CN113095139A - Infrared point target identification method based on Gaussian template matching - Google Patents
Infrared point target identification method based on Gaussian template matching Download PDFInfo
- Publication number
- CN113095139A CN113095139A CN202110265696.0A CN202110265696A CN113095139A CN 113095139 A CN113095139 A CN 113095139A CN 202110265696 A CN202110265696 A CN 202110265696A CN 113095139 A CN113095139 A CN 113095139A
- Authority
- CN
- China
- Prior art keywords
- template
- center
- gaussian
- pixel
- matching
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 239000011159 matrix material Substances 0.000 claims description 10
- 230000000694 effects Effects 0.000 abstract description 2
- 230000008447 perception Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 4
- 238000003331 infrared imaging Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an infrared point target identification method based on Gaussian template matching, belonging to the technical field of information perception and identification, and the method comprises the following steps: the method comprises the following steps: performing Gaussian fitting by using known data offline; digitizing and normalizing the Gaussian template obtained by fitting; step three: matching the area to be matched with the standard template to obtain a corresponding matching coefficient; step four: and judging whether the current area is a target or not according to the matching coefficient. According to the infrared point target identification method based on Gaussian template matching, the method is good in effect when cloud edges and other interferences which do not meet Gaussian distribution are eliminated.
Description
Technical Field
The invention belongs to the technical field of information perception and identification, and particularly relates to an infrared point target identification method based on Gaussian template matching.
Background
The infrared imaging system converts a thermal signal into a corresponding electric signal by receiving thermal radiation, the size of the electric signal corresponds to the level of radiation energy, the electric signal is converted into an image and then the image is output, the infrared image is obtained, the infrared imaging is mainly influenced by the gray level of the target, the material of the target, the atmospheric attenuation and other factors, the background is also imaged by the imaging system while the target is imaged, and the background comprises a sky background, a ground background, a sea background and the like according to the different environments of the target.
The energy of an ideal point target is dispersed after the diffraction of an optical system, and when the ideal point target is finally imaged on a detector, the ideal point target is not a point with a solitary zero, but a light spot with approximate Gaussian distribution, and other backgrounds do not have the characteristic. The object can thus be identified from the background on the basis of this characteristic.
Disclosure of Invention
The technical problem solved by the invention is as follows: aiming at the content, the invention provides an infrared point target identification method based on Gaussian template matching, and the method has better effect when cloud edges and other interferences which do not meet Gaussian distribution exist in the sky.
The technical scheme of the invention is as follows: an infrared point target identification method based on Gaussian template matching comprises the following steps:
the method comprises the following steps: performing Gaussian fitting by using known data offline to obtain corresponding Gaussian parameters;
step two: digitizing and normalizing the Gaussian template obtained by fitting;
step three: matching by adopting a correlation method, and matching the areas to be matched with the 5 normalized templates one by one to obtain 5 corresponding matching coefficients cov1-cov 5;
step four: and carrying out target identification on the current area according to the matching coefficient.
The specific process of the step one is as follows:
selecting a standard template of 5 x 5, selecting a corresponding image sequence, and obtaining a corresponding Gaussian parameter by utilizing Gaussian fitting; the two-dimensional gaussian equation is written in the form:
wherein G is the gray scale of the center of Gaussian, (x)0，y0) Is a central coordinate of the Gaussian template, sigmaxAnd σyRespectively, the standard deviations in two directions are obtained, the (x, y) is the coordinates of pixel points participating in fitting, the f (x, y) is the gray level of corresponding pixel points, and the logarithm is taken on two sides of the equation and multiplied by the f to obtain:
setting the number of the pixel points participating in fitting as N, wherein f represents the gray level of the pixel point corresponding to the coordinate, and the equation is written into a matrix form:
F＝BC
wherein F is a matrix of N x 1, B is a matrix of N x 5, and C is a matrix of 5 x 1, and the least square method is utilized to obtain:
C＝(BTB)-1BTF
g, x0, y0, σ are obtainedx，σyThe values of the five parameters, i.e., the gaussian parameters in step one, have 5 parameters, so that the number N of pixels that need to participate in fitting is at least 5.
The specific process of the second step is as follows:
digitizing and normalizing 5 different Gaussian templates to obtain digitized and normalized templates:
a) the Gaussian center completely falls in the center of the template of 5 x 5;
b) the Gaussian center falls on the lower right four pixels of the center of the template, and the four pixels respectively comprise: in the form
A center pixel, a template center horizontal right pixel, a template center vertical down pixel, and a template center diagonal down pixel;
c) the Gaussian center falls on the upper left four pixels of the center of the template, and the four pixels respectively comprise: in the form
A center pixel, a horizontal left pixel at the center of the template, a vertical upward pixel at the center of the template, and an upward pixel diagonally to the center of the template;
d) four pixels with a Gaussian center falling on the upper right of the center of the template respectively comprise: in the form
A center pixel, a template center horizontal right pixel, a template center vertical up pixel, and a template center diagonal up pixel;
e) the Gaussian center falls on the lower left four pixels of the center of the template, and the four pixels respectively comprise: in the form
A center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
The third step is specifically as follows:
setting a template as T, setting an image area to be matched as S, wherein the matching is to cover the template T on the S for translation, and the correlation coefficient of the template matching is as follows, wherein M is 5, and N is 5:
wherein i is 1,2,3,4, 5.
The specific process of the step four is as follows:
if the 5 matching coefficients meet the condition that one of the matching coefficients is larger than the currently set threshold value covT, the current region to be matched is considered to be matched with the template, and the target is determined;
if cov1> cov2 and cov1> cov3 and cov1> cov4 and cov1> cov5, it indicates that the Gaussian center of the current target falls completely at the center of the template;
if cov2> cov1 and cov2> cov3 and cov2> cov4 and cov2> cov5, then it indicates that the Gaussian center falls below and to the right of the center of the template;
if cov3> cov1 and cov3> cov2 and cov3> cov4 and cov3> cov5, then it indicates that the Gaussian center falls on the top left of the template center;
if cov4> cov1 and cov4> cov2 and cov4> cov3 and cov4> cov5, then it indicates that the Gaussian center falls on the top right of the center of the template;
if cov5> cov1 and cov5> cov2 and cov5> cov3 and cov5> cov4, then it indicates that the Gaussian center falls below and to the left of the center of the template;
if none of the 5 matching coefficients satisfies the above condition, the current region is not the target and is deleted.
Compared with the prior art, the invention has the advantages that: (1) the prior art generally adopts an online fitting Gaussian template, has large calculation amount and can not meet the real-time requirement in engineering application, and the method adopts an offline calculation mode of the Gaussian template, thereby reducing the calculation amount; (2) in the prior art, only one standard template with a Gaussian center falling in the center of the template is generally obtained, 5 Gaussian templates are obtained by discretization and normalization, and the condition that the Gaussian center deviates can be effectively identified; (3) the Gaussian fitting adopted by the invention can eliminate the interference and cloud edges which do not meet the Gaussian distribution.
Drawings
FIG. 1: the flow chart of the invention;
FIG. 2: the standard template with the Gaussian center falling in the center of the template is provided by the invention;
FIG. 3: the Gaussian center provided by the invention falls on the standard template below the center of the template;
FIG. 4: the Gaussian center provided by the invention falls on the standard template on the left upper part of the template center;
FIG. 5: the Gaussian center provided by the invention falls on the standard template on the right upper part of the template center;
FIG. 6: the Gaussian center provided by the invention falls on the standard template below the center of the template;
Detailed Description
The invention relates to an infrared point target identification method based on Gaussian template matching, which specifically comprises the following steps as shown in FIG. 1:
the method comprises the following steps: performing Gaussian fitting by using known data offline to obtain corresponding Gaussian parameters;
selecting a standard template of 5 x 5, selecting a corresponding image sequence, and obtaining the standard template by Gaussian fitting.
The two-dimensional gaussian equation can be written in the form:
wherein G is the gray scale of the center of Gaussian, x0，y0Is the center of the Gaussian template, σxAnd σyRespectively, the standard deviations in two directions are obtained, the (x, y) is the coordinates of pixel points participating in fitting, the f (x, y) is the gray level of corresponding pixel points, and the logarithm is taken on two sides of the equation and multiplied by the f to obtain:
taking logarithms of two sides of the equation and multiplying the logarithms by f to obtain:
the number of the pixel points participating in fitting is set as N, and the equation can be written into a matrix form:
F＝BC
the least square method is utilized to obtain:
C＝(BTB)-1BTF
g, x0, y0, sigma can be obtainedx，σyThe values of the five parameters are the parameters of the Gaussian template.
Step two: digitizing and normalizing the Gaussian template obtained by fitting;
the gaussian template is digitized and normalized to obtain 5 different digitized and normalized templates:
(1) the Gaussian center completely falls in the center of the template of 5 x 5;
(2) the Gaussian center falls on the lower right four pixels of the center of the template, and the four pixels respectively comprise: template center pixel, template center horizontal right pixel, template center vertical down pixel, and template center diagonal down pixel.
(3) The Gaussian center falls on the upper left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
(4) Four pixels with a Gaussian center falling on the upper right of the center of the template respectively comprise: template center pixel, template center horizontal right pixel, template center vertical up pixel, and template center diagonal up pixel.
(5) The Gaussian center falls on the lower left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
Step three: matching is carried out by adopting a correlation method, and the areas to be matched are matched with the 5 normalized templates one by one, so that 5 corresponding matching coefficients cov1 (the Gaussian center is completely positioned at the center of the template), cov2 (the Gaussian center is positioned at the lower right part of the center of the template), cov3 (the Gaussian center is positioned at the upper left part of the center of the template), cov4 (the Gaussian center is positioned at the upper right part of the center of the template), and cov5 (the Gaussian center is positioned at the lower left part of the center of the template) are obtained;
setting a template as T, setting an image area to be matched as S, wherein the matching is to cover the template T on the S for translation, and the correlation coefficient of the template matching is as follows, wherein M is 5, and N is 5:
wherein i is 1,2,3,4, 5.
Step four: judging whether the current area is a target or not according to the matching coefficient;
(1) the current area is a target;
if the 5 matching coefficients meet the condition that one of the matching coefficients is larger than the currently set threshold value covT, the current region to be matched is considered to be matched with the template, and the target is determined;
if cov1> cov2 and cov1> cov3 and cov1> cov4 and cov1> cov5, it indicates that the gaussian center of the current target falls completely at the center of the template;
if cov2> cov1 and cov2> cov3 and cov2> cov4 and cov2> cov5, it indicates that the gaussian center falls below and to the right of the center of the template;
if cov3> cov1 and cov3> cov2 and cov3> cov4 and cov3> cov5, it indicates that the gaussian center falls on the top left of the template center;
if cov4> cov1 and cov4> cov2 and cov4> cov3 and cov4> cov5, it indicates that the gaussian center falls on the top right of the center of the template;
if cov5> cov1 and cov5> cov2 and cov5> cov3 and cov5> cov4, it indicates that the gaussian center falls below and to the left of the template center;
(2) the current region is not a target;
if none of the 5 matching coefficients satisfies the above condition, the current region is not the target and is deleted.
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The invention provides an infrared point target identification method based on Gaussian template matching, which specifically comprises the following steps as shown in FIG. 1:
the method comprises the following steps: performing Gaussian fitting by using known data offline to obtain corresponding Gaussian parameters;
and selecting 5 x 5 standard templates, selecting corresponding image sequences, and obtaining the standard templates by Gaussian fitting.
The two-dimensional gaussian equation can be written in the form:
wherein G is the gray scale of the center of Gaussian, x0，y0Is highCenter of the template, σxAnd σyRespectively, the standard deviations in two directions are obtained, the (x, y) is the coordinates of pixel points participating in fitting, the f (x, y) is the gray level of corresponding pixel points, and the logarithm is taken on two sides of the equation and multiplied by the f to obtain:
the number of the pixel points participating in fitting is set as N, and the equation can be written into a matrix form:
F＝BC
the least square method is utilized to obtain:
C＝(BTB)-1BTF
g, x0, y0, sigma can be obtainedx，σyThe values of the five parameters are the gaussian template parameters.
Step two: digitizing and normalizing the Gaussian template obtained by fitting;
the gaussian template is digitized and normalized to obtain 5 different digitized and normalized templates:
(1) the Gaussian center completely falls in the center of the template of 5 x 5;
(2) the Gaussian center falls on the lower right four pixels of the center of the template, and the four pixels respectively comprise: template center pixel, template center horizontal right pixel, template center vertical down pixel, and template center diagonal down pixel.
(3) The Gaussian center falls on the upper left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
(4) Four pixels with a Gaussian center falling on the upper right of the center of the template respectively comprise: template center pixel, template center horizontal right pixel, template center vertical up pixel, and template center diagonal up pixel.
(5) The Gaussian center falls on the lower left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
Step three: matching is carried out by adopting a correlation method, and the areas to be matched are matched with the 5 normalized templates one by one, so that 5 corresponding matching coefficients cov1 (the Gaussian center is completely positioned at the center of the template), cov2 (the Gaussian center is positioned at the lower right part of the center of the template), cov3 (the Gaussian center is positioned at the upper left part of the center of the template), cov4 (the Gaussian center is positioned at the upper right part of the center of the template), and cov5 (the Gaussian center is positioned at the lower left part of the center of the template) are obtained;
setting a template as T, setting an image area to be matched as S, and performing matching by covering the template T on the S for translation, wherein the calculation method of the correlation coefficient of the template matching is as follows, wherein M is 5, and N is 5:
wherein i is 1,2,3,4, 5.
Step four: judging whether the current area is a target or not according to the matching coefficient;
(1) the current area is a target;
if the 5 matching coefficients meet the condition that one of the matching coefficients is larger than the currently set threshold value covT, the current region to be matched is considered to be matched with the template, and the target is determined;
if cov1> cov2 and cov1> cov3 and cov1> cov4 and cov1> cov5, it indicates that the gaussian center of the current target falls completely at the center of the template;
if cov2> cov1 and cov2> cov3 and cov2> cov4 and cov2> cov5, it indicates that the gaussian center falls below and to the right of the center of the template;
if cov3> cov1 and cov3> cov2 and cov3> cov4 and cov3> cov5, it indicates that the gaussian center falls on the top left of the template center;
if cov4> cov1 and cov4> cov2 and cov4> cov3 and cov4> cov5, it indicates that the gaussian center falls on the top right of the center of the template;
if cov5> cov1 and cov5> cov2 and cov5> cov3 and cov5> cov4, it indicates that the gaussian center falls below and to the left of the template center;
(2) the current region is not a target;
if none of the 5 matching coefficients satisfies the above condition, the current region is not the target and is deleted.
Claims (5)
1. An infrared point target identification method based on Gaussian template matching is characterized in that: the method comprises the following steps:
the method comprises the following steps: performing Gaussian fitting by using known data offline to obtain corresponding Gaussian parameters;
step two: digitizing and normalizing the Gaussian template obtained by fitting;
step three: matching by adopting a correlation method, and matching the areas to be matched with the 5 normalized templates one by one to obtain 5 corresponding matching coefficients cov1-cov 5;
step four: and carrying out target identification on the current area according to the matching coefficient.
2. The infrared point target identification method based on Gaussian template matching as claimed in claim 1, characterized in that: the specific process of the step one is as follows:
selecting a standard template of 5 x 5, selecting a corresponding image sequence, and obtaining a corresponding Gaussian parameter by utilizing Gaussian fitting; the two-dimensional gaussian equation is written in the form:
wherein G is the gray scale of the center of Gaussian, (x)0，y0) Is a central coordinate of the Gaussian template, sigmaxAnd σyRespectively, the standard deviations in two directions are obtained, the (x, y) is the coordinates of pixel points participating in fitting, the f (x, y) is the gray level of corresponding pixel points, and the logarithm is taken on two sides of the equation and multiplied by the f to obtain:
setting the number of the pixel points participating in fitting as N, wherein f represents the gray level of the pixel point corresponding to the coordinate, and the equation is written into a matrix form:
F＝BC
wherein F is a matrix of N x 1, B is a matrix of N x 5, and C is a matrix of 5 x 1, and the least square method is utilized to obtain:
C＝(BTB)-1BTF
g, x0, y0, σ are obtainedx，σyThe values of the five parameters, i.e., the gaussian parameters in step one, have 5 parameters, so that the number N of pixels that need to participate in fitting is at least 5.
3. The infrared point target identification method based on Gaussian template matching as claimed in claim 1, characterized in that: the specific process of the second step is as follows:
digitizing and normalizing 5 different Gaussian templates to obtain digitized and normalized templates:
a) the Gaussian center completely falls in the center of the template of 5 x 5;
b) the Gaussian center falls on the lower right four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontal right pixel, a template center vertical down pixel, and a template center diagonal down pixel;
c) the Gaussian center falls on the upper left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontal left pixel, a template center vertical up pixel, and a template center diagonal up pixel;
d) four pixels with a Gaussian center falling on the upper right of the center of the template respectively comprise: a template center pixel, a template center horizontal right pixel, a template center vertical up pixel, and a template center diagonal up pixel;
e) the Gaussian center falls on the lower left four pixels of the center of the template, and the four pixels respectively comprise: a template center pixel, a template center horizontally left pixel, a template center vertically up pixel, and a template center diagonally up pixel.
4. The infrared point target identification method based on Gaussian template matching as claimed in claim 3, characterized in that: the third step is specifically as follows:
setting a template as T, setting an image area to be matched as S, wherein the matching is to cover the template T on the S for translation, and the correlation coefficient of the template matching is as follows, wherein M is 5, and N is 5:
wherein i is 1,2,3,4, 5.
5. The infrared point target identification method based on Gaussian template matching as claimed in claim 4, characterized in that: the specific process of the step four is as follows:
if the 5 matching coefficients meet the condition that one of the matching coefficients is larger than the currently set threshold value covT, the current region to be matched is considered to be matched with the template, and the target is determined;
if cov1> cov2 and cov1> cov3 and cov1> cov4 and cov1> cov5, it indicates that the gaussian center of the current target falls completely at the center of the template;
if cov2> cov1 and cov2> cov3 and cov2> cov4 and cov2> cov5, it indicates that the gaussian center falls below and to the right of the center of the template;
if cov3> cov1 and cov3> cov2 and cov3> cov4 and cov3> cov5, it indicates that the gaussian center falls on the top left of the template center;
if cov4> cov1 and cov4> cov2 and cov4> cov3 and cov4> cov5, it indicates that the gaussian center falls on the top right of the center of the template;
if cov5> cov1 and cov5> cov2 and cov5> cov3 and cov5> cov4, it indicates that the gaussian center falls below and to the left of the template center;
if none of the 5 matching coefficients satisfies the above condition, the current region is not the target and is deleted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110265696.0A CN113095139B (en) | 2021-03-11 | 2021-03-11 | Infrared point target identification method based on Gaussian template matching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110265696.0A CN113095139B (en) | 2021-03-11 | 2021-03-11 | Infrared point target identification method based on Gaussian template matching |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113095139A true CN113095139A (en) | 2021-07-09 |
| CN113095139B CN113095139B (en) | 2022-07-05 |
Family
ID=76666902
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110265696.0A Active CN113095139B (en) | 2021-03-11 | 2021-03-11 | Infrared point target identification method based on Gaussian template matching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113095139B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5329599A (en) * | 1991-12-20 | 1994-07-12 | Xerox Corporation | Enhanced fidelity reproduction of images by hierarchical template matching |
| CN102567994A (en) * | 2011-12-31 | 2012-07-11 | 南京理工大学 | Infrared small target detection method based on angular point gaussian characteristic analysis |
| CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
| US20150206004A1 (en) * | 2014-01-20 | 2015-07-23 | Ricoh Company, Ltd. | Object tracking method and device |
-
2021
- 2021-03-11 CN CN202110265696.0A patent/CN113095139B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5329599A (en) * | 1991-12-20 | 1994-07-12 | Xerox Corporation | Enhanced fidelity reproduction of images by hierarchical template matching |
| CN102567994A (en) * | 2011-12-31 | 2012-07-11 | 南京理工大学 | Infrared small target detection method based on angular point gaussian characteristic analysis |
| CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
| US20150206004A1 (en) * | 2014-01-20 | 2015-07-23 | Ricoh Company, Ltd. | Object tracking method and device |
Non-Patent Citations (8)
| Title |
|---|
| 李国强: "高斯曲面拟合在对空红外点目标检测中的应用", 《红外技术》 * |
| 李国强: "高斯曲面拟合在对空红外点目标检测中的应用", 《红外技术》, vol. 35, no. 10, 31 October 2013 (2013-10-31), pages 638 - 641 * |
| 王博 等: "红外运动弱小目标的动态规划检测", 《电子科技大学学报》 * |
| 王博 等: "红外运动弱小目标的动态规划检测", 《电子科技大学学报》, vol. 38, no. 4, 31 July 2009 (2009-07-31), pages 613 - 616 * |
| 聂洪山 等: "基于高斯模板分层匹配的红外点目标检测算法", 《计算机工程与应用》 * |
| 聂洪山 等: "基于高斯模板分层匹配的红外点目标检测算法", 《计算机工程与应用》, vol. 44, no. 33, 31 December 2008 (2008-12-31), pages 191 - 193 * |
| 顾静良 等: "基于自适应模板匹配的红外弱小目标检测", 《计算机应用》 * |
| 顾静良 等: "基于自适应模板匹配的红外弱小目标检测", 《计算机应用》, 31 December 2005 (2005-12-31), pages 5 - 7 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113095139B (en) | 2022-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109753929B (en) | An image recognition method for high-speed rail insulator inspection | |
| CN107680054B (en) | Multi-source image fusion method in haze environment | |
| CN109308718B (en) | Space personnel positioning device and method based on multiple depth cameras | |
| CN109242870A (en) | A kind of sea horizon detection method divided based on image with textural characteristics | |
| CN111783773B (en) | Correction method for angle-inclined telegraph pole signboard | |
| CN115131325A (en) | Breaker fault operation and maintenance monitoring method and system based on image recognition and analysis | |
| CN107256547A (en) | A kind of face crack recognition methods detected based on conspicuousness | |
| CN110889374A (en) | Seal image processing method and device, computer and storage medium | |
| CN110276371B (en) | Container corner fitting identification method based on deep learning | |
| CN104182976B (en) | Field moving object fining extraction method | |
| CN113095139B (en) | Infrared point target identification method based on Gaussian template matching | |
| CN119027517B (en) | A rooftop distributed photovoltaic identification method, system, device and storage medium | |
| CN109461171B (en) | Infrared dim target detection algorithm based on multi-channel improved DoG filtering | |
| CN113888420B (en) | Underwater image restoration method, device and storage medium based on correction model | |
| CN113989260A (en) | An edge segmentation method of photovoltaic panels based on UAV infrared images | |
| CN117037064B (en) | Detection method and system for illegal land occupation and tillage actions based on improved SSD | |
| CN113256587A (en) | Ionized layer plasma bubble airglow image automatic processing method and device | |
| CN115035446B (en) | A kitchen rat detection and recognition method based on local MSR and target matching | |
| CN111192281A (en) | A method for automatic image generation of electrical equipment based on visual perception mechanism | |
| CN116758447A (en) | Digital twinning-based substation equipment fault analysis system | |
| CN112329796B (en) | Infrared imaging cloud detection method and device based on visual saliency | |
| CN111524098B (en) | Neural network output layer cutting and template frame size determining method based on self-organizing clustering | |
| CN111783580B (en) | Pedestrian identification method based on human leg detection | |
| CN114049622A (en) | License plate recognition method and system | |
| CN113963017A (en) | A real-time infrared weak and small target detection method, device and computer equipment |
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 |