CN103218782A - Infrared image strengthening method based on multiscale fractal characteristics - Google Patents
Infrared image strengthening method based on multiscale fractal characteristics Download PDFInfo
- Publication number
- CN103218782A CN103218782A CN201310126452XA CN201310126452A CN103218782A CN 103218782 A CN103218782 A CN 103218782A CN 201310126452X A CN201310126452X A CN 201310126452XA CN 201310126452 A CN201310126452 A CN 201310126452A CN 103218782 A CN103218782 A CN 103218782A
- Authority
- CN
- China
- Prior art keywords
- infrared image
- image
- pixel
- fractal
- yardstick
- 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
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to an infrared image strengthening method based on multiscale fractal characteristics. According to the fact that natural backgrounds in an infrared image have fractal characteristics and an artificial target has no the fractal characteristics, the infrared image strengthening method based on the multiscale fractal characteristics utilizes differences of the fractal characteristics to strengthen the artificial target, and restrains the natural background to achieve the goal of infrared strengthening. The infrared image is introduced and the infrared image is converted into a bitmap; measurement thinking for calculating fractal parameters in a carpet covering method is applied to a two-dimensional image surface, and three-dimensional texture surface formed by a pixel gray value is covered by a carpet with the thickness being 2r; calculation of each pixel in the image is based on a fractal parameter variation metric function of variance; and an obtained image is an infrared image after being strengthened. The infrared image strengthening method based on the multiscale fractal characteristics has real-time performance and robustness. The image after being processed by the method can reach good infrared image strengthening effects.
Description
Technical field
The present invention suppresses the complicated natural background in the infrared image and the application demand of man-made target enhancing towards military field.A kind of infrared image Enhancement Method based on Multi-scale Fractal is proposed.
Background technology
In the modern war, war environment is increasingly sophisticated, a large amount of between ourselves and the enemy means such as camouflage, hidden, deception and interference that adopt, because target is wanted emittance inevitably, infrared imaging sensor is by obtaining the target infrared radiation, what write down is the infrared radiation information of target self, comes the detection and Identification target by the heat radiation difference between the detection of a target and the background, has special detection and Identification camouflage ability; In addition, because its not outside emittance, thereby be difficult for being investigated or locating, stronger antijamming capability had; Have geneogenous " four anti-abilities " and the high characteristics of angle tracking accuracy.Therefore, infrared imaging sensor is widely used in military field, takes aim at IRST(infrared search-track system in tool and infrared search-track system, each combat vessel of naval as the middle heat of the various equipments such as tank of ground force), gondola, capstan head, fixed forward looking infrared system, the various missile guidances of Second Artillery Force, the cruise missile equipped in the various aircrafts of air force detect early warning etc.But in actual applications, infrared image often exists: target and surrounding environment contrast are not obvious, lack the prior imformation of target and background feature; The signal to noise ratio snr of infrared gray level image is low; Objective contour, shape, texture and boundary characteristic are generally not obvious; Infrared gray level image resolution is low; Target is partly or entirely blocked; Ambient light changes, problems such as background complexity.Therefore, how by effective infrared image Enhancement Method, complicated natural background in the infrared image is suppressed simultaneously man-made target to be strengthened, overcome the various deficiencies that infrared image exists, and then application such as infrared target detection, tracking and identification towards military field are provided support be one and be badly in need of strengthening and the problem of solution.
Summary of the invention
The present invention is directed in the military field the complicated natural background inhibition of infrared image and the application demand of man-made target enhancing.Because Fractal Geometry Model is unsuitable for describing man-made target, but the surface and the space structure of natural background can be described well in certain range scale.Utilize Multi-scale Fractal different of natural background in the infrared image and man-made target, propose a kind of infrared image Enhancement Method based on Multi-scale Fractal.The present invention includes following steps:
(x, y), this step is the basis of image subsequent treatment and operation to need after step 2, infrared image import that 32 or 8 bitmap files are converted to 24 bitmap file f.
Step 3, definition
Obtained out to out during for the calculating fractal parameter,
For more than or equal to 2 positive integer (
), the out to out number is set
=4.
Step 4, " tolerance " thought of carpet covering method (covering blanket method) the point counting shape parameter of falling into a trap is applied to the two dimensional image surface, the three-D grain surface that grey scale pixel value is constituted is " carpet " covering of 2r with thickness, defining this grain surface upper surface is U (x, y, 0), lower surface is B (x, y, 0).Yardstick r=0 is set, the upper and lower surface under 0 yardstick is carried out initialization, obtain after the initialization U (x, y, 0) of each pixel and B (x, y, 0) in the image.
The U (x, y, 0) and B (x, y, 0) of step 5, each pixel of obtaining according to step 4, calculate each yardstick r (
) under, the U of each pixel in the image (x, y, r) with B (x, y, r), U (x, y, r) with B (x, y are that yardstick gets 1,2 respectively successively r) ...,
The time, the gray-scale value of upper and lower surface.
(r) (x, y r), calculate under each yardstick r the U of step 6, each pixel of obtaining according to step 5 with B for x, y, the V of each pixel in the image (x, y, r) with A (x, y, r), V (x wherein, y, the r) volume between the upper and lower surface, (x, y r) are the shared whole surface area of texture to A.Wherein
(1)
Step 7, (r) (x, y r), calculate under each yardstick r, and (r), wherein (x, y are fractal parameters r) to K to the K of each pixel in the image, are called D dimension area for x, y with A for x, y to obtain the V of each pixel according to step 6.Wherein
The K of each pixel in step 8, the image that obtains according to step 7 (x, y, r) calculate branch shape parameter measure of variation function MFFK based on variance (X, Y).Wherein
(4)
MFFK be
Range scale in D dimension area (
K) intensity of variation, utilize MFFK to realize outstanding man-made target and the difference of natural background on fractal characteristic.The MFFK image that obtains is the infrared image after the enhancing.
Step 9, the infrared image after will strengthening are derived.
The invention has the beneficial effects as follows: have fractal characteristic according to the natural background in the infrared image, and man-made target does not have this feature of fractal characteristic, utilize the difference of this fractal characteristic, man-made target is strengthened, simultaneously natural background is suppressed, reach the purpose of infrared enhancing, checking shows by experiment, when the out to out value gets 4, infrared image strengthens 0.52 second time, compare with other existing 4 kinds of fractal characteristics, the MFFK fractal characteristic has good infrared image and strengthens effect, and the while is requirement of real time also.Robustness shows as: under different weather conditions, target type, detection range, target sizes condition, algorithm can both well strengthen infrared image.
Description of drawings
Fig. 1 is the inventive method process flow diagram.
Fig. 2 is figure image intensifying result and the required time under the different out to out value, wherein Fig. 2 (a) is original infrared image 1(oil tanker), Fig. 2 (b) is the enhancing design sketch of original infrared image 1 in out to out value=4 o'clock, enhancing time t=0.4s, Fig. 2 (c) is the enhancing design sketch of original infrared image 1 in out to out value=7 o'clock, enhancing time t=1.2s, Fig. 2 (d) are the enhancing design sketch of original infrared image 1 in out to out value=10 o'clock, strengthen time t=2.6s.
Fig. 3 is that the infrared image of different fractal characteristics strengthens effect relatively, wherein Fig. 3 (a) is original infrared image 2(passenger steamer), Fig. 3 (b) is that FD strengthens design sketch, strengthens time t=0.72s, Fig. 3 (c) is that K strengthens design sketch, enhancing time t=0.72s, Fig. 3 (d) are that FMFE strengthens design sketch, strengthen time t=0.75s, Fig. 3 (e) is that MFFD strengthens design sketch, enhancing time t=0.50s, Fig. 3 (f) are that MFFK strengthens design sketch, strengthen time t=0.52s.
Fig. 4 is that the infrared image enhanced robust compares, wherein Fig. 4 (a) is original infrared image 3(tank), Fig. 4 (b) is original infrared image 4(fighter plane), Fig. 4 (c) is that original infrared image 3 strengthens design sketch, Fig. 4 (d) is that original infrared image 4 strengthens design sketch, Fig. 4 (e) is original infrared image 5(people, building), Fig. 4 (f) is original infrared image 6(naval vessel), Fig. 4 (g) is that original infrared image 5 strengthens design sketch, Fig. 4 (h) is that original infrared image 6 strengthens design sketch, Fig. 4 (i) is original infrared image 7 (automobiles, advertisement), Fig. 4 (j) is original infrared image 8 (boats and ships, bridge), Fig. 4 (k) is that original infrared image 7 strengthens design sketch, and Fig. 4 (l) is that original infrared image 8 strengthens design sketch.
Fig. 5 strengthens the inefficacy example for infrared image.Wherein Fig. 5 (a) is original infrared image 9, and Fig. 5 (b) is original infrared image 10, and Fig. 5 (c) is that original infrared image 9 strengthens design sketch, and Fig. 5 (d) is that original infrared image 10 strengthens design sketch.
Embodiment
The invention will be further described below in conjunction with accompanying drawing.
As shown in Figure 1, the present invention includes following steps:
(x, y), this module is the basis of image subsequent treatment and operation to need after step 2, infrared image import that 32 or 8 bitmap files are converted to 24 bitmap file f.
Step 3, definition
Obtained out to out during for the calculating fractal parameter,
For more than or equal to 2 positive integer (
), the out to out number is set
=4.
Step 4, " tolerance " thought of carpet covering method (covering blanket method) the point counting shape parameter of falling into a trap is applied to the two dimensional image surface, the three-D grain surface that grey scale pixel value is constituted is " carpet " covering of 2r with thickness, defining this grain surface upper surface is U (x, y, 0), lower surface is B (x, y, 0).Yardstick r=0 is set, the upper and lower surface under 0 yardstick is carried out initialization, obtain after the initialization U (x, y, 0) of each pixel and B (x, y, 0) in the image.
The U (x, y, 0) and B (x, y, 0) of step 5, each pixel of obtaining according to step 4 calculate under each yardstick r, the U of each pixel in the image (x, y, r) with B (x, y, r), U (x, y, r) with B (x, y are that yardstick gets 1,2 respectively successively r) ...,
The time, the gray-scale value of upper and lower surface.
The U of step 6, each pixel of obtaining according to step 5 (x, y, r) with B (x, y, r), calculate each yardstick r (1<=r<=
) under, the V of each pixel in the image (x, y, r) with A (x, y, r), wherein V (x, y, the r) volume between the upper and lower surface, (x, y r) are the shared whole surface area of texture to A.Wherein
Step 7, (r) (x, y r), calculate under each yardstick r, and (r), wherein (x, y are fractal parameters r) to K to the K of each pixel in the image, are called D dimension area for x, y with A for x, y to obtain the V of each pixel according to step 6.Wherein
The K of each pixel in step 8, the image that obtains according to step 7 (x, y, r) calculate branch shape parameter measure of variation function MFFK based on variance (X, Y).Wherein
MFFK be
Range scale in D dimension area (
K) intensity of variation, utilize MFFK to realize outstanding man-made target and the difference of natural background on fractal characteristic.The MFFK image that obtains is the infrared image after the enhancing.
Step 9, the infrared image after will strengthening are derived.
Say that further in the step 1, infrared image importing process concrete steps are as follows:
Need to select the infrared image of importing, next the infrared image type is carried out validity checking, reexamine if non-compliant words turn back to last layer, the infrared image inspection is removed existing data by the back.
In the step 2, the infrared image that imports is converted to 24 bit image f (x, y) concrete steps are as follows: the reading images file data, obtain information header pointer and pixel pointer, next initialisation image information, for pixel data application storage space and do number of colours conversion, change 8,32 information header unifications into 24 and handle.
1), according to the thought of being retouched in the carpet covering method, it is that " carpet " of 2r covers that two dimensional image can be supposed with a thickness, carpet be by its upper surface U (x, y, r) and lower surface B (x, y, r) definition.
2), (x y) is arbitrarily pixel (x, y) grey scale pixel value of place correspondence in the infrared image in the step 2 to f
3), respectively be 1,2 when yardstick ..., when r changed, the definition of upper and lower surface gray-scale value was respectively:
Be the out to out of being got when calculating fractal parameter,
, the out to out number is set here
=4.
In the step 4, and the U of each pixel of initialization (x, y, 0) and B (x, y, 0)=f (x, y); Concrete steps are as follows: it is 0 that yardstick r is set, and the upper and lower surface under 0 yardstick is carried out initialization, has
In the step 5, calculate under each yardstick r, the U of each pixel in the image (x, y, r) with B (x, y, r); Concrete steps are as follows:
A r at concrete asks for the whole U(x under this yardstick, y, r) and B(x, y, r).
What AA represented is top-surface camber (x, y, r-1) maximum gradation value in eight fields of picture element.
Top-surface camber (x, y, gray-scale value r-1) be top-surface camber (x, y, r-1)+1 and the maximal value of AA in the two.
What BB represented is lower surface camber (x, y, r-1) minimum gradation value in eight fields of picture element.
Lower surface camber (x, y, gray-scale value r-1) be lower surface camber (x, y, r-1)-1 and the minimum value of BB in the two.
In the step 6, calculate under each yardstick r, the V of each pixel in the image (x, y, r) with A (r) concrete steps are as follows for x, y:
At a concrete r, ask for whole V under this yardstick (x, y, r) and A (x, y, r), at first make V (x, y, r)=0, the x scope:
+ 1 to M-
, the y scope:
+ 1 to N-
, next according to the U that obtains in the step 5 (x, y, r) with B (x, y, r) calculate each pixel in the image V (x, y, r) with A (x, y, r).
V(x,y,r)= V(x,y,r+1)+U(x,y,r)-B(x,y,r) (1)
A(x,y,r)= V(x,y,r)/2*r (2)
In the step 7, calculate under each yardstick r, the K of each pixel in the image (r) concrete steps are as follows for x, y:
1), according to Fractal Geometry Theory, the relation between fractal measure and the yardstick be can be described as by the Richardson law:
Wherein:
The expression yardstick,
Be illustrated in estimating of yardstick r;
The expression fractal dimension;
KIt also is a fractal parameter;
It is topological dimension.
2), for the two dimensional gray image, it can be described as:
Wherein:
The expression yardstick,
Be illustrated in yardstick
Under the surface area on gradation of image surface estimate;
Fractal dimension under the expression different scale;
Also be a fractal parameter, be also referred to as D dimension area.
According to Fractal Geometry Theory as can be known, for one desirable fractal, its fractal dimension
FDBeing and the irrelevant amount of all yardsticks, is a constant all the time.Therefore, suppose under the different scales of measurement, in formula (3) and (4)
FDBe constant, and establish
, then can get by formula (3) and (4), when the scale of measurement is r, its corresponding D dimension area
For,
In the step 8, calculate branch shape parameter measure of variation function MFFK based on variance (X, Y) concrete steps are as follows:
1), for the D dimension area of outstanding natural scene and man-made target (
K) with yardstick
The difference that changes and show define a fractal parameter measure of variation function, that is: with D tie up area (
K) relevant Multi-scale Fractal (
MFFK, multi-scale fractal feature related with
K) be,
Wherein:
The out to out of being got when being the actual computation fractal parameter,
MFFKCan be understood as
Range scale in D dimension area (
K) intensity of variation, utilize
MFFKRealize outstanding man-made target and the difference of natural background on fractal characteristic.
2), by being provided with after the suitable out to out, to each pixel in the original infrared image, carry out Multi-scale Fractal
MFFKCalculate, and generate corresponding Multi-scale Fractal
MFFKImage can realize that infrared image strengthens.
In the step 9, it is as follows that infrared image is derived concrete steps:
1), judges whether image processes, if also not processed then turn back to step 3-8 and carry out figure image intensifying and dividing processing.
2), if infrared image processes, be 8 or 32 infrared images with image transitions then, the preservation of enhancing image after will change then.
As shown in Figure 2, be figure image intensifying result and required time under the different out to out values, the algorithm computation complexity depends on the out to out number.Therefore, how to guarantee that infrared image enhancing effect satisfies under the prerequisite of algorithm real-time requirement simultaneously, determining that suitable out to out number is crucial.Show by a large amount of experiments: along with the increase of out to out number, infrared image strengthens the not significant increase of effect, but operand is increasing.In fact, after the out to out value was greater than 10, along with the increase of out to out number, infrared image strengthened effect on the contrary worse and worse.This is because the statistical nature of fractal Brown motion model is just better identical with the statistical nature of practical natural scene image in the small scale scope.
When out to out value value is 4, can obtain best infrared image; View picture infrared image (320 * 240) was strengthened required time 0.52 second, can satisfy in the practical application per second and handle 1 two field picture real-time requirement.
The infrared image that is illustrated in figure 3 as different fractal characteristics strengthens effect relatively, compared to 4 kinds of fractal characteristics that existed,
MFFKFractal characteristic has best infrared image and strengthens effect, and the while is requirement of real time also.
Be illustrated in figure 4 as the analysis of infrared image enhanced robust, robustness shows as: under different weather conditions, target type, detection range, target sizes condition, algorithm can both finely strengthen infrared image,
1), the robustness of target type.As Fig. 2, Fig. 3 and figure image intensifying result shown in Figure 4, this algorithm all has good increase effect to man-made targets such as people, automobile, naval vessel, aircraft, tank, buildings; Obtained good restraining for natural background (mountain range, cloud, the water surface, plant, topography and geomorphology etc.).Thereby verified for dissimilar man-made target enhanced robust.
2), the robustness of target sizes.Point target (at a distance): as Fig. 4 (d) with (h), for detection range situation of (aircraft, naval vessel, guided missile etc.) greater than more than 5 kilometers the time, man-made target is rendered as point target, algorithm has for point target and very significantly strengthens effect, and the number of pixels that shows as the expression target increases and the enhancing of target gray area calibration; Appearance mark (middle distance): as Fig. 4 (c), (g), (k) with (l), for detection range situation of (aircraft, naval vessel, tank, building, billboard, bridge etc.) 3 ~ 5 kilometers the time, man-made target is rendered as the appearance mark, algorithm has equally and very significantly strengthens effect, strengthens effect and shows as and provide appearance target profile.Thereby verified enhancing robustness for the big or small target of difference.
3), weather robustness.Infrared image among Fig. 2, Fig. 3 and Fig. 4 is gathered respectively from different weather conditions such as daytime, night, the moon, fine, dense fog, middle mist, mist, light rain, experiment show algorithm have robustness for weather.
Be illustrated in figure 5 as infrared image and strengthen the analysis of effect applicable elements, as a kind of infrared image Enhancement Method, it has its limitation equally.For detection range during less than 3 kilometers (building, naval vessel, bridge, tank etc.), our imaging of man-made target interested in image occupies big zone, when also having a large amount of artificial background (as building, bunding etc.) situation in the image simultaneously, large tracts of land all is the man-made target zone in the view picture infrared image, and natural background only occupies few zone.As shown in Figure 5, infrared image at this moment strengthens poor effect.
This is because be applicable to European geometric description for man-made target, and natural background is applicable to the fractal geometry description.When the natural background in the image is taken as the leading factor, strengthen principle, desirable good enhancing effect according to infrared image based on Multi-scale Fractal; Otherwise, when the man-made target in the image is taken as the leading factor, the good enhancing effect that can not get.
Through a large amount of experiment and statistical study, when natural background in the infrared image (occupying 60% above pixel of image) or man-made target (occupy and be less than 30% pixel), use this paper method can obtain extraordinary infrared image and strengthen effect.Most military affairs, civil area application scenarios as security protection of missile defence, IRST system, land and sea border defense, ship collision prevention, key area etc., satisfy above-mentioned applicable elements.
Claims (2)
1. based on the infrared image Enhancement Method of Multi-scale Fractal, it is characterized in that this method may further comprise the steps:
Step 1, the infrared image that detects is handled, specifically: import infrared image, the infrared image that takes out need be imported scene;
After step 2, infrared image import 32 or 8 bitmap files be converted to 24 bitmap file f (x, y), this step is the basis of image subsequent treatment and operation;
Step 3, definition
Obtained out to out during for the calculating fractal parameter,
For more than or equal to 2 positive integer, the out to out number is set
=4;
Step 4, fall into a trap " tolerance " thought of point counting shape parameter of carpet covering method is applied to the two dimensional image surface, the three-D grain surface that grey scale pixel value is constituted is " carpet " covering of 2r with thickness, and defining this grain surface upper surface is U (x, y, 0), lower surface is B (x, y, 0); Yardstick r=0 is set, the upper and lower surface under 0 yardstick is carried out initialization, obtain after the initialization U (x, y, 0) of each pixel and B (x, y, 0) in the image;
The U (x, y, 0) and B (x, y, 0) of step 5, each pixel of obtaining according to step 4 calculate under each yardstick r, the U of each pixel in the image (x, y, r) with B (x, y, r), U (x, y, r) with B (x, y are that yardstick gets 1,2 respectively successively r) ...,
The time, the gray-scale value of upper and lower surface,
(r) (x, y r), calculate under each yardstick r the U of step 6, each pixel of obtaining according to step 5 with B for x, y, the V of each pixel in the image (x, y, r) with A (x, y, r), V (x wherein, y r) is volume between the upper and lower surface, and (x, y r) are the shared whole surface area of texture to A;
Step 7, (r) (x, y r), calculate under each yardstick r, and (r), wherein (x, y are fractal parameters r) to K to the K of each pixel in the image, are called D dimension area, wherein for x, y with A for x, y to obtain the V of each pixel according to step 6
The K of each pixel in step 8, the image that obtains according to step 7 (x, y, r) calculate branch shape parameter measure of variation function MFFK based on variance (X, Y); Wherein
MFFK be
Range scale in D dimension area (
K) intensity of variation, utilize MFFK to realize outstanding man-made target and the difference of natural background on fractal characteristic; The MFFK image that obtains is the infrared image after the enhancing;
Step 9, the infrared image after will strengthening are derived.
2. according to right 1 described infrared image Enhancement Method, it is characterized in that: during out to out value value 4, obtain best infrared image based on Multi-scale Fractal; It is 0.52 second that view picture infrared image 320 * 240 is strengthened required time.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310126452XA CN103218782A (en) | 2013-04-11 | 2013-04-11 | Infrared image strengthening method based on multiscale fractal characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310126452XA CN103218782A (en) | 2013-04-11 | 2013-04-11 | Infrared image strengthening method based on multiscale fractal characteristics |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103218782A true CN103218782A (en) | 2013-07-24 |
Family
ID=48816534
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310126452XA Pending CN103218782A (en) | 2013-04-11 | 2013-04-11 | Infrared image strengthening method based on multiscale fractal characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103218782A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107016652A (en) * | 2017-03-14 | 2017-08-04 | 中国人民解放军空军工程大学 | A kind of infrared noise based on Analysis On Multi-scale Features suppresses and detail enhancing method |
CN108038856A (en) * | 2017-12-22 | 2018-05-15 | 杭州电子科技大学 | Based on the infrared small target detection method for improving Multi-scale Fractal enhancing |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102426698A (en) * | 2011-11-04 | 2012-04-25 | 中国航空工业集团公司洛阳电光设备研究所 | Method for enhancing infrared image |
-
2013
- 2013-04-11 CN CN201310126452XA patent/CN103218782A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102426698A (en) * | 2011-11-04 | 2012-04-25 | 中国航空工业集团公司洛阳电光设备研究所 | Method for enhancing infrared image |
Non-Patent Citations (1)
Title |
---|
刘俊: "面向船舶避碰预警的红外运动船舶检测与跟踪", 《光电工程》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107016652A (en) * | 2017-03-14 | 2017-08-04 | 中国人民解放军空军工程大学 | A kind of infrared noise based on Analysis On Multi-scale Features suppresses and detail enhancing method |
CN108038856A (en) * | 2017-12-22 | 2018-05-15 | 杭州电子科技大学 | Based on the infrared small target detection method for improving Multi-scale Fractal enhancing |
CN108038856B (en) * | 2017-12-22 | 2020-08-04 | 杭州电子科技大学 | Infrared small target detection method based on improved multi-scale fractal enhancement |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111027496B (en) | Infrared dim target detection method based on space-time joint local contrast | |
Huang et al. | A novel method for speckle noise reduction and ship target detection in SAR images | |
CN102109597B (en) | Method for identifying type of ship according to high-resolution SAR (synthetic aperture radar) image of ship | |
CN107505614A (en) | A kind of object detection method based on the fusion of ATI SAR image width phase information | |
Bürgmann et al. | Matching of TerraSAR-X derived ground control points to optical image patches using deep learning | |
Panico et al. | SAR-based vessel velocity estimation from partially imaged Kelvin pattern | |
CN104166838B (en) | Layering High Resolution SAR image Ship Detection based on multi-layer C FAR | |
US10473429B1 (en) | Projectile detection system and method | |
CN103049905B (en) | Utilize and singly drill the three-component diameter radar image method for registering of signal | |
Sablina et al. | An approach to the image superimposition problem in multispectral computer vision systems | |
Shaoqing et al. | The comparative study of three methods of remote sensing image change detection | |
EP2054835B1 (en) | Target orientation | |
Demir et al. | Extraction of coastlines with fuzzy approach using SENTINEL-1 SAR image | |
Xu et al. | The comparative study of three methods of remote sensing image change detection | |
CN103218782A (en) | Infrared image strengthening method based on multiscale fractal characteristics | |
Hwang et al. | An efficient ship detection method for KOMPSAT-5 synthetic aperture radar imagery based on adaptive filtering approach | |
CN106910178B (en) | Multi-angle SAR image fusion method based on tone statistical characteristic classification | |
CN107729903A (en) | SAR image object detection method based on area probability statistics and significance analysis | |
CN111105419A (en) | Vehicle and ship detection method and device based on polarized SAR image | |
Heymann et al. | Radar image processing and AIS target fusion | |
Wang et al. | Terrain matching by fusing HOG with Zernike moments | |
Zhang et al. | A novel method for estimation of the target rotation angle in SAR image | |
Steinvall et al. | Simulating the performance of laser imaging and range profiling of small surface vessels | |
Zhang et al. | Target detection in sar images based on sub-aperture coherence and phase congruency | |
Garestier et al. | Anisotropic surface detection over coastal environment using near-IR LiDAR Intensity maps |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130724 |