CN110223258A - A kind of multi-mode fast video image defogging method and device - Google Patents
A kind of multi-mode fast video image defogging method and device Download PDFInfo
- Publication number
- CN110223258A CN110223258A CN201910508122.4A CN201910508122A CN110223258A CN 110223258 A CN110223258 A CN 110223258A CN 201910508122 A CN201910508122 A CN 201910508122A CN 110223258 A CN110223258 A CN 110223258A
- Authority
- CN
- China
- Prior art keywords
- video image
- module
- image data
- video
- dark
- 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 32
- 238000010586 diagram Methods 0.000 claims description 52
- 239000011159 matrix material Substances 0.000 claims description 25
- 238000011084 recovery Methods 0.000 claims description 19
- 230000004927 fusion Effects 0.000 claims description 6
- 238000006073 displacement reaction Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract 2
- 238000003909 pattern recognition Methods 0.000 abstract 2
- 238000000053 physical method Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000015271 coagulation Effects 0.000 description 2
- 238000005345 coagulation Methods 0.000 description 2
- 230000009977 dual effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 239000003595 mist Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000005202 decontamination Methods 0.000 description 1
- 230000003588 decontaminative effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000007787 solid Substances 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Abstract
The invention discloses a kind of multi-mode fast video image defogging method and devices, comprising: the first image collecting device;Second image collecting device;First object image information inputting apparatus;Second target image information input unit;First pattern recognition device;Second pattern recognition device;Image comparison analytical equipment.The application of the present invention directly acquires image information of the height with double heels by the first image collecting device and the second image collecting device, and after carrying out corresponding image recognition, obtain recognition result, it realizes spring heel height and matches double automatic detections, and, it is blocked by fixture the partial region that fixture clamps heel, the information of this partial region can not be collected, the pressure information in the region is acquired by first pressure sensor array and second pressure sensor array, and then judge the practical height in the region with Double Data, so that the spring heel detected in a kind of multi-mode fast video image defogging method and device is without detection dead angle, further reduce error.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a kind of multi-mode fast video image defogging method and dress
It sets.
Background technique
The video image method that outdoor monitoring system obtains is easy to be influenced by outdoor weather, and wherein mist and haze are that China is normal
The weather phenomenon seen.Image acquired in the greasy weather, since there are a large amount of media to seriously affect light in muddy atmosphere
Scattering and absorb, cause the contrast of image to reduce, reduced dynamic range, clarity it is inadequate, the detailed information of image is unobvious,
Many features are capped or fuzzy, increase the difficulty for extracting information.
Physical method and non-physical method are broadly divided into the method for outdoor monitoring video sharpening at present, such as in camera
Preceding to add polarizing film again, using using software mode to realize in PC machine, defogging is handled, but these two kinds of methods are adaptable weak, defogging
The problems such as time is long, and recovery effect is not significant.
Therefore, it is proposed to which a kind of demister based on programming device, can carry out video with 80 frames/second speed
Defogging processing, delay time only 3us.
Summary of the invention
It is an object of that present invention to provide a kind of multi-mode fast video image defogging method and devices, for solving existing skill
Physical method and non-physical method are broadly divided into the method for outdoor monitoring video sharpening in art, it is such as before camera plus inclined again
Shake piece, and using using software mode to realize in PC machine, defogging is handled, but these two kinds of methods are adaptable weak, and the defogging time is long,
The inapparent problem of recovery effect.
To achieve the above object, the technical scheme adopted by the invention is that:
A kind of multi-mode fast video image demister, comprising: video frequency collection card, central processing unit, external storage
Device, video output display device and mode selection button;The central processing unit is deposited with the video frequency collection card, outside respectively
Reservoir, video output display device are connected with mode selection button;
The video frequency collection card carries out video image data is realized in coding fusion for acquiring video image data
Coagulation;
The external memory is used for store video images data;
The video output display device is for exporting display video image data;
The mode selection button is used to select the operating mode of the multi-mode fast video image demister;
The central processing unit includes dark algoritic module, Steerable filter module, atmosphere light according to A module, recovery calculation
Method module and color correction module;Wherein, the video image data of the video frequency collection card acquisition is transmitted to the dark and calculates
Method resume module;By the dark algoritic module treated video image data, be on the one hand be directly transferred to it is described
Color correction module is handled, and is on the other hand to transmit again by the atmosphere light according to A module or the restoration algorithm module
It is handled to the color correction module;Video image data after the color correction resume module is the centre
Manage the video image data of unit output.
Preferably, the video frequency collection card uses OV5640 camera.
Preferably, the external memory use two panels model Micron, bus width for 32bit, size 4GB
MT41J256M16HA-125 composition, highest data rate 1066Mbps.
Preferably, the video display is connect using HDMI interface with the output end of the central processing unit.
Preferably, the work step of the dark algoritic module are as follows:, will be next after updating dark channel diagram order and arriving
Frame is as key frame, and the minimum value of each pixel in solution tri- channels RGB is as minimum value figure first, then by minimum value figure
It is input in matrix generator, generates the matrix of 5*5, by solving minimum value twice in succession, obtain the minimum of entire 5*5 matrix
Value, generates dark channel diagram, and model is
Preferably, the work step of the Steerable filter module are as follows: generated using pipeline system and 4*4 matrix generator
4*4 matrix seeks the sum of every a line first, then in the sum for seeking continuous four column, solves to obtain finally by the mode of displacement
Value;Using model be Q=a*I+b, algorithm realize dark refinement, a and b are parameter, and I is guide image, whereinThe regular parameter ε introduced during solving a, ε solve model be ε=|
var(Ik)-var(Gk)|。
Preferably, work step of the atmosphere light according to A module are as follows: during solving atmosphere light according to A, when atmosphere light is shone
More newer command arrive when, solve maximum pixel of the maximum value of every a line as every a line in the dark channel diagram of crucial frame
Point is maximized as a whole big after the maximum value of last line, which solves, to be come from every a line maximum pixel point value
Gas illumination.
Preferably, the work step of the restoration algorithm module are as follows: by by dark channel diagram after Steerable filter module and
Foggy image after delay is input in restoration module, realizes each pixel to three channels using parallel and pipelining
Recovery operation is carried out, wherein restoration model is
Preferably, the work step of the color correction module are as follows: image carries out micronization processes to treated, right first
The pixel point value in each channel is pre-processed, i.e., takes 255 when pixel value is greater than 255, takes zero when pixel value is negative;
Then do to the pixel of three channel RGB of restored image enhances in proportion, obtains final restored image.
A kind of multi-mode fast video image defogging method, comprising the following steps:
S1: acquisition video image data, and carry out the first order processing that video image data is realized in coding fusion;
S2: dark channel diagram processing will be updated by the video image data of first order processing;
S3: Steerable filter processing will be carried out by the video image data for updating dark channel diagram processing processing;
S4: atmosphere lighting process will be updated by the video image data of Steerable filter processing;
S5: recovery operation will be carried out by the video image data for updating atmosphere lighting process;
S6: color correction processing will be carried out by the video image data of recovery operation;
S7: the video image data output by color correction processing is displayed and store.
The method have the benefit that: (1) DDR3 storage chip is connected the high-speed interface of ZYNQ chip by this programme
Place is reduced since circulation reads delay caused by data.The frame speed for carrying out video image defogging is 80 frames/second, the function of complete machine
Rate is less than 10W.
(2) dark channel diagram is brought directly in restoration model, reduce due to transmission plot solve bring time delay with
And the waste of buffer zone, reach simplified decontamination mist model, improve the flexibility ratio of algorithm, adaptability, reduce defogging delay time,
Present apparatus delay time is 2.075us.
(3) optimize down-sampling technology, reduce and cache waste as caused by caching dark channel diagram, skill is restored by circulation
Art is realized the fast reading and writing operation of dark channel diagram, reduces delay caused by being sought as dark, the present apparatus is being sought helping secretly
The delay time of road figure is 0.05us.
(4) using SoPC chip as central processing element, FPGA hardware exploitation, which both may be implemented, can also pass through load
System on chip realizes other image procossings, establishes solid foundation for follow-up system upgrading exploitation.
Detailed description of the invention
Fig. 1 is shown as the structural schematic diagram of one embodiment of the present of invention.
Fig. 2 is shown as the central processing unit structural schematic diagram of one embodiment of the present of invention.
The Steerable filter module that Fig. 3 is shown as one embodiment of the present of invention realizes block diagram.
Fig. 4 is shown as the workflow schematic diagram of one embodiment of the present of invention.
Specific embodiment
Below with reference to attached drawing 1-3 of the invention, technical solution in the embodiment of the present invention is clearly and completely retouched
It states, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based on the present invention
In embodiment, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
Embodiment 1:
As depicted in figs. 1 and 2, a kind of multi-mode fast video image demister, comprising: video frequency collection card, centre
Manage unit, external memory, video output display device and mode selection button;The central processing unit respectively with the view
Frequency capture card, external memory, video output display device are connected with mode selection button;
The video frequency collection card carries out video image data is realized in coding fusion for acquiring video image data
Coagulation;More accurate data are provided for the processing of subsequent video;
Central processing unit is mainly made of the ZYNQ chip of Xilinx and its surrounding attached circuit, main to complete to video
The selection of defogging processing, the read-write of image data and mode of image;
The external memory is used for store video images data;
The video output display device is for exporting display video image data;
The mode selection button is used to select the operating mode of the multi-mode fast video image demister;This part
It is made of two keys, the work mainly completed is that A key is realized video defogging, picture defogging and do not carried out at defogging to image
Three models selection is managed, B key realizes the selection to video output size;
The central processing unit includes dark algoritic module, Steerable filter module, atmosphere light according to A module, recovery calculation
Method module and color correction module;Wherein, the video image data of the video frequency collection card acquisition is transmitted to the dark and calculates
Method resume module;By the dark algoritic module treated video image data, be on the one hand be directly transferred to it is described
Color correction module is handled, and is on the other hand to transmit again by the atmosphere light according to A module or the restoration algorithm module
It is handled to the color correction module;Video image data after the color correction resume module is the centre
Manage the video image data of unit output.
Preferably, the video frequency collection card uses OV5640 camera.
Preferably, the external memory use two panels model Micron, bus width for 32bit, size 4GB
MT41J256M16HA-125 composition, highest data rate 1066Mbps.
Preferably, the video display is connect using HDMI interface with the output end of the central processing unit.
Preferably, the work step of the dark algoritic module are as follows:, will be next after updating dark channel diagram order and arriving
Frame is as key frame, and the minimum value of each pixel in solution tri- channels RGB is as minimum value figure first, then by minimum value figure
It is input in matrix generator, generates the matrix of 5*5, by solving minimum value twice in succession, obtain the minimum of entire 5*5 matrix
Value, generates dark channel diagram, and model isDue to needing to be deposited as key message in dark channel diagram
Storage, but dark channel diagram storage will cause and significantly postpone into external memory.Therefore, by the central point of 5*5 matrix
Crucial pixel storage as matrix can greatly reduce the consumption stored in piece, finally by each picture to dual port RAM
Vegetarian refreshments value generates the matrix of a 5*5, ultimately produces complete dark channel diagram.There are two RAM of the same size (A and B) conducts
Dark channel diagram is stored, the value in A memory is read when more newer command does not arrive as dark channel diagram, when more newer command
After arrival, dark channel diagram participates in subsequent recovery operation using the value in B memory as dark channel diagram, while updating A storage
Value in device stores 1/4 summation of the general and original value of the value newly solved into A as new dark channel value.When more
After newer command is completed, the value of A memory is read as new dark channel diagram, while updating the value in B memory.Guarantee obtains
Dark channel diagram be more nearly the dark channel diagram in fogless situation, and delay time is 0.05us.
Preferably, as shown in figure 3, the work step of the Steerable filter module are as follows: use pipeline system and 4*4 matrix
Generator generates 4*4 matrix, seeks the sum of every a line first, then in the sum for seeking continuous four column, finally by the mode of displacement
Solution obtains mean value;Be Q=a*I+b using model, algorithm realize the refinement of dark, a and b are parameter, and I is guidance
Image, whereinThe regular parameter ε, ε introduced during solving a solves mould
Type be ε=| var (Ik)-var(Gk)|.The solution procedure for simplifying regular parameter, the dark channel diagram details after also ensuring optimization
Information is more abundant.
Preferably, work step of the atmosphere light according to A module are as follows: during solving atmosphere light according to A, when atmosphere light is shone
More newer command arrive when, solve maximum pixel of the maximum value of every a line as every a line in the dark channel diagram of crucial frame
Point is maximized as a whole big after the maximum value of last line, which solves, to be come from every a line maximum pixel point value
Gas illumination participates in subsequent image restoration.
Preferably, the work step of the restoration algorithm module are as follows: by by dark channel diagram after Steerable filter module and
Foggy image after delay is input in restoration module, realizes each pixel to three channels using parallel and pipelining
Recovery operation is carried out, wherein restoration model isReduce processing consumption when
Between, entire delay time is 0.97us.
Preferably, the work step of the color correction module are as follows: image carries out micronization processes to treated, right first
The pixel point value in each channel is pre-processed, i.e., takes 255 when pixel value is greater than 255, takes zero when pixel value is negative;
Then do to the pixel of three channel RGB of restored image enhances in proportion, obtains final restored image.
Embodiment 2:
As shown in figure 4, a kind of multi-mode fast video image defogging method, comprising the following steps:
S1: acquisition video image data, and carry out the first order processing that video image data is realized in coding fusion;
S2: it will be updated dark channel diagram processing by the video image data of first order processing, is updating dark channel diagram
In processing, after updating dark channel diagram order arrival, using next frame as key frame, each pixel in tri- channels RGB is solved first
Then minimum value figure is input in matrix generator as minimum value figure, generates the matrix of 5*5, pass through company by the minimum value of point
It is continuous to solve minimum value twice, the minimum value of entire 5*5 matrix is obtained, dark channel diagram is generated, model isIt stores due to needing to be stored as key message in dark channel diagram, but by dark channel diagram to outer
In portion's memory, it will cause and significantly postpone.Therefore, it is deposited the central point of 5*5 matrix as the crucial pixel of matrix
Dual port RAM is stored up, the consumption stored in piece can be greatly reduced, each pixel point value is finally generated to the square of a 5*5
Battle array, ultimately produces complete dark channel diagram.There are two RAM of the same size (A and B) as storage dark channel diagram, orders when updating
Enable value when arrival in reading A memory as dark channel diagram, after more newer command arrives, dark channel diagram is with B
Value in memory participates in subsequent recovery operation as dark channel diagram, while updating the value in A memory, the value that will newly solve
General and original value 1/4 summation as new dark channel value storage into A.After more newer command is completed, reads A and deposit
The value of reservoir updates the value in B memory as new dark channel diagram.It is fogless to guarantee that obtained dark channel diagram is more nearly
In the case of dark channel diagram, and delay time be 0.05us;
S3: Steerable filter processing will be carried out by the video image data for updating dark channel diagram processing processing, is being led
Into filtering processing, 4*4 matrix is generated using pipeline system and 4*4 matrix generator, seeks the sum of every a line first, then
In the sum for seeking continuous four column, solve to obtain mean value finally by the mode of displacement;Be Q=a*I+b using model, algorithm it is real
The refinement of existing dark, a and b are parameter, and I is guide image, wherein?
The regular parameter ε introduced during solving a, ε solve model be ε=| var (Ik)-var(Gk)|.Simplify asking for regular parameter
Solution preocess, the dark channel diagram detailed information after also ensuring optimization are more abundant;
S4: will be updated atmosphere lighting process by the video image data of Steerable filter processing, big being updated
In gas lighting process, when atmosphere light arrives according to more newer command, every a line is solved most in the dark channel diagram of crucial frame
Maximum pixel point of the big value as every a line, after the maximum value of last line, which solves, to be come, from every a line maximum pixel point
The atmosphere light being maximized in value as a whole is shone, and subsequent image restoration is participated in;
S5: recovery operation will be carried out by the video image data for updating atmosphere lighting process, is carrying out recovery behaviour
In work, it will be input in restoration module, adopted by the foggy image after the dark channel diagram and delay after Steerable filter module
Recovery operation is carried out to each pixel in three channels with the realization of parallel and pipelining, wherein restoration model isIt reduces the time of the consumption of processing, entire delay time is 0.97us;
S6: will carry out color correction processing by the video image data of recovery operation, in carrying out color correction processing,
To treated, image carries out micronization processes, pre-processes first to the pixel point value in each channel, i.e., when pixel value is big
255 are taken when 255, takes zero when pixel value is negative;Then the pixel of three channel RGB of restored image is done in proportion
Enhancing, obtains final restored image;
S7: the video image data output by color correction processing is displayed and store.
In the description of the present invention, it is to be understood that, term " counterclockwise ", " clockwise " " longitudinal direction ", " transverse direction ",
The orientation of the instructions such as "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outside" or
Positional relationship is to be based on the orientation or positional relationship shown in the drawings, and is merely for convenience of the description present invention, rather than is indicated or dark
Show that signified device or element must have a particular orientation, be constructed and operated in a specific orientation, therefore should not be understood as pair
Limitation of the invention.
Claims (10)
1. a kind of multi-mode fast video image demister characterized by comprising video frequency collection card, central processing unit,
External memory, video output display device and mode selection button;The central processing unit respectively with the video acquisition
Card, external memory, video output display device are connected with mode selection button;
The video frequency collection card carries out the first order that video image data is realized in coding fusion for acquiring video image data
Processing;
The external memory is used for store video images data;
The video output display device is for exporting display video image data;
The mode selection button is used to select the operating mode of the multi-mode fast video image demister;
The central processing unit includes dark algoritic module, Steerable filter module, atmosphere light according to A module, restoration algorithm mould
Block and color correction module;Wherein, the video image data of the video frequency collection card acquisition is transmitted to the dark algorithm mould
Block processing;It is on the one hand to be directly transferred to the color by the dark algoritic module treated video image data
Correction module is handled, and is on the other hand to be transmitted to institute again according to A module or the restoration algorithm module by the atmosphere light
Color correction module is stated to be handled;Video image data after the color correction resume module is the central processing list
The video image data of member output.
2. a kind of multi-mode fast video image demister according to claim 1, which is characterized in that the video is adopted
Truck uses OV5640 camera.
3. a kind of multi-mode fast video image demister according to claim 1, which is characterized in that deposit the outside
The MT41J256M16HA-125 composition that reservoir uses two panels model Micron, the width of bus is 4GB for 32bit, size,
Its highest data rate 1066Mbps.
4. a kind of multi-mode fast video image demister according to claim 1, which is characterized in that the video is aobvious
Show that device is connect using HDMI interface with the output end of the central processing unit.
5. a kind of multi-mode fast video image demister according to claim 1, which is characterized in that the dark
The work step of algoritic module are as follows: after updating dark channel diagram order and arriving, using next frame as key frame, first solution RGB
Then minimum value figure is input in matrix generator by the minimum value of each pixel in three channels as minimum value figure, generate
The matrix of 5*5 obtains the minimum value of entire 5*5 matrix, generates dark channel diagram, model is by solving minimum value twice in succession
6. a kind of multi-mode fast video image demister according to claim 5, which is characterized in that the guiding filter
The work step of wave module are as follows: 4*4 matrix is generated using pipeline system and 4*4 matrix generator, seeks every a line first
With, then asking it is continuous four column sums, solve to obtain mean value finally by the mode of displacement;It is being Q=a*I+b using model,
Algorithm realize dark refinement, a and b are parameter, and I is guide image, wherein The regular parameter ε introduced during solving a, ε solve model be ε=| var (Ik)-var(Gk)|。
7. a kind of multi-mode fast video image demister according to claim 6, which is characterized in that the atmosphere light
According to the work step of A module are as follows: during solving atmosphere light according to A, when atmosphere light arrives according to more newer command, in key
Maximum pixel point of the maximum value as every a line that every a line is solved in the dark channel diagram of frame, when the maximum value of last line solves
After out, the atmosphere light being maximized from every a line maximum pixel point value as a whole is shone.
8. a kind of multi-mode fast video image demister according to claim 7, which is characterized in that the recovery is calculated
The work step of method module are as follows: recovery will be input to by the foggy image after the dark channel diagram and delay after Steerable filter module
In module, is realized using parallel and pipelining and recovery operation is carried out to each pixel in three channels, wherein restoration model
For
9. a kind of multi-mode fast video image demister according to claim 8, which is characterized in that the color is repaired
The work step of positive module are as follows: image carries out micronization processes to treated, carries out first to the pixel point value in each channel
Pretreatment takes 255 when pixel value is greater than 255, take zero when pixel value is negative;Then to three channels of restored image
The pixel of RGB is done to be enhanced in proportion, obtains final restored image.
10. a kind of multi-mode fast video image defogging method, which comprises the following steps:
S1: acquisition video image data, and carry out the first order processing that video image data is realized in coding fusion;
S2: dark channel diagram processing will be updated by the video image data of first order processing;
S3: Steerable filter processing will be carried out by the video image data for updating dark channel diagram processing processing;
S4: atmosphere lighting process will be updated by the video image data of Steerable filter processing;
S5: recovery operation will be carried out by the video image data for updating atmosphere lighting process;
S6: color correction processing will be carried out by the video image data of recovery operation;
S7: the video image data output by color correction processing is displayed and store.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910508122.4A CN110223258A (en) | 2019-06-12 | 2019-06-12 | A kind of multi-mode fast video image defogging method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910508122.4A CN110223258A (en) | 2019-06-12 | 2019-06-12 | A kind of multi-mode fast video image defogging method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110223258A true CN110223258A (en) | 2019-09-10 |
Family
ID=67816783
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910508122.4A Pending CN110223258A (en) | 2019-06-12 | 2019-06-12 | A kind of multi-mode fast video image defogging method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110223258A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112184597A (en) * | 2020-11-05 | 2021-01-05 | 温州大学大数据与信息技术研究院 | Image restoration device and method |
CN112561776A (en) * | 2019-09-25 | 2021-03-26 | 杭州海康威视数字技术股份有限公司 | Image processing method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104252698A (en) * | 2014-06-25 | 2014-12-31 | 西南科技大学 | Semi-inverse method-based rapid single image dehazing algorithm |
CN104253930A (en) * | 2014-04-10 | 2014-12-31 | 西南科技大学 | Real-time video defogging method |
US20170084042A1 (en) * | 2014-06-12 | 2017-03-23 | Eizo Corporation | Fog removing device and image generating method |
CN107644431A (en) * | 2017-08-16 | 2018-01-30 | 孙战里 | Effective method for concentration of video monitoring in the case of a kind of low-light (level) |
CN108022225A (en) * | 2017-12-26 | 2018-05-11 | 哈尔滨理工大学 | Based on the improved dark channel prior image defogging algorithm of quick Steerable filter |
CN109255759A (en) * | 2018-08-02 | 2019-01-22 | 辽宁师范大学 | Image defogging method based on sky segmentation and transmissivity adaptive correction |
-
2019
- 2019-06-12 CN CN201910508122.4A patent/CN110223258A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104253930A (en) * | 2014-04-10 | 2014-12-31 | 西南科技大学 | Real-time video defogging method |
US20170084042A1 (en) * | 2014-06-12 | 2017-03-23 | Eizo Corporation | Fog removing device and image generating method |
CN104252698A (en) * | 2014-06-25 | 2014-12-31 | 西南科技大学 | Semi-inverse method-based rapid single image dehazing algorithm |
CN107644431A (en) * | 2017-08-16 | 2018-01-30 | 孙战里 | Effective method for concentration of video monitoring in the case of a kind of low-light (level) |
CN108022225A (en) * | 2017-12-26 | 2018-05-11 | 哈尔滨理工大学 | Based on the improved dark channel prior image defogging algorithm of quick Steerable filter |
CN109255759A (en) * | 2018-08-02 | 2019-01-22 | 辽宁师范大学 | Image defogging method based on sky segmentation and transmissivity adaptive correction |
Non-Patent Citations (2)
Title |
---|
曹慧: "实时视频去雾算法的研究及其FPGA实现", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
葛馨阳: "基于暗通道的快速图像去雾方法的研究与实现", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112561776A (en) * | 2019-09-25 | 2021-03-26 | 杭州海康威视数字技术股份有限公司 | Image processing method |
CN112561776B (en) * | 2019-09-25 | 2023-08-22 | 杭州海康威视数字技术股份有限公司 | image processing method |
CN112184597A (en) * | 2020-11-05 | 2021-01-05 | 温州大学大数据与信息技术研究院 | Image restoration device and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111460967B (en) | Illegal building identification method, device, equipment and storage medium | |
CN103310443B (en) | High-resolution remote sensing image immediate processing method and system | |
CN102393906B (en) | Computer vision inspection technology-based method for rapid detection and extraction of two-dimensional code | |
CN110852190B (en) | Driving behavior recognition method and system integrating target detection and gesture recognition | |
CN110472529A (en) | Target identification navigation methods and systems | |
CN110223258A (en) | A kind of multi-mode fast video image defogging method and device | |
CN111028267B (en) | Monocular vision following system and method for mobile robot | |
CN105913435A (en) | Multidimensional remote sensing image matching method and multidirectional remote sensing image matching system suitable for large area | |
CN109272110A (en) | Photoelectricity based on photon neural network chip merges intelligent signal processing system | |
CN104660951A (en) | Super-resolution amplification method of ultra-high definition video image converted from high definition video image | |
CN114972312A (en) | Improved insulator defect detection method based on YOLOv4-Tiny | |
CN105260995A (en) | Image repairing and denoising method and system | |
CN111882620A (en) | Road drivable area segmentation method based on multi-scale information | |
CN102905136A (en) | Video coding and decoding method and system | |
CN110889854B (en) | Sketch part segmentation method, system, device and storage medium based on multi-scale deep learning | |
CN116342953A (en) | Dual-mode target detection model and method based on residual shrinkage attention network | |
CN106469440B (en) | Dark defogging parallel optimization method based on OpenCL | |
CN111523439A (en) | Method, system, device and medium for target detection based on deep learning | |
CN117037198A (en) | Bank statement identification method | |
CN110738624B (en) | Area-adaptive image defogging system and method | |
CN110889858A (en) | Automobile part segmentation method and device based on point regression | |
Zhang et al. | Spatial-information Guided Adaptive Context-aware Network for Efficient RGB-D Semantic Segmentation | |
Wang et al. | Improving facade parsing with vision transformers and line integration | |
CN114937153B (en) | Visual characteristic processing system and method based on neural network in weak texture environment | |
CN116258854A (en) | Pavement crack segmentation method based on dense connection strategy and image gradient |
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: 20190910 |