CN105139433B - Infrared DIM-small Target Image sequence emulation mode based on mean value model - Google Patents
Infrared DIM-small Target Image sequence emulation mode based on mean value model Download PDFInfo
- Publication number
- CN105139433B CN105139433B CN201510473934.1A CN201510473934A CN105139433B CN 105139433 B CN105139433 B CN 105139433B CN 201510473934 A CN201510473934 A CN 201510473934A CN 105139433 B CN105139433 B CN 105139433B
- Authority
- CN
- China
- Prior art keywords
- infrared
- background
- target image
- small
- image
- 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
Landscapes
- Transforming Light Signals Into Electric Signals (AREA)
Abstract
The invention discloses a kind of Infrared DIM-small Target Image sequence emulation mode based on mean value model.Specific steps include：1. gather image background；2. simulation objectives image background；3. mirror-extended；4. create object module；5. targetpath is set；6. synthesize target image；7. a couple emulation Infrared DIM-small Target Image background frame number i carries out cumulative 1 operation；8. judge whether cumulative 1 post-simulation Infrared DIM-small Target Image background frame number is equal to emulation Infrared DIM-small Target Image sequence totalframes L, if it is, performing step 9；Otherwise, step 2 is performed；9. output emulation Infrared DIM-small Target Image sequence.The present invention has amount of calculation small, emulates Infrared DIM-small Target Image validity height, and texture characteristics enrich and infrared small object size, infrared small object signal to noise ratio, infrared small object movement velocity, the controllable advantage of infrared small object flight path.
Description
Technical Field
The invention belongs to the technical field of image processing, and further relates to an infrared small and weak target image sequence simulation method based on a mean value model in the technical field of infrared image processing. The invention can provide a large number of infrared weak and small target test image sequences for performance evaluation of the high-altitude long-distance infrared weak and small target detection and tracking algorithm, and is used for verifying the effectiveness of the detection and tracking algorithm.
Background
The infrared imaging sensor is widely applied to modern weaponry, the infrared weak and small target image sequence actually shot by the infrared imaging sensor cannot know the specific position of the infrared weak and small target in advance, and the acquisition of a large number of infrared weak and small target image sequences under different types of infrared backgrounds has great difficulty. But a large number of infrared dim target image sequences with known target positions are needed when the performance of the infrared dim target detection and tracking algorithm is verified. The problem needs to be solved by using an infrared weak and small target image sequence simulation technology.
At present, it is an effective technical method to generate the corresponding infrared dim target image sequence by computer modeling simulation software. The infrared small and weak target image sequence generated by the method lacks texture detail information, and has a large error with a real infrared small and weak target image sequence. And an effective infrared small target image sequence cannot be provided for the performance evaluation of the subsequent infrared small target detection and tracking algorithm.
A large amount of manpower and material resources are consumed in the process of outfield experiments generally, stability and repeatability are poor, weather influence is large, a semi-physical simulation platform is built through a computer, infrared dim target image sequences of infrared dim targets under different infrared backgrounds can be simulated, the simulation platform can be realized under laboratory conditions, convenience and rapidness are realized, and manpower and material resources are saved.
The patent of Huazhong science and technology university "an infrared scene image generation method based on scene classification" (patent application No. 201110110316.2, application publication No. CN 102270355A) discloses an infrared scene image generation method based on scene classification, which belongs to a method for generating and simulating infrared scene images through a visible light remote sensing image, a terrain digital elevation model and a target three-dimensional model, and the method firstly classifies scenes and specifies texture types; then, establishing a scene three-dimensional model for the classified scene images and the digital elevation model; then, carrying out automatic mapping of texture materials in batches on the sequence texture file of the scene three-dimensional model and generating a corresponding sequence texture material mapping file; and finally, loading an atmospheric parameter model, an imaging model and atmospheric parameter conditions to finish the simulation output of the infrared scene image, wherein the atmospheric parameter model is obtained by calculating through mature commercial software. The invention can provide infrared scene data required by simulation training for various aircrafts using infrared imaging technology, reduce flight training cost, provide algorithm experiment and verification image for development of infrared imaging guided weapon system, and improve development efficiency, and the method has the following defects: when the infrared scene image is generated through simulation, the scene is required to be classified and the texture type is specified, a scene three-dimensional model is established, a mapping file is generated, an atmospheric parameter model, an imaging model and atmospheric parameter conditions are loaded, simulation output of the infrared scene image is completed, the steps are complex, the calculation amount is large, the generated infrared weak and small target image sequence lacks texture detail characteristics, and the generated infrared weak and small target image sequence has larger distortion compared with a real infrared weak and small target image.
A patent applied by shenyang automation research institute of china academy of sciences, "a real-time simulation method for infrared dynamic scenes of multiple targets in a marine background" (patent application No. 201110447686.5, application publication No. CN 103186906A) discloses a real-time simulation method for infrared dynamic scenes of multiple targets in a marine background. The method comprises the steps of infrared modeling, scene construction, infrared calculation, infrared atmospheric transmission calculation, real-time infrared scene generation and final rendering to generate an infrared scene real-time simulation image. The method has rich simulation scene contents, and can contain sea surface, sky, ships and airplanes of various models in the same scene. The real-time requirements can still be met under the condition of a complex scene, a dynamic scene is built in real time according to current data, and the sound field infrared scene image is simulated in real time. The method has the following defects: the size of the infrared dim target, the signal-to-noise ratio of the infrared dim target, the movement speed of the infrared dim target and the track of the infrared dim target cannot be controlled.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, and provides an infrared small and weak target image sequence simulation method based on a mean value model.
In order to achieve the purpose, the method comprises the following specific steps:
(1) Acquiring an image background:
(1a) Acquiring a real infrared image background sequence by using a thermal infrared imager to obtain an original infrared image background sequence;
(1b) Assigning the frame number i of the original infrared image background sequence to be 1;
(2) Simulating the background of the infrared small dim target image:
(2a) Reading in an ith original infrared image background in the original infrared image background sequence, wherein i is more than 0 and less than or equal to R, and R represents the total frame number of the original infrared image background sequence;
(2b) Setting an initial row and an initial column for cutting the background of the ith frame of original infrared image;
(2c) Cutting an ith frame of original infrared image background to obtain a simulated infrared small target image background with the size of MxN, wherein M represents the line number of the simulated infrared small target image background, N represents the column number of the simulated infrared small target image background, M is more than or equal to 0 and less than or equal to A, N is more than or equal to 0 and less than or equal to B, A represents the line number of each frame of image in the original infrared image background sequence, and B represents the column number of each frame of image in the original infrared image background sequence;
(3) Carrying out mirror image expansion on the background edge of the simulated infrared small and weak target image;
(4) Creating an infrared small target model:
(4a) Setting the signal-to-noise ratio D of the infrared weak target, wherein D is an integer value within the range of more than 0 and less than or equal to 20;
(4b) Calculating the gray value of the infrared weak and small target pixel according to the following formula:
s＝D×σ+μ
wherein s represents the gray value of the pixel of the infrared small dim target, D represents the set signal-to-noise ratio of the infrared small dim target, sigma represents the standard deviation of all pixels in a local neighborhood taking the centroid of the infrared small dim target as the center, and mu represents the mean value of all pixels in a local neighborhood range taking the centroid of the infrared small dim target as the center;
(4c) Setting the size of the infrared dim target as m multiplied by m pixels, wherein m is an integer value within the range of more than 0 and less than or equal to 10;
(5) Setting an infrared small target track:
(5a) Setting the moving speed of the infrared weak and small target as delta x pixels/frame, wherein delta x is valued in the range of 0< delta x < 10;
(5b) Taking a pixel point at the top left corner vertex of the background of the ith frame of simulated infrared dim target image as an origin, taking the horizontal rightward direction as an x axis and taking the vertical downward direction as a y axis, and establishing a background track coordinate system of the ith frame of simulated infrared dim target image;
(5c) Judging whether the background frame number of the simulated infrared dim target image is 1, if so, executing the step (5 d); otherwise, executing step (5 e);
(5d) Setting the abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system as x 1 ,0＜x 1 N is less than or equal to N, wherein N represents the column number of the background of the simulated infrared small target image, and a linear model of the background track of the 1 st frame of the simulated infrared small target image is established according to the following formula:
y 1 ＝ax 1 +b
wherein, y 1 The vertical coordinate value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and y is more than 0 1 M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, x 1 The abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 1 N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, a represents the slope of the linear model of the background track of the simulated infrared dim target image, b represents the intercept of the linear model of the background track of the simulated infrared dim target image, and the value ranges of a and b are as follows: 0< ax 1 + b is less than or equal to M, and M represents the line number of the background of the simulated infrared weak and small target image;
(5e) According to the following formula, establishing a straight line model of the background track of the simulation infrared dim target image except the 1 st frame:
wherein x is i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, wherein N represents the column number, x, of the background of the simulated infrared dim target image i-1 The abscissa value of the infrared dim target in the i-1 frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i-1 N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, deltax represents the movement speed of the infrared dim target, and Deltax is 0<Δx&Value of y within 10 i The longitudinal coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, and y is more than 0 i M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, a represents the slope of the linear model of the infrared dim target track, b represents the intercept of the linear model of the infrared dim target track, and the value ranges of a and b are as follows: 0< ax i +b≤M；
(6) Synthesizing a simulated infrared weak and small target image:
(6a) And setting the total frame number of the simulated infrared dim target image sequence according to the following formula:
0＜L≤min((N-x 1 )/Δx,R)
wherein, L represents the total frame number of the simulated infrared dim target image sequence, min (·) represents the minimum operation, N represents the column number of the background of the simulated infrared dim target image, and x 1 The abscissa value of the infrared dim target in the 1 st frame of the simulated infrared dim target image sequence is represented, the delta x represents the movement speed of the infrared dim target, and the delta x is 0<Δx&Taking values within 10 ranges, wherein R represents the total frame number of the original infrared image background sequence;
(6b) Assigning the gray value s of the infrared weak and small target pixel in the step (4 b) to the pixel in the m multiplied by m pixel block, wherein the m multiplied by m pixel block is (x) i ，y i ) The pixel blocks are pixel blocks of the infrared weak small target centroid, m is an integer value within the range of more than 0 and less than or equal to 10, and a simulated infrared weak small target image is obtained, x i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, y i The longitudinal coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, and y is more than 0 i If the number of lines is less than M, M represents the number of lines of the background of the simulated infrared dim target image;
(7) Performing accumulation 1 operation on the background frame number i of the simulated infrared small and weak target image;
(8) Judging whether the background frame number of the simulated infrared dim target image after 1 accumulation is equal to the total frame number L +1 of the simulated infrared dim target image sequence, if so, executing the step (9); otherwise, executing the step (2);
(9) And outputting a simulated infrared small target image sequence.
Compared with the prior art, the invention has the following advantages:
firstly, because the invention uses the high-performance thermal infrared imager to collect the real infrared image background sequence as the original infrared image background sequence, the invention overcomes the defects that in the prior art, when an infrared scene image is generated by simulation, the scene is required to be classified and the texture type is specified, a scene three-dimensional model is established, a mapping file is generated, an atmospheric parameter model, an imaging model and atmospheric parameter conditions are loaded, the simulation output of the infrared scene image is completed, the steps are complex, the calculated amount is large, the generated infrared weak and small target image sequence lacks the texture detail characteristics, and the defect of larger distortion exists compared with the real infrared weak and small target image, so that the invention has the advantages of simple steps, small calculated amount, high fidelity of the simulated infrared weak and small target image, and rich texture detail characteristics.
Secondly, because the modeling simulation method is used, the defects that the size of the infrared dim target, the signal-to-noise ratio of the infrared dim target, the movement speed of the infrared dim target and the flight path of the infrared dim target cannot be controlled in the prior art are overcome, so that the method has the advantages that the size of the infrared dim target, the signal-to-noise ratio of the infrared dim target, the movement speed of the infrared dim target and the flight path of the infrared dim target can be controlled.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a background image of the original IR image of frame 50 according to the present invention;
FIG. 3 is a schematic view of a flight path of a simulated infrared small dim target according to the present invention;
FIG. 4 is a schematic diagram of a neighborhood of 9 × 9 size centered on the centroid of the small infrared target in the 50 th frame of small infrared target image according to the present invention;
fig. 5 is a schematic diagram of a 50 th frame of simulated infrared small and weak target image according to the present invention.
Detailed Description
The present invention is described in further detail below with reference to the attached drawing figures.
Referring to fig. 1, the specific steps of the present invention are as follows.
Step 1, collecting an infrared image background.
And step 1, acquiring a real infrared image background sequence by using a thermal infrared imager to obtain an original infrared image background sequence.
In the implementation of the invention, the real infrared image background sequence is an infrared sky background image sequence. Fig. 2 shows a 50 th frame of infrared sky background image in the infrared sky background image sequence, where the sky background shown in fig. 2 has rich texture features.
And step 2, assigning the frame number i of the original infrared image background sequence as 1.
And 2, simulating the infrared dim target image background.
Reading in an ith original infrared image background in the original infrared image background sequence, wherein i is more than 0 and less than or equal to R, and R represents the total frame number of the original infrared image background sequence.
In the implementation of the invention, the total frame number R of the original infrared image background sequence is 1000.
And 2, setting a starting row and a starting column for cutting the background of the ith frame of original infrared image, wherein the starting row for cutting the background of the ith frame of original infrared image is more than 0 and less than the row number of the background of the original infrared image, and the starting column for cutting the background of the ith frame of original infrared image is more than 0 and less than the column number of the background of the original infrared image.
In the implementation of the present invention, an initial row 150 for the background cropping of the i-th frame of original infrared image is set, and an initial column 150 for the background cropping of the i-th frame of original infrared image is set.
And 3, cutting the original infrared image background of the ith frame to obtain an artificial infrared weak and small target image background with the size of MxN, wherein M represents the line number of the artificial infrared weak and small target image background, N represents the column number of the artificial infrared weak and small target image background, M is more than or equal to 0 and less than or equal to A, N is more than or equal to 0 and less than or equal to B, A represents the line number of each frame image in the original infrared image background sequence, and B represents the column number of each frame image in the original infrared image background sequence.
In the implementation of the invention, the original infrared image background with the size of 480 multiplied by 625 of the ith frame is cut to obtain the simulated infrared weak and small target image background with the size of 256 multiplied by 256.
And 3, carrying out mirror image expansion on the background edge of the simulated infrared small dim target image.
Step 1, initializing a mirror image expansion image matrix to be (M +2 t) x (N +2 t) to obtain an expansion image matrix, wherein t represents the width of the background expansion of the simulation infrared small and weak target image, M represents the number of rows of the simulation infrared small and weak target image background, N represents the number of columns of the simulation infrared small and weak target image background, and 0 to t & ltmin (M/4, N/4) and min (·) represents minimum operation.
In the practice of the present invention, the mirror expanded image matrix is initialized to 276 x 276, resulting in an expanded image matrix of 276 x 276 in size.
Assigning the simulated infrared small target image background to the t +1 st row to t + M row of the extended image matrix, within the range from the t +1 st row to the t + N row, assigning data of the extended image matrix from 2t to t +1 th row to the 1 st row of the extended image matrix, assigning data of the extended image matrix from M +1 st to t + M +1 th row to the 2t + M row of the extended image matrix, assigning data of the extended image matrix from 2t to t +1 th row to the 1 st column of the extended image matrix, assigning data of the extended image matrix from N +1 st to t + N column to the 2t + N +1 th row of the extended image matrix, so as to obtain the expanded simulated infrared small target image background, wherein M represents the number of rows of the simulated infrared small target image background, N represents the number of columns of the simulated infrared small target image background, t represents the expanded width of the simulated infrared small target image background, 0 t/min (M/4 min), and N/4 operation is the minimum value (M/N/4.).
In the implementation of the invention, the simulated infrared weak small target image background is assigned to the 11 th row to the 266 th row of the extended image matrix, the 11 th column to the 266 th column, the 20 th row to the 11 th row of the extended image matrix are assigned to the 1 st row to the 10 th row of the extended image matrix respectively, the 257 th row to the 266 th row of the extended image matrix are assigned to the 276 th row to the 267 th row of the extended image matrix respectively, the 20 th column to the 11 th column data of the extended image matrix are assigned to the 1 st column to the 10 th column of the extended image matrix respectively, the 257 th column to the 266 th column data of the extended image matrix are assigned to the 276 th column to the 267 th column of the extended image matrix respectively, and the expanded simulated infrared weak target image background with the size of 276 x 276 is obtained.
And 4, creating an infrared small target model.
Step 1, setting the signal-to-noise ratio D of the infrared weak target, wherein D is an integer value within the range that D is more than 0 and less than or equal to 20.
In the implementation of the invention, the signal-to-noise ratio of the infrared weak target is set to be 6.
Step 2, calculating the gray value of the infrared weak and small target pixel according to the following formula:
s＝D×σ+μ
wherein s represents the gray value of the pixel of the infrared small dim target, D represents the set signal-to-noise ratio of the infrared small dim target, sigma represents the standard deviation of all pixels in a local neighborhood taking the centroid of the infrared small dim target as the center, and mu represents the mean value of all pixels in a local neighborhood taking the centroid of the infrared small dim target as the center.
In the implementation of the present invention, the size of the local neighborhood centered on the centroid of the infrared weak small target is 9 × 9, which includes 81 pixels, σ represents the standard deviation of all pixels in the 9 × 9 range centered on the centroid of the infrared weak small target, and μ represents the mean of all pixels in the 9 × 9 range centered on the centroid of the infrared weak small target.
And 3, setting the size of the infrared small target as m multiplied by m pixels, wherein m is an integer value within the range of more than 0 and less than or equal to 10.
In the implementation of the invention, the size of the infrared weak and small target is set to be 3 × 3 pixels.
And 5, setting an infrared small target track.
Step 1, setting the movement speed of the infrared weak and small target as delta x pixels/frame, wherein delta x takes values within the range of 0< delta x < 10.
In the implementation of the invention, the difference value of the horizontal coordinates of the infrared small and weak target between two adjacent frames in the infrared small and weak target image sequence sets the movement speed of the infrared small and weak target to 1 pixel/frame.
And 2, taking the pixel point at the top left corner vertex of the background of the ith frame of simulated infrared dim target image as an origin, taking the horizontal rightward direction as an x axis, taking the vertical downward direction as a y axis, and establishing a background track coordinate system of the ith frame of simulated infrared dim target image.
Step 3, judging whether the background frame number of the simulation infrared small and weak target image is 1, if so, executing the step 4 in the step 5; otherwise, executing step 5 in step 5.
Step 4, setting the abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system as x 1 ,0＜x 1 N is less than or equal to N, wherein N represents the column number of the background of the simulated infrared small target image, and a linear model of the background track of the 1 st frame of the simulated infrared small target image is established according to the following formula:
y 1 ＝ax 1 +b
wherein, y 1 The vertical coordinate value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and y is more than 0 1 M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, x 1 The abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 1 N is less than or equal to N, N represents the column number of the simulation infrared small dim target image background, a represents the slope of the linear model of the simulation infrared small dim target image background track, b represents the intercept of the linear model of the simulation infrared small dim target image background track, and the value ranges of a and b are as follows: 0< ax 1 And + b is less than or equal to M, wherein M represents the number of lines of the background of the simulated infrared dim target image, in the implementation of the invention, the abscissa value of the infrared dim target in the 1 st frame of simulated infrared dim target image background track coordinate system is set as 30, the slope of the linear model for simulating the infrared dim target image background track is 1, the intercept of the linear model for simulating the infrared dim target image background track is 10, and the linear model for simulating the infrared dim target image background track is y 1 ＝x 1 +10, the coordinates of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system are (30, 40), and step 6 is executed after the 4 th step in step 5 is finished.
And 5, establishing a linear model of the background track of the simulated infrared small dim target image except the 1 st frame according to the following formula:
wherein x is i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, wherein N represents the column number, x, of the background of the simulated infrared dim target image i-1 The abscissa value of the infrared dim target in the i-1 frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i-1 N is less than or equal to N, N represents the column number of the background of the simulated infrared small target image, deltax represents the moving speed of the infrared small target, and Deltax is 0<Δx&Value of y within 10 i The longitudinal coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, and y is more than 0 i M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, a represents the slope of the linear model of the infrared dim target track, b represents the intercept of the linear model of the infrared dim target track, and the value ranges of a and b are as follows: 0< ax i And + b is less than or equal to M, in the implementation of the invention, the slope of the linear model simulating the background track of the infrared dim small target image is 1, the intercept of the linear model simulating the background track of the infrared dim small target image is 10, and the linear models simulating the background track of the infrared dim small target image except for the 1 st frame are as follows:
the coordinate of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is (x) i ，y i ) Fig. 3 is a schematic view of a flight path of a simulated infrared small dim target according to the present invention.
And 6, synthesizing a simulated infrared small target image.
Step 1, setting the total frame number of the simulated infrared weak and small target image sequence according to the following formula:
0＜L≤min((N-x 1 )/Δx,R)
wherein L represents the total image sequence of the simulated infrared weak and small targetFrame number, min (-) denotes minimum operation, N denotes column number of background of simulated infrared dim target image, x 1 The abscissa value of the infrared dim target in the 1 st frame of the simulated infrared dim target image sequence is represented, the delta x represents the movement speed of the infrared dim target, and the delta x is 0<Δx&And (4) taking a value within the range of 10, wherein R represents the total frame number of the original infrared image background sequence.
In the implementation of the invention, the total frame number L of the simulated infrared dim target image sequence is set to be 226.
Step 2, assigning the gray value s of the infrared weak and small target pixel in the step 4b to the pixel in the m multiplied by m pixel block, wherein the m multiplied by m pixel block is (x) i ，y i ) The pixel blocks are pixel blocks of the heart of the infrared weak small target, m is an integer value within the range of more than 0 and less than or equal to 10, and a simulated infrared weak small target image, x i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, N represents the column number of the background of the simulated infrared small target image, y i The vertical coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, y is more than 0 i And < M, wherein M represents the line number of the background of the simulated infrared weak and small target image.
In the implementation of the invention, the gray value s of the infrared weak and small target pixel in the step (4 b) is assigned to the pixel in the 3 × 3 pixel block, and the 3 × 3 pixel block is (x) i ，y i ) The pixel blocks are pixel blocks of the infrared weak small target type heart. The centroid coordinates of the infrared weak small target in the 50 th frame of simulated infrared weak small target image are (79, 89).
Fig. 4 is a schematic diagram of a neighborhood of 9 × 9 size centered around (79, 89) in the 50 th frame of infrared small object image, where the gray values of 9 pixels in a size range of 3 × 3 centered around the centroid of the infrared small object are all s. As can be seen from fig. 4, the infrared small target is a mean value model.
Fig. 5 is a schematic diagram of a 50 th frame of simulated infrared small and weak target image, and it can be seen from fig. 5 that the simulated infrared small and weak target image is generated by simulating and embedding the infrared small and weak target on the basis of fig. 3.
And 7, accumulating 1 operation on the background frame number i of the simulation infrared small and weak target image.
Step 8, judging whether the background frame number of the simulated infrared weak and small target image after 1 accumulation is equal to the total frame number L +1 of the simulated infrared weak and small target image sequence, if so, executing step 9; otherwise, step 2 is executed.
And 9, outputting a simulated infrared small target image sequence.
The total frame number of the simulated infrared small target image sequence is 226, and the size of each frame of the simulated infrared small target image is 256 multiplied by 256.
The effect of the present invention will be further described with reference to simulation experiments.
1. Simulation conditions are as follows:
the running system of the simulation experiment is an Intel (R) Core (TM) i3-4130@3.20GHz and 32-bit Windows operating system, and MATLAB R2012a software is adopted as simulation software.
2. The experimental contents are as follows:
in order to verify the effectiveness of the infrared dim target image sequence simulation method based on the mean value model, an original infrared image background with the total frame number of 226 frames of an original infrared image background sequence is selected for carrying out a simulation experiment.
As shown in fig. 3, the size of the original infrared image background is 480 × 625 pixels, a simulated infrared weak small target image background of 256 × 256 pixels is intercepted from the 150 th row and 150 th column of the original infrared image background to the lower right of the original infrared image background, and the size of the infrared weak small target is set to 3 × 3 pixels; the horizontal movement speed of the infrared small and weak target is set to be 1 pixel/frame, and the signal-to-noise ratio of the infrared small and weak target is set to be 6; the motion horizontal starting point of the infrared dim target is set to be 30, and the track of the infrared dim target is a straight line model. Fig. 5 shows a simulated infrared weak and small target image.
Comparing fig. 3 and fig. 5, it can be seen that the invention effectively realizes the generation simulation of the infrared dim target image, and provides a large number of testable infrared dim target image sequences for the subsequent research of the detection and tracking algorithm of the infrared dim target.
In conclusion, the infrared weak and small target image sequence generated by the method provided by the invention has the advantages of being closer to a real infrared scene, controllable in infrared weak and small target parameters and the like. Therefore, the method has wide engineering application value.
Claims (4)
1. An infrared small and weak target image sequence simulation method based on a mean value model comprises the following steps:
(1) Acquiring an infrared image background:
(1a) Acquiring a real infrared image background sequence by using a thermal infrared imager to obtain an original infrared image background sequence;
(1b) Assigning the frame number i of the original infrared image background sequence to be 1;
(2) Simulating the background of the infrared weak and small target image:
(2a) Reading in an ith frame of original infrared image background in the original infrared image background sequence, wherein i is more than 0 and less than or equal to R, and R represents the total frame number of the original infrared image background sequence;
(2b) Setting an initial row and an initial column for cutting the background of the ith frame of original infrared image;
(2c) Cutting the original infrared image background of the ith frame to obtain a simulated infrared weak and small target image background with the size of M multiplied by N, wherein M represents the line number of the simulated infrared weak and small target image background, N represents the column number of the simulated infrared weak and small target image background, M is required to be more than or equal to 0 and less than or equal to A, N is required to be more than or equal to 0 and less than or equal to B, A represents the line number of each frame image in the original infrared image background sequence, and B represents the column number of each frame image in the original infrared image background sequence;
(3) Carrying out mirror image expansion on the background edge of the simulated infrared small and weak target image;
(4) Creating an infrared small target model:
(4a) Setting the signal-to-noise ratio D of the infrared weak target, and taking an integer value when D is more than 0 and less than or equal to 20;
(4b) Calculating the gray value of the infrared weak and small target pixel according to the following formula:
s＝D×σ+μ
wherein s represents the gray value of the pixel of the infrared weak and small target, D represents the set signal-to-noise ratio of the infrared weak and small target, sigma represents the standard deviation of all pixels in a local neighborhood taking the centroid of the infrared weak and small target as the center, and mu represents the mean value of all pixels in a local neighborhood taking the centroid of the infrared weak and small target as the center;
(4c) Setting the size of the infrared dim target as m multiplied by m pixels, wherein m is an integer value within the range of more than 0 and less than or equal to 10;
(5) Setting an infrared small target track:
(5a) Setting the moving speed of the infrared weak and small target as delta x pixels/frame, wherein delta x is valued in the range of 0< delta x < 10;
(5b) Taking a pixel point at the top left corner vertex of the background of the ith frame of simulated infrared small dim target image as an origin, taking a horizontal right direction as an x axis and a vertical downward direction as a y axis, and establishing a background track coordinate system of the ith frame of simulated infrared small dim target image;
(5c) Judging whether the background frame number of the simulation infrared dim target image is 1, if so, executing the step (5 d); otherwise, executing step (5 e);
(5d) Setting the abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system as x 1 ,0＜x 1 N is less than or equal to N, wherein N represents the column number of the background of the simulated infrared dim target image, and a linear model of the background track of the 1 st frame of the simulated infrared dim target image is established according to the following formula:
y 1 ＝ax 1 +b
wherein, y 1 The vertical coordinate value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and y is more than 0 1 M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, x 1 The abscissa value of the infrared dim target in the 1 st frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 1 N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, a represents the slope of the linear model of the background track of the simulated infrared dim target image, b represents the intercept of the linear model of the background track of the simulated infrared dim target image, and the value ranges of a and b are as follows: 0< ax 1 + b is less than or equal to M, and M represents the line number of the background of the simulated infrared weak and small target image;
(5e) According to the following formula, establishing a straight line model of the background track of the simulation infrared dim target image except the 1 st frame:
wherein x is i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, wherein N represents the column number, x, of the background of the simulated infrared dim target image i-1 The abscissa value of the infrared dim target in the i-1 frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i-1 N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, deltax represents the movement speed of the infrared dim target, and Deltax is 0<Δx&Value of y within 10 i The longitudinal coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, and y is more than 0 i M is less than or equal to M, M represents the line number of the background of the simulated infrared dim target image, a represents the slope of a linear model of the infrared dim target track, b represents the intercept of the linear model of the infrared dim target track, and the value ranges of a and b are as follows: 0< ax i +b≤M；
(6) Synthesizing a simulated infrared small target image:
(6a) Setting the total frame number of the simulation infrared dim target image sequence according to the following formula:
0＜L≤min((N-x 1 )/Δx,R)
wherein, L represents the total frame number of the simulated infrared dim target image sequence, min (·) represents the minimum operation, N represents the column number of the background of the simulated infrared dim target image, and x 1 The abscissa value of the infrared small and weak target in the 1 st frame of the simulated infrared small and weak target image sequence is represented, deltax represents the movement speed of the infrared small and weak target, and deltax is 0<Δx&Taking values within 10 ranges, wherein R represents the total frame number of the original infrared image background sequence;
(6b) Assigning the gray value s of the infrared weak and small target pixel in the step (4 b) to the pixel in the m multiplied by m pixel block, wherein the m multiplied by m pixel block is (x m pixel block) i ，y i ) Is a pixel block of infrared weak small target type center, m is 0Taking an integer value within the range of less than or equal to m and less than or equal to 10 to obtain a simulated infrared dim target image, x i The abscissa value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is represented, and x is more than 0 i N is less than or equal to N, N represents the column number of the background of the simulated infrared dim target image, y i The vertical coordinate value of the infrared dim target in the ith frame simulation infrared dim target image background track coordinate system is expressed, y is more than 0 i M represents the number of lines of the background of the simulated infrared dim target image;
(7) Performing accumulation 1 operation on the background frame number i of the simulated infrared small and weak target image;
(8) Judging whether the background frame number of the simulated infrared small dim target image is equal to the total frame number L +1 of the simulated infrared small dim target image sequence after 1 is accumulated, and if so, executing the step (9); otherwise, executing the step (2);
(9) And outputting a simulated infrared weak and small target image sequence.
2. The method for simulating an infrared weak small target image sequence based on a mean value model as claimed in claim 1, wherein the range of values of the starting row and the starting column of the i frame original infrared image background clipping in step (2 b) is that the starting row of the i frame original infrared image background clipping is greater than 0 and less than the row of the original infrared image background, and the starting column of the i frame original infrared image background clipping is greater than 0 and less than the column of the original infrared image background.
3. The mean value model-based infrared small and weak target image sequence simulation method according to claim 1, wherein the specific steps of simulating background edge mirror image expansion of the infrared small and weak target image in the step (3) are as follows:
initializing a mirror image expansion image matrix to be (M +2 t) x (N +2 t) to obtain an expansion image matrix, wherein t represents the width of expansion of the background of the simulation infrared small and weak target image, M represents the row number of the background of the simulation infrared small and weak target image, N represents the column number of the background of the simulation infrared small and weak target image, 0< -t < -min (M/4, N/4) and min (·) represents minimum operation;
assigning the simulated infrared small target image background to the t +1 th row to the t + M row of the expanded image matrix, within the range from the t +1 th row to the t + N row, assigning the 2t to t +1 th row data of the expanded image matrix to the 1 st to t th rows of the expanded image matrix, assigning the M +1 to t + M th row data of the expanded image matrix to the 2t + M to t + M +1 th rows of the expanded image matrix, assigning the 2t to t +1 th column data of the expanded image matrix to the 1 st to t th rows of the expanded image matrix, assigning the N +1 to t + N th column data of the expanded image matrix to the 2t + N to t + N +1 th columns of the expanded image matrix to obtain the expanded simulated infrared small target image background, wherein M represents the row number of the simulated infrared small target image background, N represents the column number of the simulated infrared small target image background, t represents the expanded width of the simulated infrared small target image background, 0 t/min (M/4/min), and N represents the minimum operation (M/N) of the operation.
4. The mean value model-based infrared small and weak target image sequence simulation method according to claim 1, wherein the moving speed of the infrared small and weak target in the step (5 a) is a difference value of horizontal coordinates of the infrared small and weak target between two adjacent frames in the infrared small and weak target image sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510473934.1A CN105139433B (en) | 2015-08-05 | 2015-08-05 | Infrared DIM-small Target Image sequence emulation mode based on mean value model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510473934.1A CN105139433B (en) | 2015-08-05 | 2015-08-05 | Infrared DIM-small Target Image sequence emulation mode based on mean value model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105139433A CN105139433A (en) | 2015-12-09 |
| CN105139433B true CN105139433B (en) | 2018-03-06 |
Family
ID=54724765
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510473934.1A Active CN105139433B (en) | 2015-08-05 | 2015-08-05 | Infrared DIM-small Target Image sequence emulation mode based on mean value model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105139433B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875420B (en) * | 2016-12-29 | 2019-08-23 | 北京理工雷科电子信息技术有限公司 | A kind of aerial moving-target sequence image emulation mode based on remote sensing platform |
| CN108491847A (en) * | 2018-02-02 | 2018-09-04 | 口碑(上海)信息技术有限公司 | Method for expanding image and device |
| CN109584370B (en) * | 2018-11-22 | 2023-03-21 | 北京未尔锐创科技有限公司 | Target and background infrared scene simulation method |
| CN114581345B (en) * | 2022-05-07 | 2022-07-05 | 广州骏天科技有限公司 | Image enhancement method and system based on self-adaptive linear graying |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009171362A (en) * | 2008-01-17 | 2009-07-30 | Toshiba Corp | Photographing apparatus, and photographing method |
| CN103186906A (en) * | 2011-12-28 | 2013-07-03 | 中国科学院沈阳自动化研究所 | Real-time infrared dynamic scene simulation method for multiple objects in sea and sky background |
| CN104156929A (en) * | 2014-09-05 | 2014-11-19 | 西安电子科技大学 | Infrared weak and small target background inhibiting method and device on basis of global filtering |
| CN104268835A (en) * | 2014-09-23 | 2015-01-07 | 西安电子科技大学 | Weak and small infrared target strengthening method based on image geometric separation |
-
2015
- 2015-08-05 CN CN201510473934.1A patent/CN105139433B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009171362A (en) * | 2008-01-17 | 2009-07-30 | Toshiba Corp | Photographing apparatus, and photographing method |
| CN103186906A (en) * | 2011-12-28 | 2013-07-03 | 中国科学院沈阳自动化研究所 | Real-time infrared dynamic scene simulation method for multiple objects in sea and sky background |
| CN104156929A (en) * | 2014-09-05 | 2014-11-19 | 西安电子科技大学 | Infrared weak and small target background inhibiting method and device on basis of global filtering |
| CN104268835A (en) * | 2014-09-23 | 2015-01-07 | 西安电子科技大学 | Weak and small infrared target strengthening method based on image geometric separation |
Non-Patent Citations (3)
| Title |
|---|
| Tongsheng Shen等.IR image generation of space target based on openGL.《Automatic Target Recognition and Image alysis;and Multispectral Image Acquisition》.2007, * |
| 基于多尺度各向异性扩散方程的红外弱小目标检测算法;周慧鑫等;《光子学报》;20150706;全文 * |
| 基于奇异值分解的红外弱小目标背景抑制;秦翰林等;《半导体光电》;20090630;第30卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105139433A (en) | 2015-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110348376B (en) | Pedestrian real-time detection method based on neural network | |
| CN110032949A (en) | A kind of target detection and localization method based on lightweight convolutional neural networks | |
| CN112070729B (en) | Anchor-free remote sensing image target detection method and system based on scene enhancement | |
| CN105139433B (en) | Infrared DIM-small Target Image sequence emulation mode based on mean value model | |
| CN102494675B (en) | High-speed visual capturing method of moving target features | |
| CN103440389B (en) | Based on infrared camouflage effect emulation system and the appraisal procedure of full link | |
| CN105786016A (en) | Unmanned plane and RGBD image processing method | |
| CN111640125A (en) | Mask R-CNN-based aerial photograph building detection and segmentation method and device | |
| CN111027505B (en) | Hierarchical multi-target tracking method based on significance detection | |
| CN105225230A (en) | A kind of method and device identifying foreground target object | |
| CN104794737A (en) | Depth-information-aided particle filter tracking method | |
| CN103942786B (en) | The self adaptation block objects detection method of unmanned plane visible ray and infrared image | |
| CN109871829A (en) | A kind of detection model training method and device based on deep learning | |
| Li et al. | Blinkflow: A dataset to push the limits of event-based optical flow estimation | |
| CN105139432B (en) | Infrared DIM-small Target Image emulation mode based on Gauss model | |
| CN117475416A (en) | Thermal power station pointer type instrument reading identification method, system, equipment and medium | |
| CN105023255B (en) | The Infrared DIM-small Target Image sequence emulation mode of Gauss model area array cameras shake | |
| Ren et al. | Environment influences on uncertainty of object detection for automated driving systems | |
| CN115965658A (en) | Ball motion trajectory prediction method and system, electronic device and storage medium | |
| CN105072311B (en) | The Infrared DIM-small Target Image sequence emulation mode of scan IR shake | |
| CN110490165B (en) | Dynamic gesture tracking method based on convolutional neural network | |
| CN111738099B (en) | Face automatic detection method based on video image scene understanding | |
| CN110059544B (en) | Pedestrian detection method and system based on road scene | |
| CN113656918B (en) | Four-rotor simulation test method applied to finished product overhead warehouse scene | |
| CN116051808A (en) | YOLOv 5-based lightweight part identification and positioning method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhou Huixin Inventor after: Du Juan Inventor after: Song Shangzhen Inventor after: Li Huan Inventor after: Wang Bingjian Inventor after: Zhao Dong Inventor after: Yu Yue Inventor after: Jin Yangqun Inventor after: Qian Kun Inventor after: Yao Bo Inventor after: Qin Hanlin Inventor after: Rong Shenghui Inventor after: Lai Rui Inventor before: Zhou Huixin Inventor before: Zhao Ying Inventor before: Ni Man Inventor before: Yin Kuan Inventor before: Zhao Dong Inventor before: Wang Bingjian Inventor before: Pang Yingming Inventor before: Qin Hanlin Inventor before: Rong Shenghui Inventor before: Li Xiao Inventor before: Cheng Kuanhong Inventor before: Qian Kun |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |