
Space between 1 and second revetment 3 of revetment is straight trip track, and second revetment 3 combines revetment for extension type, by described
Mobile device carries out the second revetment 3 the mobile third revetment 11-1 formed for bend；11-2, the third revetment 11-1；
11-2 and the first revetment 1, the second revetment 3 form the turning track of track.
Preferably, by storage treated video data, module 26 is sent to far the storage unit by radio communication
Journey equipment realizes long-range detection in real time；
Preferably, the inside of the electronic capsule unit 22 also sets up chemical spraying device, and the chemical spraying device includes accommodating medicine
The containers of product, nozzle, spray pipeline and spraying control device, the container are connected by the pipeline that sprays with nozzle, when needing to spray
Then nozzle valve is controlled to open and close by controller during medicine, realize the ejection of liquid.
Preferably, the spraying control device also sets up timer, passes through the Time And Frequency of timer setting spray.
Preferably, the bottom of the electronic capsule unit 22 is additionally provided with the driving dress of idler wheel and the driving idler wheel
Put, the idler wheel driven to run by driving device, the driving device by be set to the battery in electronic capsule unit into
Row power supply.
As shown in figure 5, the execution flow chart of the present invention is as shown in figure 5, its specific technical solution is as follows：
1st, original foggy image D (x) is obtained.
2nd, original foggy image is enhanced using POSHE algorithms, makes that the histogram distribution of mist figure is more balanced, dynamic model
Expansion is enclosed, in order to the processing of follow-up defogging.The execution step of POSHE algorithms is as follows：
(1) for a width input picture D (x, y), size is M × N, defines it and exports image as H (x, y), accumulative frequency
Forcount；
(2) the sliding window f that a size is m × n is definedT, and set its horizontal moving step length with vertical direction and distinguish
For w and h, the initial position of sliding window is located at the image upper left corner；
(3) image currently covered to sliding window carries out histogram equalization, and result is added to correspondence after equilibrium
On pixel, i.e. HT=HT+T(fT), while accumulative frequency adds 1, i.e. countx,y=countx,y+1；
(4) sliding window is continued into the step-length w that moves right, if window without departing from image boundary, jumps to step (3),
Otherwise it goes in next step；
(5) sliding window is continued to move down step-length h, if window without departing from image boundary, jumps to step (3),
Otherwise it goes in next step；
(6) value of each pixel divided by corresponding operation times in image will be exported and obtain final output image, i.e.,
3rd, scheme enhanced image as input, the transmissivity after optimization is asked for using dark channel prior principleIt is right
In a width open air fog free images, dark channel diagram may be defined as：
In above formula, x is the coordinate of input image pixels point, and Ω (x) is a certain regional area centered on x, Hc(x) it is
The intensity value of a certain pixel c channels.Therefore using dark channel prior principle, the formula for seeking transmissivity is：
Here for more true scene is restored, picture is made to show Deep Canvas, introduce corrected parameter ω, value
Range is [0,1], and value is smaller, and defog effect gets over unobvious.In the present invention, ω takes 0.95.
4th, in Markov random field model, t (~x) is modeled, the transmissivity t after further refinement can be acquired
(x).Markov model is non-directed graph, the set Y of stochastic variable be by the described Markov random fields of non-directed graph G, this
When stochastic variable Y joint probability distribution be Gibbs distributions, be expressed as：
In above formula,yFor the value of stochastic variable,gTo scheme the group of G,γFor model parameter,μFor normalization coefficient, VgFor a gesture
Function.Since potential function is just, we are at the group's of definition energy function：
eg(yg| γ)=- logVg(yg|γ)
Then the energy function for obtaining Markov random field is：
By upper, the maximum a-posteriori estimation problem of stochastic variable Y is converted for energy function minimization problem, i.e.,：
Therefore, the cost function of MRF models is expressed as form by us：
Wherein,For single order data item, value is that node i value is yiWhen cost, for weighing stochastic variable yiIt takes
Difference degree between value and observation data；For the smooth item of second order, value is that adjacent node i, j value are yi,yjWhen
Cost, for adjacent node in constraint space value as close possible to.Image defogging can rely on MRF model foundations adjacent to picture
Relationship between element, so as to achieve the purpose that transmissivity optimizes.When optimizing transmissivity with MRF, mainly consider in terms of two：
First, constraint t (x) and the estimated value asked for beforeBetween relationship, ensure that error between the two is small as far as possible；Second is that
The gradient of t (x) is constrained, for describing the relationship in field between pixel.It is embodied in shown in following energy function：
Wherein,
This is data item, describes the difference between the true value of transmissivity and estimated value so that estimated value is more nearly very
Real value.
This is smooth item, describes the difference between the pixel of field so that the transmissivity variation after optimization is more smooth.
The transmissivity after final refinement can be acquired as a result,.The transmittance figure obtained by this method becomes at depth of field mutation
Must be more smooth, CONSTRUCTED SPECIFICATION information is more abundant.
5th, according to the three of sky domain notable features air light value A is asked for using partitioned searching method.Sky areas has following
Three notable features：A, brightness value is higher；B, three channel pixel values approach；C, position is on the upper side.With reference to this feature, using as follows
Partitioned searching method is asked for：
(1) atmosphere light region is determined first.Since infinite point is normally above image, we choose figure
As the region of top 25% is as the candidate region for asking for air light value；
(2) candidate region is divided into four grade blocks, and calculates the average brightness in each block region, then by average value most
Big part continues to be divided into quarter, repeats this process, until the number of pixel in gained block is less than preset threshold
Value；
(3) estimated value of the average brightness as atmosphere light of the sub-block is finally calculated.
As above method can balance out the highlight regions of non-air light source and the influence of large area white area so that institute
Ask air light value more accurate, defogging recovery effect is more natural.
6th, in the transmissivity t (x) asked in above-mentioned steps and air light value A generations, are returned into Misty Image degradation model, inverting
Obtain final defogging figure H (x).The degradation model is expressed as：
D (x)=H (x) t (x)+A (1-t (x))
In above formula, D (x) represents foggy image, and H (x) represents the clear image after defogging, and A represents air luminous intensity, t (x)
Represent transmittivity, when air is uniform, transmissivity is represented by t (x)=e-rd(x), d (x) represent equipment receiving point away from
The distance of scene point, r represent atmospheric scattering coefficient.
Therefore, obtaining final mist elimination image is：
When the ratio of pixel component value and air luminous intensity is more than 1, transmissivity0 may be less than, this can make image
Comprising noise, therefore set a threshold value t0If transmissivity is acquired less than t0, enable t=t0, then above formula can be modified to：
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.
After the content for having read the record of the present invention, technical staff can make various changes or modifications the present invention, these equivalent changes
Change and modification equally falls into the scope of the claims in the present invention.
The preferred embodiment of the present invention described in detail above.It should be appreciated that those of ordinary skill in the art without
Creative work is needed according to the present invention can to conceive and makes many modifications and variations.Therefore, all technologies in the art
Personnel are available by logical analysis, reasoning, or a limited experiment on the basis of existing technology under this invention's idea
Technical solution, all should be in the protection domain being defined in the patent claims.
Claims (7)
1. a kind of defogging method based on floor heating capsule moving image, which is characterized in that include the following steps：
The floor heating capsule track includes several capsule heating units, at least one hot glue capsule unit and at least one of being not added with
Electronic capsule unit (22), the capsule heating unit include housing (14), the first noise damping device are provided with outside the housing (14)
Structure, the housing (14) is interior to be provided with fever oil (15) for heating, and the first magnetism is additionally provided in the housing (14) and is filled
(14) and the second magnetic devices (16) are put, when several capsule heating units are navigated in floor heating track, by adding to capsule
Hot cell is arranged using polar repulsion, i.e. the head adjacent row of the tail portion of the first capsule heating unit and the second capsule heating unit
During row, the head magnetic pole of the tail portion magnetic pole of the first capsule heating unit and the second capsule heating unit is on the contrary, pass through transmission device
First capsule heating unit is pushed, first capsule heating unit moves forward, and then pushes remaining capsule successively
Heating unit moves in orbit；The hot glue capsule unit that is not added with includes hollow housing, the second noise elimination mechanism, camera and control
Chip, the second noise elimination mechanism, camera are arranged on outside hollow housing, and movement foggy image D (x) is obtained by camera；
The second noise elimination mechanism is rotatable Unclosing structure, and the control chip is set in hollow housing, and the control chip leads to
The video crossed in the floor heating track for shooting camera carries out defogging processing, stores and is filled by radioing to external display
Put carry out real-time display；The electronic capsule unit (22) is using self-contained power control structure, by the electronic capsule unit
(22) it is controlled, remaining capsule heating unit is pushed to move in orbit；The step of control chip carries out defogging processing
It specifically includes：1) the movement foggy image D (x) in floor heating capsule track, is obtained first；
2), original motion foggy image D (x) is enhanced using the sub-block local histogram equalization algorithm that partly overlaps, specifically
Including step：(1) for a width input picture D (x, y), size is M × N, defines it and exports image as H (x, y), accumulative frequency
For count；
(2) the sliding window f that a size is m × n is definedT, and set its horizontal moving step length with vertical direction be respectively w and
H, the initial position of sliding window are located at the image upper left corner；
(3) image currently covered to sliding window carries out histogram equalization, and result is added to respective pixel after equilibrium
On point, i.e. HT=HT+T(fT), while accumulative frequency adds 1, i.e. countx,y=countx,y+1；
(4) sliding window is continued into the step-length w that moves right, if window goes to step (3), otherwise turn without departing from image boundary
To in next step；
(5) sliding window is continued to move down step-length h, if window goes to step (3), otherwise turn without departing from image boundary
To in next step；
(6) value of each pixel divided by corresponding operation times in image will be exported and obtain final output image, i.e.,
3), scheme enhanced image as input, transmissivity is asked for using dark channel prior principle
4), in Markov random field model, to transmissivityMRF model modelings are carried out, that is, utilize each picture of MRF model foundations
Restriction relation formula between the transmittance values of vegetarian refreshments, so as to acquire the transmissivity t (x) after refinement；
5), air light value A is asked for using partitioned searching method；
6), in the above-mentioned transmissivity t (x) asked for and air light value A generations, are returned in foggy image degradation model, inverting is finally gone
Mist figure H (x).
2. the defogging method according to claim 1 based on floor heating capsule moving image, which is characterized in that the control core
Piece includes video defogging processing unit (24), alarm unit (25), wireless communication module (26), processor (23) and storage unit
(27), the video defogging processing unit (24) for the video in the floor heating track that is shot to camera include denoising,
Pretreatment including gray processing processing, and be transferred to processor (23) and carry out processing analysis, capsule unit is obtained in floor heating track
Interior moving image, the alarm unit (25) is for detecting the heating work state of capsule heating unit, when detecting capsule
It is then alarmed by alarm unit (25) when heating unit is less than the efficiency value set and is handled to processor (23), the processing
The alarm signal that device (23) sends alarm unit (25) analyzes and processes, and controls the executing agency for being not added with hot glue capsule unit
Selection markers go out the capsule heating unit to be detected, take out the capsule heating unit to be detected, and the storage unit is used to store
Processor treated video data.
3. the defogging method according to claim 1 based on floor heating capsule moving image, which is characterized in that the step 3
It is middle to scheme enhanced image as input, in conjunction with dark channel prior principle, ask for the transmissivity after initial optimization and specifically wrap
It includes：For a width fog free images, dark channel diagram may be defined as：
In above formula, x is the coordinate of input image pixels point, and Ω (x) is a certain regional area centered on x, Hc(x) it is a certain
The intensity value of pixel c channels, therefore dark channel prior principle is utilized, the formula for seeking transmissivity is：
Here for more true scene is restored, picture is made to show Deep Canvas, introduce corrected parameter ω, value range
It is [0,1], value is smaller, and defog effect gets over unobvious.
4. the defogging method based on floor heating capsule moving image according to one of claim 1-3, which is characterized in that described
Step 5 asks for air light value A using partitioned searching method to be asked for according to three notable features of sky domain, day dead zone
There are following three notable features in domain：A, brightness value is higher；B, three channel pixel values approach；C, position is on the upper side, and step is：
(1) atmosphere light region is determined first, chooses the region of image top 25% as the candidate regions for asking for air light value
Domain；
(2) candidate region is divided into four grade blocks, and calculates the average brightness in each block region, then average value is maximum
Part continues to be divided into quarter, repeats this process, until the number of pixel in gained block is less than preset threshold value；
(3) estimated value of the average brightness as atmosphere light of the sub-block is finally calculated.
5. the defogging method according to claim 1 based on floor heating capsule moving image, which is characterized in that the step 6
In the transmissivity t (x) asked in above-mentioned steps and air light value A generations, are returned into Misty Image degradation model, inverting is finally gone
Mist figure H (x) is specifically included：Misty Image degradation model is：
D (x)=H (x) t (x)+A (1-t (x))
In above formula, D (x) represents foggy image, and H (x) represents the clear image after defogging, and A represents air luminous intensity, and t (x) is represented
Transmittivity, when air is uniform, transmissivity is represented by t (x)=e-rd(x), d (x) expression equipment receiving points are away from scene
The distance of point, r represent atmospheric scattering coefficient；
Therefore, obtaining final mist elimination image is：
6. the defogging method according to claim 1 based on floor heating capsule moving image, which is characterized in that the step 4
Middle Markov model is non-directed graph, and the set Y of stochastic variable is by the described Markov random fields of non-directed graph G, at this time
The joint probability distribution of stochastic variable Y is distributed for Gibbs, is expressed as：
In above formula, y is the value of stochastic variable, and g be the group of figure G, and γ is model parameter, and μ is normalization coefficient, VgFor a gesture letter
Number, since potential function is that just, definition group energy function is：
eg(yg| γ)=- logVg(yg|γ)
Then the energy function for obtaining Markov random field is：
By upper, the maximum a-posteriori estimation problem of stochastic variable Y is converted for energy function minimization problem, i.e.,：
Therefore, the cost function of MRF models is expressed as form by us：
Wherein,For single order data item, value is that node i value is yiWhen cost, for weighing stochastic variable yiValue with
Observe the difference degree between data；For the smooth item of second order, value is that adjacent node i, j value are yi,yjWhen cost,
For adjacent node in constraint space value as close possible to.
7. the defogging method according to claim 1 based on floor heating capsule moving image, which is characterized in that the electric colloidal
The inside of capsule unit (22) also sets up chemical spraying device, the chemical spraying device include accommodating the container of drug, nozzle, spray pipeline and
Spraying control device, the container are connected by the pipeline that sprays with nozzle, when need spray when nozzle is then controlled by controller
Valve is opened and is closed, and realizes the ejection of liquid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810182947.7A CN108198153A (en) | 2018-03-06 | 2018-03-06 | A kind of defogging method based on floor heating capsule moving image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810182947.7A CN108198153A (en) | 2018-03-06 | 2018-03-06 | A kind of defogging method based on floor heating capsule moving image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108198153A true CN108198153A (en) | 2018-06-22 |
Family
ID=62594658
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810182947.7A Pending CN108198153A (en) | 2018-03-06 | 2018-03-06 | A kind of defogging method based on floor heating capsule moving image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108198153A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108061334A (en) * | 2018-02-09 | 2018-05-22 | 重庆中贝毅科科技有限公司 | Floor heating track controls command device with ground |
| CN115170443A (en) * | 2022-09-08 | 2022-10-11 | 荣耀终端有限公司 | Image processing method, shooting method and electronic equipment |
-
2018
- 2018-03-06 CN CN201810182947.7A patent/CN108198153A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108061334A (en) * | 2018-02-09 | 2018-05-22 | 重庆中贝毅科科技有限公司 | Floor heating track controls command device with ground |
| CN115170443A (en) * | 2022-09-08 | 2022-10-11 | 荣耀终端有限公司 | Image processing method, shooting method and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Manivasagam et al. | Lidarsim: Realistic lidar simulation by leveraging the real world | |
| CN108830150A (en) | One kind being based on 3 D human body Attitude estimation method and device | |
| CN108198153A (en) | A kind of defogging method based on floor heating capsule moving image | |
| CN107507126A (en) | A kind of method that 3D scenes are reduced using RGB image | |
| CN109711261A (en) | A kind of indoor occupant localization method and device, building intelligence lighting system | |
| Saroya et al. | Online exploration of tunnel networks leveraging topological CNN-based world predictions | |
| Xu et al. | Real-time detecting method of marine small object with underwater robot vision | |
| CN112949452B (en) | Robot low-light environment grabbing detection method based on multitask shared network | |
| CN107833239A (en) | A kind of searching of optimal matching method for tracking target based on weighted model constraint | |
| CN108596243A (en) | The eye movement for watching figure and condition random field attentively based on classification watches figure prediction technique attentively | |
| CN102999923A (en) | Motion capture data key frame extraction method based on adaptive threshold | |
| Yang et al. | Embodied visual recognition | |
| Tung et al. | 3d-oes: Viewpoint-invariant object-factorized environment simulators | |
| CN106872943A (en) | A kind of mobile object indoor tracking and positioning method based on visible light communication | |
| Cheng et al. | The design and application of a track-type autonomous inspection robot for electrical distribution room | |
| Doretto et al. | Dynamic shape and appearance models | |
| Li et al. | Blinkflow: A dataset to push the limits of event-based optical flow estimation | |
| Wu et al. | Multi-modal visual place recognition in dynamics-invariant perception space | |
| Li et al. | 3D semantic mapping based on convolutional neural networks | |
| CN108665424A (en) | The moving image defogging method of capsule-type ground heating system | |
| Hu et al. | Modeling realistic 3D agricultural vegetations using a photometric-based approach and its application to weed detection | |
| CN108765300A (en) | The capsule-type floor heating infrared image enhancing method of view-based access control model contrast resolution | |
| Mahé et al. | Semantic-only Visual Odometry based on dense class-level segmentation | |
| Mousavi et al. | OpenWaters: Photorealistic simulations for underwater computer vision | |
| Zhao et al. | Toward the resilient SLAM in adverse illumination environments with learning-based image transformation |
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: 20180622 |