CN109903235A - A Method for Eliminating Stripe Noise in Infrared Image - Google Patents
A Method for Eliminating Stripe Noise in Infrared Image Download PDFInfo
- Publication number
- CN109903235A CN109903235A CN201910051808.5A CN201910051808A CN109903235A CN 109903235 A CN109903235 A CN 109903235A CN 201910051808 A CN201910051808 A CN 201910051808A CN 109903235 A CN109903235 A CN 109903235A
- Authority
- CN
- China
- Prior art keywords
- column
- noise
- frame image
- infrared single
- infrared
- 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 29
- 238000012937 correction Methods 0.000 claims abstract description 12
- 238000012804 iterative process Methods 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims 4
- 238000003379 elimination reaction Methods 0.000 claims 4
- 238000003384 imaging method Methods 0.000 description 2
- 238000003331 infrared imaging Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Transforming Light Signals Into Electric Signals (AREA)
Abstract
本发明公开了一种红外图像条纹噪声的消除方法，包括步骤：第一步、获取包含列条纹噪声的原始红外单帧图像；第二步、根据原始红外单帧图像，建立原始红外单帧图像中相邻列的列条纹噪声之间偏差函数；第三步、对偏差函数进行最小化，然后利用递归方式，估计出每一列的列条纹噪声的噪声值；第四步、将包含列条纹噪声的原始红外单帧图像，减去所述估计出的每一列的列条纹噪声的噪声值，从而实现对原始红外单帧图像的校正，获得已消除条纹噪声的红外单帧图像。本发明公开的一种红外图像条纹噪声的消除方法，其能够及时有效地对红外单帧图像中存在的条纹噪声进行校正，提升最终获得的红外单帧图像的图像质量，具有重大的实践意义。
The invention discloses a method for eliminating stripe noise of an infrared image, comprising the steps of: first, acquiring an original infrared single-frame image containing column stripe noise; and secondly, establishing an original infrared single-frame image according to the original infrared single-frame image The deviation function between the column fringe noises of adjacent columns in ; the third step is to minimize the deviation function, and then use the recursive method to estimate the noise value of the column fringe noise of each column; the fourth step, will include the column fringe noise The original infrared single-frame image is subtracted from the estimated noise value of the column stripe noise of each column, so as to realize the correction of the original infrared single-frame image, and obtain an infrared single-frame image whose stripe noise has been eliminated. The invention discloses a method for eliminating stripe noise of an infrared image, which can timely and effectively correct the stripe noise existing in an infrared single-frame image, and improve the image quality of the finally obtained infrared single-frame image, which has great practical significance.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of removing method of infrared image fringes noise.
Background technique
Currently, infrared imaging is extremely important in industry, military and medical applications, because its available people's is infrared
Information.
Infrared focal plane array (IRFPA) is mainly used for infrared imaging system, and infrared focal plane array IRPFA is by small-scale transfer
Sensor and reading circuit composition, belong on infrared optical system focal plane, can make in entire visual field each pixel of scenery with
One sensitive first corresponding polynary planar array infrared detector.
For infrared focal plane array IRPFA, due to the variation of parameter in the mismatch and manufacturing process of element, pixel
The bias voltage of roomage response and reading circuit is different.Inhomogeneities leads to heterogeneity noise, and referred to as fixed mode is made an uproar
Sound (FPN), its presence seriously reduce the quality of infrared image collected.Wherein, in view of in infrared focal plane array IRPFA
In, reading circuit is usually same row pixel or shares the same output circuit with a line pixel.Due to column output circuit
Bias voltage is not quite identical, causes comprising the fringes noise with longitudinal stripe (i.e. column striped) for main feature in image, i.e.,
For non-uniform noise.
In order to promote the quality of infrared image, carry out Nonuniformity Correction (non-uniformity correction,
It NUC), is the steps necessary for improving infrared image.
Detector a row or column, which exports, shares an amplifier, therefore can generate heterogeneity item laterally or longitudinally
Line, this fringes noise heterogeneity noise special as one kind is especially prominent in infrared focal plane imaging system, commonly
Some Nonuniformity Corrections or two o'clock asymmetric correction method can not all filter out such noise.
In early stage research, by using statistics, it is assumed that by the mean value of each sensor and standard deviation and gain and deviation
Parameter is associated, it is assumed that input irradiation level is equally distributed stochastic variable, and all pixels sensor is having the same
Value and standard deviation, in this way can be by using linear model, to estimate the biasing and gain of each sensor.But this vacation
If it is only very long in image sequence and move it is sufficiently large so that when many different pieces of the inswept scene of each sensor
It is reasonable.
Non-uniform correction method of the another kind based on algebra scene, this method is independent of any system about scene temperature
Meter or scene diversity are it is assumed that it is used to extract about deviation using the estimation of the interframe sub-pixel shift in image sequence with linear
The interpolation model of the movement of heteropical information.The gain of each sensor and bias modeling are Gauss-by scheme also
Markov stochastic variable, and update the gain of these sensors using Kalman filter and bias estimated value.It is basic herein
On, by disabling Kalman filter parallel, according to the dynamic model parameters of itself, to estimate the gain and partially of each sensor
Difference.Realize that gain and the final of deviation are estimated by forming the weighted superposition of all estimations presented by each Kalman filter
Meter.
But the existing treatment scheme to infrared single-frame images, complex, real-time is poor, has one
Fixed limitation can not timely and effectively be corrected fringes noise present in infrared single-frame images, promote final obtain
Infrared single-frame images picture quality.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of removing method of infrared image fringes noise, it can be timely
Effectively fringes noise present in infrared single-frame images is corrected, promotes the image of the infrared single-frame images finally obtained
Quality has great practice significance.
For this purpose, the present invention provides a kind of removing methods of infrared image fringes noise, comprising the following steps:
The first step obtains the original infrared single-frame images comprising column fringes noise；
Second step, according to original infrared single-frame images, establish the column fringes noise of adjacent column in original infrared single-frame images
Between departure function；
Third step minimizes departure function, then utilizes recursive fashion, estimates the column fringes noise of each column
Noise figure；
4th step, by the original infrared single-frame images comprising column fringes noise, subtract the column of each column estimated
The noise figure of fringes noise obtains the infrared list for having eliminated fringes noise to realize the correction to original infrared single-frame images
Frame image.
Wherein, in the first step, it from the output signal of infrared sensor, obtains original infrared comprising column fringes noise
Single-frame images.
Wherein, in the first step, the formula of original infrared single-frame images z (i, j) indicates are as follows:
Z (i, j)=u (i, j)+b (j), formula (1)；
In the equation above, i, j are the line number and columns of pixel respectively, and b (j) is in infrared focal plane array IRFPA
Jth row reading circuit biasing, z is original infrared single-frame images, and u is ideal true infrared image.
Wherein, in second step, in original infrared single-frame images between the column fringes noise of adjacent column departure function E public affairs
Formula, specific as follows:
In the equation above, M is the line number of image, and u (j) is the average value of jth column pixel, ux(j) be u (j) ladder
Degree.
Wherein, by solution Euler-Lagrange equation, carry out the minimum of function to achieve the objective E, specific formula is as follows:
zxx(j)-bxx(j)-λ [b (j)]=0；
In the equation above, zxx(j)、bxxIt (j) is z (j) respectively, the second dervative of deviation b (j) is iterative process
Intermediate quantity.
By the above technical solution provided by the invention as it can be seen that compared with prior art, the present invention provides a kind of infrared
The removing method of image stripe noise timely and effectively can carry out school to fringes noise present in infrared single-frame images
Just, the picture quality of the infrared single-frame images finally obtained is promoted, there is great practice significance.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the removing method of infrared image fringes noise provided by the invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, with reference to the accompanying drawing with embodiment to this
Invention is described in further detail.
Referring to Fig. 1, it is based on the red of recursive fashion that the present invention, which provides a kind of removing method of infrared image fringes noise,
The removing method of outer image stripe noise, specifically includes the following steps:
The first step obtains the original infrared single-frame images comprising column fringes noise；
In the first step, it in specific implementation, can directly obtain from the output signal of infrared sensor comprising column striped
The original infrared single-frame images of noise.
Second step, according to original infrared single-frame images, establish the column fringes noise of adjacent column in original infrared single-frame images
Between departure function；
Third step minimizes departure function, then utilizes recursive fashion, estimates the column fringes noise of each column
Noise figure；
4th step, by the original infrared single-frame images comprising column fringes noise, subtract the column of each column estimated
The noise figure (i.e. noise vector value) of fringes noise, to realize the correction to original infrared single-frame images, item has been eliminated in acquisition
The infrared single-frame images of line noise.
In the present invention, about the linear imaging model of infrared sensor.The output signal of infrared sensor is (i.e. practical defeated
Infrared image and original infrared single-frame images out) formula of z (i, j) can indicate are as follows:
Z (i, j)=u (i, j)+b (j), formula (1)；
In the equation above, i, j are the line number and columns of pixel respectively, and b (j) is infrared focal plane array (IRFPA)
In jth row reading circuit biasing, z is noisy reality output infrared image (i.e. original infrared single-frame images), and u is reason
The true infrared image (not having noisy infrared single-frame images, also the infrared single-frame images to have eliminated fringes noise) thought.
It should be noted that for the present invention, according to formula (1), in order to restore incident in image z from being destroyed
Infra-red radiation, should estimate first bias b (j).In infrared image, the infra-red radiation of adjacent pixel is typically considered in space
It is highly relevant, therefore the average value of the column pixel of true picture u should be slowly varying signal.And deviation b (j) is only
Vertical random noise, it will lead to quick localized variation.Therefore, the notable difference between the column of reality output infrared image z
Mainly generated by deviation b (j), and the difference between the column of true infrared image u should be smaller, therefore, original infrared single frames
In image between the column fringes noise of adjacent column departure function E (also referred to as objective function) formula, it is specific as follows:
In the equation above, M is the line number of image, and u (j) is the average value of jth column pixel.ux(j) be u (j) ladder
Degree.
On the other hand, it is not enough to change the dynamic range of true infrared image u, therefore, adjacent column in view of deviation b (j)
Departure function (also referred to as objective function) E can be between column fringes noise is defined as:
In the equation above, z (j) is the average value that jth arranges in reality output infrared image, zx(j) be z (j) ladder
Degree, bxIt (j) is to bias the gradient of b (j), and know based on formula (1): u (j)=z (j) b (j).By minimizing target letter
Number (i.e. minimum departure function) carrys out estimated bias b (j).Wherein, the first item of objective function is known as smooth item, it is intended to minimize
Correct the difference between the column of image.Section 2 is known as bound term, and for guarding against deviations, b (j) excessively significantly changes image.Ginseng
Number λ controls the balance between two.
In the present invention, the minimum of function to achieve the objective E can be carried out, specifically by solving Euler-Lagrange equation
Formula is as follows:
zxx(j)-bxx(j)-λ [b (j)]=0, formula (4)；
In the equation above, zxx(j)、bxxIt (j) is z (j) respectively, the second dervative of deviation b (j) is iterative process
Intermediate quantity.The numerical value solution of formula (4) is as follows:
In the equation above, bo(i, j) is primary data, and n is the number of iterations.Δ t is the iteration of control convergence speed
?.Iteration may cannot be restrained since step-length is excessive.Show the step Δ t in [0.1,0.01] range most by experiment
It is suitable for this method.Using second-order central finite difference calculus, the second dervative for obtaining z (i, j) and b (i, j) can be calculated.
In the present invention, the deviation of estimation can be used, the reality output infrared image Z (i, j) that Lai Jiaozheng is observed is
Meet following formula:
WhereinInfrared image to use the above method to estimate obtains each column deviation,It indicates using estimation
Deviation correction after close to ideal image.Since deviation b (j) is changed over time slowly, it can be considered that adjacent frame
Between deviation be constant.The estimation of biasing b (j) does not need to realize within the time of a frame.Bias the recursive estimation of b (j)
It is written as follows:
In the equation above, t indicates frame number.This recursive method can independently iterate to calculate the deviation of every frame,
It is significant in a frame time to reduce calculating, it is very suitable to the application of real-time display.
It was proved that parameter lambda is appropriate value in [0.01,0.1] range.It, can be by λ in specific implementation
It is respectively set to 0.1 with Δ t, the number of iterations is fixed on 200 times.
It should be noted that being based on above technical scheme it is found that considering the single-frame images based on scene for the present invention
Processing, and algorithm complexity is reduced, there is positive help to the promotion of Nonuniformity Correction.It is provided by the invention to be based on
The removing method of the infrared image fringes noise of recursive fashion.This method can estimate striped deviation based on scene image, thus
Fringes noise is corrected, and is suitable for the non-uniform situation of striped.
Compared with prior art, for the removing method of infrared image fringes noise provided by the invention, have with
Under advantageous effects:
1, output valve is considered as linear model with the variation of row reading circuit deviation by the present invention.It, will be infrared according to the model
Difference in image between adjacent column is described as departure function, by finding deviation, keeps difference as small as possible.To infrared image into
Row correction eliminates fringes noise using the deviation of estimation.
2, the present invention can substantially reduce calculation amount in a frame time, improve algorithm speed by recursive mode.
In conclusion compared with prior art, a kind of removing method of infrared image fringes noise provided by the invention,
It can timely and effectively be corrected fringes noise present in infrared single-frame images, promote the infrared single frames finally obtained
The picture quality of image has great practice significance.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (5)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910051808.5A CN109903235A (en) | 2019-01-21 | 2019-01-21 | A Method for Eliminating Stripe Noise in Infrared Image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910051808.5A CN109903235A (en) | 2019-01-21 | 2019-01-21 | A Method for Eliminating Stripe Noise in Infrared Image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109903235A true CN109903235A (en) | 2019-06-18 |
Family
ID=66943934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910051808.5A Pending CN109903235A (en) | 2019-01-21 | 2019-01-21 | A Method for Eliminating Stripe Noise in Infrared Image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109903235A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110400271A (en) * | 2019-07-09 | 2019-11-01 | 浙江大华技术股份有限公司 | A kind of striped asymmetric correction method, device, electronic equipment and storage medium |
| CN110910324A (en) * | 2019-11-19 | 2020-03-24 | 山东神戎电子股份有限公司 | How to remove vertical stripes from infrared video |
| CN111161172A (en) * | 2019-12-18 | 2020-05-15 | 北京波谱华光科技有限公司 | Infrared image column direction stripe eliminating method, system and computer storage medium |
| CN111524057A (en) * | 2020-04-14 | 2020-08-11 | 烟台艾睿光电科技有限公司 | Infrared image generation method, device and equipment and infrared thermal imaging system |
| CN111784599A (en) * | 2020-06-24 | 2020-10-16 | 西北工业大学 | A method for eliminating streak noise in infrared images |
| CN111986171A (en) * | 2020-08-14 | 2020-11-24 | 西安应用光学研究所 | Abnormal element detection method for infrared linear array detector |
| WO2021022779A1 (en) * | 2019-08-02 | 2021-02-11 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for noise reduction |
| US20220189043A1 (en) * | 2020-12-15 | 2022-06-16 | Microsoft Technology Licensing, Llc | Correcting line bias in an image |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104657958A (en) * | 2015-03-18 | 2015-05-27 | 西安科技大学 | Infrared image stripe noise elimination method |
| CN105931203A (en) * | 2016-04-26 | 2016-09-07 | 成都市晶林科技有限公司 | Infrared image stripe filtering method based on statistical relative stripe removal method |
| CN106846275A (en) * | 2017-01-24 | 2017-06-13 | 西安科技大学 | A kind of real-time removing method of Infrared video image strip noise |
| CN106934771A (en) * | 2017-02-16 | 2017-07-07 | 武汉镭英科技有限公司 | A kind of infrared image fringes noise minimizing technology based on local correlations |
-
2019
- 2019-01-21 CN CN201910051808.5A patent/CN109903235A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104657958A (en) * | 2015-03-18 | 2015-05-27 | 西安科技大学 | Infrared image stripe noise elimination method |
| CN105931203A (en) * | 2016-04-26 | 2016-09-07 | 成都市晶林科技有限公司 | Infrared image stripe filtering method based on statistical relative stripe removal method |
| CN106846275A (en) * | 2017-01-24 | 2017-06-13 | 西安科技大学 | A kind of real-time removing method of Infrared video image strip noise |
| CN106934771A (en) * | 2017-02-16 | 2017-07-07 | 武汉镭英科技有限公司 | A kind of infrared image fringes noise minimizing technology based on local correlations |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110400271B (en) * | 2019-07-09 | 2021-06-15 | 浙江大华技术股份有限公司 | Stripe non-uniformity correction method and device, electronic equipment and storage medium |
| CN110400271A (en) * | 2019-07-09 | 2019-11-01 | 浙江大华技术股份有限公司 | A kind of striped asymmetric correction method, device, electronic equipment and storage medium |
| US12205305B2 (en) | 2019-08-02 | 2025-01-21 | Zhejiang Pixfra Technology Co., Ltd. | Information processing method and system |
| WO2021022779A1 (en) * | 2019-08-02 | 2021-02-11 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for noise reduction |
| CN110910324A (en) * | 2019-11-19 | 2020-03-24 | 山东神戎电子股份有限公司 | How to remove vertical stripes from infrared video |
| CN110910324B (en) * | 2019-11-19 | 2023-04-14 | 山东神戎电子股份有限公司 | Method for removing vertical stripes in infrared video |
| CN111161172A (en) * | 2019-12-18 | 2020-05-15 | 北京波谱华光科技有限公司 | Infrared image column direction stripe eliminating method, system and computer storage medium |
| CN111161172B (en) * | 2019-12-18 | 2020-11-06 | 北京波谱华光科技有限公司 | Infrared image column direction stripe eliminating method, system and computer storage medium |
| CN111524057B (en) * | 2020-04-14 | 2023-06-02 | 烟台艾睿光电科技有限公司 | Infrared image generation method, device, equipment and infrared thermal imaging system |
| CN111524057A (en) * | 2020-04-14 | 2020-08-11 | 烟台艾睿光电科技有限公司 | Infrared image generation method, device and equipment and infrared thermal imaging system |
| CN111784599B (en) * | 2020-06-24 | 2022-04-29 | 西北工业大学 | A method for eliminating streak noise in infrared images |
| CN111784599A (en) * | 2020-06-24 | 2020-10-16 | 西北工业大学 | A method for eliminating streak noise in infrared images |
| CN111986171A (en) * | 2020-08-14 | 2020-11-24 | 西安应用光学研究所 | Abnormal element detection method for infrared linear array detector |
| CN111986171B (en) * | 2020-08-14 | 2024-02-27 | 西安应用光学研究所 | Abnormal element detection method for infrared array detector |
| US20220189043A1 (en) * | 2020-12-15 | 2022-06-16 | Microsoft Technology Licensing, Llc | Correcting line bias in an image |
| US11776138B2 (en) * | 2020-12-15 | 2023-10-03 | Microsoft Technology Licensing, Llc | Correcting line bias in an image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109903235A (en) | A Method for Eliminating Stripe Noise in Infrared Image | |
| JP6887597B2 (en) | Infrared image processing methods and image processing devices for non-uniform correction | |
| Scribner et al. | Adaptive nonuniformity correction for IR focal-plane arrays using neural networks | |
| Wang | Stripe noise removal for infrared image by minimizing difference between columns | |
| CN111539879A (en) | Video blind denoising method and device based on deep learning | |
| US20090257679A1 (en) | Scene based non-uniformity correction systems and methods | |
| Sheng-Hui et al. | Guided filter and adaptive learning rate based non-uniformity correction algorithm for infrared focal plane array | |
| CN104917936B (en) | One kind is based on the relevant time domain high pass asymmetric correction method of gray scale | |
| CN110211056A (en) | Adaptive infrared image based on value histogram in part removes striped algorithm | |
| WO2017185441A1 (en) | Infrared image stripe filtering method based on statistical relative stripe removal method | |
| CN110501025B (en) | On-orbit calibration method of rolling shutter area scan CMOS camera | |
| CN109791699B (en) | Radiation Imaging | |
| Lee et al. | Dual-branch structured de-striping convolution network using parametric noise model | |
| Ratliff et al. | Algebraic scene-based nonuniformity correction in focal plane arrays | |
| CN106846275B (en) | A Real-time Elimination Method of Strip Noise in Infrared Video Image | |
| CN109767403A (en) | A scene-based infrared focal plane array strip noise removal method | |
| CN110553739A (en) | Non-barrier-piece non-uniformity correction method for infrared thermal imaging | |
| CN109712088A (en) | A remote sensing video satellite image processing method and system based on image stabilization | |
| CN109636740B (en) | A multi-scale intelligent non-uniformity correction method for infrared images | |
| CN103997611A (en) | Method for suppressing image fixed-pattern noise based on noise template | |
| CN115082335B (en) | A method for image denoising based on noise modeling | |
| CN103985089B (en) | With reference to weight edge analysis and the image streak correction method of frame inner iteration | |
| Liu et al. | Efficient polarization demosaicking via low-cost edge-aware and inter-channel correlation | |
| CN106815820B (en) | A kind of infrared image strip noise cancellation method | |
| CN114240801A (en) | A method for non-uniform correction of remote sensing images |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190618 |