CN109949252B - Infrared image light spot removing method based on compensation coefficient fitting - Google Patents
Infrared image light spot removing method based on compensation coefficient fitting Download PDFInfo
- Publication number
- CN109949252B CN109949252B CN201910298232.2A CN201910298232A CN109949252B CN 109949252 B CN109949252 B CN 109949252B CN 201910298232 A CN201910298232 A CN 201910298232A CN 109949252 B CN109949252 B CN 109949252B
- Authority
- CN
- China
- Prior art keywords
- image
- light spot
- spot
- light
- point set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an infrared image light spot removing method based on compensation coefficient fitting, and belongs to the field of image processing. The implementation method of the invention comprises the following steps: calculating according to the matrix low-rank decomposition to obtain a light spot template image; detecting to obtain a sampling point set of a light spot position by using a threshold segmentation and connected block search technology, wherein the sampling point set of the light spot position comprises a light spot position point set and a point set adjacent to the light spot position point set; calculating and fitting to obtain a compensation coefficient of the image by using a coefficient fitting method in combination with the light spot template and the sampling point set; and performing linear compensation on the current infrared image by using the compensation coefficient and the light spot template to obtain the infrared image with the light spots removed. The invention can remove light spots in the infrared image and improve the imaging quality of the infrared camera, and can remove light spots with small area and light spots with larger area.
Description
Technical Field
The invention relates to a light spot removing method, in particular to an infrared image light spot removing method based on compensation coefficient fitting, and belongs to the field of image processing.
Background
With the development of infrared imaging technology, more and more infrared imaging technologies are applied to various industries, and relate to various aspects such as monitoring, fire fighting, military and the like. For uncooled infrared imaging devices, germanium windows are typically provided for filtering infrared light. However, due to the limitations of the manufacturing process, dust of varying sizes is often attached to germanium windows. The dust causes uneven light transmission degree at the corresponding position of the germanium window, and finally causes light spots with different intensities in the infrared image. These spots seriously affect the imaging quality and the viewing experience, and therefore, the spot removal is one of the important methods for improving the infrared imaging quality.
For infrared image spot removal, the problem is rarely researched at home and abroad. The currently available light spot removing method mainly aims at the denoising method of the infrared image. The denoising method for the infrared image is represented by a denoising method (Ni C, Li Q, Xia L Z.A novel method and edge enhancement. Signal Processing,2008,88(6):1606-1614.) based on wavelet transformation, the method utilizes the wavelet transformation and the local singularity of the image to establish a combined model, updates the edge of the image through the geometric continuity of the edge, and simultaneously introduces a two-way diffusion technology in the wavelet diffusion to eliminate noise; the method may cause information loss of the image at the corresponding position during the process of removing the light spots. In addition, the method can remove the light spot with a small area, but has poor effect on the light spot with a large area.
Disclosure of Invention
The invention discloses an infrared image light spot removing method based on compensation coefficient fitting, which aims to solve the technical problems that: the method and the device can remove the light spots in the infrared image and improve the imaging quality of the infrared camera, and can remove the light spots with small area and the light spots with larger area.
The purpose of the invention is realized by the following technical scheme:
the invention discloses an infrared image spot removing method based on compensation coefficient fitting, which comprises the steps of calculating according to matrix low-rank decomposition to obtain a spot template image; detecting by using a threshold segmentation and connected block search technology to obtain a sampling point set of the light spot position; calculating and fitting to obtain a compensation coefficient of the image by using a coefficient fitting method in combination with the light spot template and the sampling point set; and performing linear compensation on the current infrared image by using the compensation coefficient and the light spot template to obtain the infrared image with the light spots removed.
The invention discloses an infrared image light spot removing method based on compensation coefficient fitting, which comprises the following steps of:
step one, calculating according to the matrix low rank to obtain a light spot template image.
Collecting a calibration image C of the infrared imaging equipment, wherein the calibration image C is the linear superposition of the background image B and the light spot template image E, namely
C＝B+E (1)
And solving the following energy function according to matrix low-rank decomposition, and performing iterative computation to obtain a light spot template image E.
Wherein | | | purple hairFRepresents the Frobenius norm, | | | | | | purple phosphor*Represents the nuclear norm, | | | | luminance1Representing a 1 norm. μ and λ are parameters that control the weight of the energy term.
And secondly, obtaining a sampling point set of the light spot position by using threshold segmentation and connected block detection, wherein the sampling point set of the light spot position comprises a light spot position point set and a point set adjacent to the light spot position point set.
Given spot intensity division threshold th1And a spot intensity difference threshold th2. Using a threshold th for splitting the intensity of the spot1Performing binarization processing on the spot template image E obtained in the step one to obtain a binarized image Eb. Then, for the binary image EbDetecting the connected blocks to obtain a binary image EbThe n connected blocks with the largest area are taken to form a connected Block set { Block }1,Block2,…,Blockn}. For the ith connected Block BlockiGet the packet containing BlockiThe smallest Circle is taken as the corresponding light spot Circlei. Respectively processing the n connected blocks to obtain a light spot set { Circle }1,Circle2,…,CirclenAnd the circle center and the radius of the light spot set are the position and the size of the light spot, namely the position and the size of the light spot in the image are detected by combining low-rank decomposition of the matrix with threshold segmentation and connected block detection, so that the calculation of the compensation coefficient in the step three is facilitated.
For ith light spot CircleiTaking a point x and an adjacent point y on the vertical diameter, and when the absolute value of the difference between the pixel intensities of the template image E at the corresponding positions of the two points is greater than the light spot intensity difference threshold th2While, a point x on the diameter and its neighboring point y are added to the light, respectivelySpot location point sets P and Q. The same operation is carried out on the points on the vertical diameter of all the light spots in the light spot set, and finally the light spot position point set P is obtained as { x ═ x1,x2,…,xmAnd a set of points Q adjacent to the set of spot location points P (y)1,y2,…,ym}. Where m represents the number of points.
Step three: and D, obtaining the pixel intensity of the corresponding position of the light spot template according to the sampling point set in the step two, fitting the pixel intensity by using a coefficient fitting method, and calculating to obtain a compensation coefficient of the image.
And the coefficient fitting method in the third step preferably adopts a least square method to fit the pixel intensity.
When the least square method is selected as the coefficient fitting method in the third step, the third step is realized as follows:
given an infrared image D with light spots, the relationship between the image I after the light spots are finally removed and a light spot template image E is modeled as follows:
D＝I-ηE (3)
wherein η represents a compensation coefficient for controlling the image weight of the spot template.
Since in a natural image the flat area occupies most of the area of the image, the same probability is much greater than the different probabilities for the intensities of any two adjacent pixels of the speckle-free image I. Using said characteristic, for the ith point x in the spot position point set PiAnd the ith point y in the point set Q adjacent to the spot position point set PiTaking the pixel intensity value corresponding to the corresponding position of I as I (x)i) And I (y)i) Is provided with
I(xi)＝I(yi) (4)
Combining the formulas (3) and (4), the following relationship is obtained
D(xi)-D(yi)＝η[E(xi)-E(yi)] (5)
And performing least square fitting by using all m points in the spot position point set P and all m points in the point set Q adjacent to the spot position point set P to obtain a relatively accurate compensation coefficient eta. The solution formula of the compensation coefficient eta is as follows:
wherein e isTRepresenting the transpose of the matrix e. The matrices e and d are of the form:
e＝[E(x1)-E(y1) E(x2)-E(y2) … E(xm)-E(ym)]
d＝[D(x1)-D(y1) D(x2)-D(y2) … D(xm)-D(ym)] (7)
and step four, performing linear compensation on the current infrared image by using the compensation coefficient obtained in the step three and the light spot template to obtain the infrared image with the light spots removed, so as to improve the imaging quality of the infrared camera.
According to the light spot template image E obtained in the first step, the compensation coefficient eta obtained in the third step and the given infrared image D with the light spots, performing linear compensation on the infrared image D with the light spots at present according to a formula (3) to obtain an image I after the light spots are removed:
I＝D+ηE (8)
the obtained image I is the infrared image with the light spots removed corresponding to the given image D with the light spots, and the imaging quality of the infrared camera is improved.
Has the advantages that:
1. according to the infrared image light spot removing method based on compensation coefficient fitting, disclosed by the invention, the position and the size of the light spot in the image can be accurately detected by using low-rank decomposition of a matrix in combination with threshold segmentation and connected block detection, and meanwhile, the light spots with different areas can be effectively eliminated by establishing a compensation model aiming at the characteristics of the light spot.
2. The invention discloses an infrared image light spot removing method based on compensation coefficient fitting, which is characterized in that a coefficient fitting method is used for combining a light spot template and a sampling point set to calculate and fit to obtain a compensation coefficient of an image; the compensation coefficient and the light spot template are used for carrying out linear compensation on the current infrared image to obtain the infrared image with the light spots removed, so that the effective information of the corresponding position can be enhanced while the light spots are eliminated, the effective information of the image can be reserved as much as possible, and the imaging quality of the infrared camera can be effectively improved.
Drawings
FIG. 1 is a flow chart of infrared image speckle removal based on compensation coefficient fitting according to the present invention.
Fig. 2 is a schematic diagram comparing an image with light spots with an image after removing the light spots, wherein fig. 2a is the image with the light spots, and fig. 2b is the image after removing the light spots.
Detailed Description
For a better understanding of the objects and advantages of the present invention, reference should be made to the following detailed description taken in conjunction with the accompanying drawings and examples.
As shown in fig. 1, the method for removing infrared image spots based on compensation coefficient fitting disclosed in this embodiment includes the following steps:
step one, calculating according to the matrix low rank to obtain a light spot template image.
Collecting a calibration image C of the infrared imaging equipment, wherein the calibration image C is the linear superposition of the background image B and the light spot template image E, namely
C＝B+E (1)
For a specific camera, because the germanium window is packaged, the template E does not change, and the template E can be used for multiple times only by calibration calculation once. For the calibration image C, the infrared imaging device may be respectively corrected at two temperatures with a large difference, and finally an image with an obvious light spot is obtained as the calibration image.
And solving the following energy function according to matrix low-rank decomposition, and performing iterative computation to obtain a light spot template image E.
Wherein | | | purple hairFRepresents the Frobenius norm, | | | | | | purple phosphor*Represents the nuclear norm, | | | | luminance1Representing a 1 norm. Mu andλ is a parameter that controls the weight of the energy term.
And secondly, obtaining a sampling point set of the light spot position by using threshold segmentation and connected block detection, wherein the sampling point set of the light spot position comprises a light spot position point set and a point set adjacent to the light spot position point set.
Given spot intensity division threshold th1And a spot intensity difference threshold th2. Using a threshold th for splitting the intensity of the spot1Performing binarization processing on the spot template image E obtained in the step one to obtain a binarized image Eb. Then, for the binary image EbDetecting the connected blocks to obtain a binary image EbThe n connected blocks with the largest area are taken to form a connected Block set { Block }1,Block2,…,Blockn}. For the ith connected Block BlockiGet the packet containing BlockiThe smallest Circle is taken as the corresponding light spot Circlei. Respectively processing the n connected blocks to obtain a light spot set { Circle }1,Circle2,…,CirclenAnd n is set to be between 10 and 20 in actual operation, the circle center and the radius of the light spot set are the position and the size of the light spot, namely the position and the size of the light spot in the image are detected by combining low rank decomposition of the matrix with threshold segmentation and connected block detection, and the calculation of the compensation coefficient in the step three is facilitated.
For ith light spot CircleiTaking a point x and an adjacent point y on the vertical diameter, and when the absolute value of the difference between the pixel intensities of the template image E at the corresponding positions of the two points is greater than the light spot intensity difference threshold th2At the same time, the point x on the diameter and the adjacent point y are added into the spot position point sets P and Q respectively. For the selection of the adjacent point y, the corresponding position of the pixel before x is generally selected in the scanning order of the image in practical operation. The same operation is carried out on the points on the vertical diameter of all the light spots in the light spot set, and finally the light spot position point set P is obtained as { x ═ x1,x2,…,xmAnd a set of points Q adjacent to the set of spot location points P (y)1,y2,…,ym}. Where m represents the number of points, set at around 100.
Step three: and D, obtaining the pixel intensity of the corresponding position of the light spot template according to the sampling point set in the step two, fitting the pixel intensity by using a coefficient fitting method, and calculating to obtain a compensation coefficient of the image.
When the least square method is selected as the coefficient fitting method in the third step, the third step is realized as follows:
given an infrared image D with light spots, the relationship between the image I after the light spots are finally removed and a light spot template image E is modeled as follows:
D＝I-ηE (3)
wherein η represents a compensation coefficient for controlling the image weight of the spot template. In practice, the width and height of the infrared image D with the light spots are the same as those of the light spot template image E.
Since in a natural image the flat area occupies most of the area of the image, the same probability is much greater than the different probabilities for the intensities of any two adjacent pixels of the speckle-free image I. Using said characteristic, for the ith point x in the spot position point set PiAnd the ith point y in the point set Q adjacent to the spot position point set PiTaking the pixel intensity value corresponding to the corresponding position of I as I (x)i) And I (y)i) Is provided with
I(xi)＝I(yi) (4)
Combining the formulas (3) and (4), the following relationship is obtained
D(xi)-D(yi)＝η[E(xi)-E(yi)] (5)
And performing least square fitting by using all m points in the spot position point set P and all m points in the point set Q adjacent to the spot position point set P to obtain a relatively accurate compensation coefficient eta. The solution formula of the compensation coefficient eta is as follows:
wherein e isTRepresenting the transpose of the matrix e. The matrices e and d are of the form:
e＝[E(x1)-E(y1) E(x2)-E(y2) … E(xm)-E(ym)]
d＝[D(x1)-D(y1) D(x2)-D(y2) … D(xm)-D(ym)] (7)
and step four, performing linear compensation on the current infrared image by using the compensation coefficient obtained in the step three and the light spot template to obtain the infrared image with the light spots removed, so as to improve the imaging quality of the infrared camera.
According to the light spot template image E obtained in the first step, the compensation coefficient eta obtained in the third step and the given infrared image D with the light spots, performing linear compensation on the infrared image D with the light spots at present according to a formula (3) to obtain an image I after the light spots are removed:
I＝D+ηE (8)
the calculated image I is the infrared image with the light spots removed corresponding to the given image D with the light spots, and the imaging quality of the infrared camera is improved. Fig. 2 shows the effect of the spot removal method. In fig. 2, fig. 2a is an original infrared image with light spots, which can be clearly seen as having light spots with different sizes at multiple positions of the image, and the light spots seriously reduce the quality of the image. Fig. 2b is the image after the light spots are removed, the light spots in the image are eliminated, and effective information (such as edge information) covered by the light spots can be successfully recovered, so that the imaging quality of the infrared camera is improved.
The above detailed description is intended to illustrate the objects, aspects and advantages of the present invention, and it should be understood that the above detailed description is only exemplary of the present invention and is not intended to limit the scope of the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (5)
1. An infrared image light spot removing method based on compensation coefficient fitting is characterized by comprising the following steps: comprises the following steps of (a) carrying out,
step one, calculating according to matrix low-rank decomposition to obtain a light spot template image;
step two, a sampling point set of the light spot position is obtained by using threshold segmentation and connected block detection, wherein the sampling point set of the light spot position comprises a light spot position point set and a point set adjacent to the light spot position point set;
step three: obtaining the pixel intensity of the corresponding position of the light spot template according to the sampling point set in the step two, fitting the pixel intensity by using a coefficient fitting method, and calculating to obtain a compensation coefficient of the image;
step four, performing linear compensation on the current infrared image by using the compensation coefficient obtained in the step three and the light spot template to obtain the infrared image with the light spots removed, and further improving the imaging quality of the infrared camera;
wherein, the first realization method of the step is,
collecting a calibration image C of the infrared imaging equipment, wherein the calibration image C is the linear superposition of the background image B and the light spot template image E, namely
C＝B+E (1)
Solving the following energy function according to matrix low-rank decomposition, and performing iterative computation to obtain a light spot template image E;
wherein | | | purple hairFRepresents the Frobenius norm, | | | | | | purple phosphor*Represents the nuclear norm, | | | | luminance1Represents a norm of 1; μ and λ are parameters that control the weight of the energy term.
2. The infrared image spot removal method based on the fitting of the compensation coefficient as claimed in claim 1, wherein: the second step is realized by the method that,
given spot intensity division threshold th1And a spot intensity difference threshold th2(ii) a Using a threshold th for splitting the intensity of the spot1Performing binarization processing on the spot template image E obtained in the step one to obtain a binarized image Eb(ii) a Then, the binary image is processedEbDetecting the connected blocks to obtain a binary image EbThe n connected blocks with the largest area are taken to form a connected Block set { Block }1,Block2,…,Blockn}; for the ith connected Block BlockiGet the packet containing BlockiThe smallest Circle is taken as the corresponding light spot Circlei(ii) a Respectively processing the n connected blocks to obtain a light spot set { Circle }1,Circle2,…,CirclenThe circle center and the radius of the light spot set are the position and the size of the light spot, namely the position and the size of the light spot in the image are detected by combining low-rank decomposition of the matrix with threshold segmentation and connected block detection, so that the calculation of the compensation coefficient in the step three is facilitated;
for ith light spot CircleiTaking a point x and an adjacent point y on the vertical diameter, and when the absolute value of the difference between the pixel intensities of the template image E at the corresponding positions of the point x and the point y is greater than the spot intensity difference threshold th2Adding the point x on the diameter and the adjacent point y thereof into the spot position point sets P and Q respectively; the same operation is carried out on the points on the vertical diameter of all the light spots in the light spot set, and finally the light spot position point set P is obtained as { x ═ x1,x2,…,xmAnd a set of points Q adjacent to the set of spot location points P (y)1,y2,…,ym}; where m represents the number of points.
3. The infrared image spot removal method based on the fitting of the compensation coefficient as claimed in claim 2, wherein: and thirdly, fitting the pixel intensity by using a least square method according to the coefficient fitting method.
4. The infrared image spot removal method based on the fitting of the compensation coefficient as claimed in claim 3, wherein: when the least square method is selected as the coefficient fitting method in the third step, the third step is realized as follows,
given an infrared image D with light spots, the relationship between the image I after the light spots are finally removed and a light spot template image E is modeled as follows:
D＝I+ηE (3)
wherein eta represents a compensation coefficient for controlling the image weight of the light spot template;
because the flat area occupies most of the area of the image in the natural image, the same probability of the intensities of any two adjacent pixels of the facula-free image I is far greater than different probabilities; for the ith point x in the spot position point set PiAnd the ith point y in the point set Q adjacent to the spot position point set PiTaking the pixel intensity value corresponding to the corresponding position of I as I (x)i) And I (y)i) Is provided with
I(xi)＝I(yi) (4)
Combining the formulas (3) and (4), the following relationship is obtained
D(xi)-D(yi)＝η[E(xi)-E(yi)] (5)
Performing least square fitting by using all m points in the light spot position point set P and all m points in the point set Q adjacent to the light spot position point set P to obtain a relatively accurate compensation coefficient eta; the solution formula of the compensation coefficient eta is as follows:
wherein e isTRepresents the transpose of matrix e; the matrices e and d are of the form:
e＝[E(x1)-E(y1) E(x2)-E(y2) … E(xm)-E(ym)]
d＝[D(x1)-D(y1) D(x2)-D(y2) … D(xm)-D(ym)] (7)。
5. the infrared image spot removal method based on the fitting of the compensation coefficient as claimed in claim 4, wherein: according to the light spot template image E obtained in the first step, the compensation coefficient eta obtained in the third step and the given infrared image D with the light spots, performing linear compensation on the infrared image D with the light spots at present according to a formula (3) to obtain an image I after the light spots are removed:
I＝D-ηE (8)
the obtained image I is the infrared image with the light spots removed corresponding to the given image D with the light spots, and the imaging quality of the infrared camera is improved.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910298232.2A CN109949252B (en) | 2019-04-15 | 2019-04-15 | Infrared image light spot removing method based on compensation coefficient fitting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910298232.2A CN109949252B (en) | 2019-04-15 | 2019-04-15 | Infrared image light spot removing method based on compensation coefficient fitting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109949252A CN109949252A (en) | 2019-06-28 |
| CN109949252B true CN109949252B (en) | 2020-12-25 |
Family
ID=67015140
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910298232.2A Active CN109949252B (en) | 2019-04-15 | 2019-04-15 | Infrared image light spot removing method based on compensation coefficient fitting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949252B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111614903B (en) * | 2020-05-28 | 2021-08-10 | 西安航空学院 | Method for removing faculae in image shooting |
| CN114565517B (en) * | 2021-12-29 | 2023-09-29 | 骨圣元化机器人(深圳)有限公司 | Image denoising method and device of infrared camera and computer equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103049892A (en) * | 2013-01-27 | 2013-04-17 | 西安电子科技大学 | Non-local image denoising method based on similar block matrix rank minimization |
| CN103679660A (en) * | 2013-12-16 | 2014-03-26 | 清华大学 | Method and system for restoring image |
| CN103745445A (en) * | 2014-01-21 | 2014-04-23 | 中国科学院地理科学与资源研究所 | Gaussian and pulse mixed noise removing method and device |
| CN106204477A (en) * | 2016-07-06 | 2016-12-07 | 天津大学 | Video frequency sequence background restoration methods based on online low-rank background modeling |
| CN106778814A (en) * | 2016-11-24 | 2017-05-31 | 郑州航空工业管理学院 | A kind of method of the removal SAR image spot based on projection spectral clustering |
| US9858497B2 (en) * | 2014-05-06 | 2018-01-02 | Nant Holdings Ip, Llc | Image-based feature detection using edge vectors |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927730B (en) * | 2014-01-22 | 2017-05-24 | 西安电子科技大学 | Image noise reduction method based on Primal Sketch correction and matrix filling |
| CN104915940A (en) * | 2015-06-03 | 2015-09-16 | 厦门美图之家科技有限公司 | Alignment-based image denoising method and system |
| CN106504214B (en) * | 2016-10-31 | 2019-03-05 | 西京学院 | The high spectrum image Banded improvement removing method of wavelet transformation and local interpolation fusion |
-
2019
- 2019-04-15 CN CN201910298232.2A patent/CN109949252B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103049892A (en) * | 2013-01-27 | 2013-04-17 | 西安电子科技大学 | Non-local image denoising method based on similar block matrix rank minimization |
| CN103679660A (en) * | 2013-12-16 | 2014-03-26 | 清华大学 | Method and system for restoring image |
| CN103745445A (en) * | 2014-01-21 | 2014-04-23 | 中国科学院地理科学与资源研究所 | Gaussian and pulse mixed noise removing method and device |
| US9858497B2 (en) * | 2014-05-06 | 2018-01-02 | Nant Holdings Ip, Llc | Image-based feature detection using edge vectors |
| CN106204477A (en) * | 2016-07-06 | 2016-12-07 | 天津大学 | Video frequency sequence background restoration methods based on online low-rank background modeling |
| CN106778814A (en) * | 2016-11-24 | 2017-05-31 | 郑州航空工业管理学院 | A kind of method of the removal SAR image spot based on projection spectral clustering |
Non-Patent Citations (4)
| Title |
|---|
| Denoising of Hyperspectral images using Group Low-Rank Representation;mengdi wang等;《IEEE》;20160304;论文正文 * |
| 低信噪比激光光斑图像快速去噪系统设计;郑洪庆等;《激光杂志》;20181031;论文正文 * |
| 单晶锗二维转鼓的加工工艺及检测方法研究;朱可可;《中国优秀硕士学位论文全文数据库 信息科技辑》;20190115;论文正文 * |
| 基于低秩近似与常量统计的记载红外多光谱扫描仪去噪方法;杨暄等;《红外技术》;20180531;论文正文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109949252A (en) | 2019-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109949252B (en) | Infrared image light spot removing method based on compensation coefficient fitting | |
| CN110444157B (en) | Gray scale compensation relation, method for acquiring compensation value and display device | |
| CN104240229B (en) | A kind of adaptive method for correcting polar line of infrared binocular camera | |
| CN104021533B (en) | A kind of real time imaging noise-reduction method and device | |
| CN105741244A (en) | Method for removing shadows and halos under weak light through indoor polling robot | |
| CN113284070A (en) | Non-uniform fog image defogging algorithm based on attention transfer mechanism | |
| CN111915558A (en) | Pin state detection method for high-voltage transmission line | |
| Tao et al. | Electric insulator detection of UAV images based on depth learning | |
| CN103886590B (en) | A kind of push-broom type remote sensing camera Atomatic focusing method based on Wavelet Packet Energy Spectrum | |
| CN113378672A (en) | Multi-target detection method for defects of power transmission line based on improved YOLOv3 | |
| CN113096103A (en) | Intelligent smoke image sensing method for emptying torch | |
| CN105005967A (en) | Method and apparatus for correcting non-uniformity of infrared imaging based on combined space-time filtering | |
| CN104502992A (en) | Weak point target precisely positioning method and system based on space-time oversampling and scanning | |
| CN110991565A (en) | Target tracking optimization algorithm based on KCF | |
| CN116128766A (en) | Improved Retinex-Net-based infrared image enhancement method for power equipment | |
| CN112926552B (en) | Remote sensing image vehicle target recognition model and method based on deep neural network | |
| CN114882072A (en) | High-speed correlation filtering target tracking method based on multi-channel image feature fusion | |
| CN111089865B (en) | Defect cable detection method based on F-RCNN | |
| Qin et al. | Fast single image dehazing with domain transformation-based edge-preserving filter and weighted quadtree subdivision | |
| CN112819843A (en) | Method and system for extracting power line at night | |
| CN112903109A (en) | Industrial temperature measurement imaging equipment based on infrared image and image processing method thereof | |
| CN111932470A (en) | Image restoration method, device, equipment and medium based on visual selection fusion | |
| CN104915937A (en) | Quick single-lens calculating imaging method based on frequency domain matrix decomposition | |
| CN104243767A (en) | Method for removing image noise | |
| CN110853001B (en) | Transformer substation foreign matter interference prevention image recognition method, system and medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |