CN114049263B - Correlated imaging denoising method based on mean value filtering - Google Patents
Correlated imaging denoising method based on mean value filtering Download PDFInfo
- Publication number
- CN114049263B CN114049263B CN202111134437.0A CN202111134437A CN114049263B CN 114049263 B CN114049263 B CN 114049263B CN 202111134437 A CN202111134437 A CN 202111134437A CN 114049263 B CN114049263 B CN 114049263B
- Authority
- CN
- China
- Prior art keywords
- hadamard
- speckle
- measured
- imaging
- moving
- 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
- 238000003384 imaging method Methods 0.000 title claims abstract description 70
- 238000001914 filtration Methods 0.000 title claims abstract description 40
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000002596 correlated effect Effects 0.000 title claims abstract description 26
- 230000003287 optical effect Effects 0.000 claims description 8
- 230000033001 locomotion Effects 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000001678 irradiating effect Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 abstract description 9
- 238000009825 accumulation Methods 0.000 abstract 1
- 230000000875 corresponding effect Effects 0.000 abstract 1
- 238000011084 recovery Methods 0.000 description 9
- 238000005314 correlation function Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000005337 ground glass Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
Abstract
The invention provides a correlated imaging denoising method based on mean filtering. According to the method, moving speckles are designed according to the characteristic of mean filtering, the speckles are irradiated onto an unknown object through a correlation imaging system, and the image of the denoised unknown object is obtained through the second-order correlation between the sum of the accumulation of the corresponding barrel detector values and the moving speckles. The method introduces the mean filtering method in image denoising into the correlated imaging, and can effectively reduce the influence of noise on the imaging quality under the condition of Gaussian white noise. Simulation results show that compared with the original associated imaging results without mean filtering, the method can effectively reduce the influence of noise on imaging, improve the imaging quality, has the advantages of simple structure and convenience in implementation, is suitable for the condition with higher requirements on the imaging quality, and has a larger application prospect.
Description
Technical Field
The invention relates to a denoising method of correlated imaging based on mean filtering, and belongs to the field of correlated imaging.
Background
Associative Imaging (Correlated Imaging), also known as "Ghost Imaging (GI), is one of the leading edge and hot spot in the quantum optics field in recent years. The correlation imaging adopts two light paths, one is called a signal light path, and the light after speckles pass through an object is received by a barrel detector without spatial resolution; the other branch is called a reference light path, and refers to that the speckles are received by a detector with spatial resolution capability after being transmitted for a certain distance. And finally, carrying out second-order correlation on the detection results of the two light paths to obtain a restored image of the object in the reference light path. Unlike conventional imaging, correlated imaging can recover an image of an object on an optical path that does not contain the object, a property known as delocalization. With the development of related imaging, novel ghost imaging technologies such as thermo-optic ghost imaging and computing ghost imaging are continuously proposed. Meanwhile, as the performance of "ghost" imaging increases, various applications based on "ghost" imaging are also being realized. A large number of researches show that the ghost imaging can be widely applied to the fields of military, encryption, laser radar and the like.
In the 80 s of the last century, scholars d.n. klyshko proposed quantum-dependent imaging schemes based on the entanglement behavior of the spontaneous parametric down-conversion photon pairs. In 1995, pittman et al have experimentally achieved quantum correlation imaging for the first time according to Klyshko's theory. In 2002, bennink et al experimentally realized "ghost" imaging using classical thermal light sources, demonstrating that "ghost" imaging can be realized using thermo-optic alike, and that a tremendous bombarding was induced in the field of "ghost" imaging. 2008. Jeffrey h.shapiro et al theorized that computational "ghost" imaging could be an alternative technique to conventional "ghost" imaging, and 2009 Bromberg et al experimentally achieved a computational "ghost" imaging, i.e., a "imaging scheme of a single detection light path, in which the rotating ground glass was replaced by a computer-controlled Spatial Light Modulator (SLM). In 2010, ferri F proposes a differential ghost imaging scheme based on thermo-optic ghost imaging, and only differential information of an object is transmitted during imaging, so that the influence of background noise on the quality of a restored image is reduced, and the quality of imaging is improved. In 2015, zhang Z B et al proposed a method for achieving single pixel imaging by acquiring the fourier spectrum of an image, which greatly improved the quality of the reconstructed image. Meanwhile, various applications based on the correlated imaging are also widely studied, such as gradient correlated imaging proposed by Liu F et al in 2015, and edge information of an object can be directly obtained under the condition that an object image is not obtained through gradient correlated imaging.
With the development of the technology of "ghost" imaging, how to effectively reduce the influence of noise on imaging becomes one of the research hotspots. The concept of mean filtering is introduced into the correlated imaging process, so that the influence of Gaussian white noise on imaging can be effectively reduced, and an object image with higher imaging quality can be obtained.
In view of the foregoing, it is necessary to propose a denoising method for correlated imaging based on mean filtering to solve the above-mentioned problems.
Disclosure of Invention
The invention aims to provide a denoising method of correlated imaging based on mean filtering, which is used for reducing the influence of Gaussian white noise on image recovery.
In order to achieve the above object, the present invention provides a denoising method for correlated imaging based on mean filtering, which mainly comprises the following steps:
step one, generating a required Hadamard speckle pattern by a computer according to the size of an object;
selecting a template with mean filtering, moving the coordinate position of the generated speckle to obtain a speckle pattern containing mean filtering information, namely a moving speckle pattern, and loading the moving speckle pattern onto a digital micromirror array;
Step three, the light source irradiates the digital micro-mirror array, and the mobile Hadamard speckle required by the associated imaging can be obtained through the modulation of the digital micro-mirror array; sequentially irradiating the moving Hadamard speckles on an object to be detected, and receiving an optical signal acquired from the object to be detected by utilizing a barrel detector through transmission or reflection to obtain the barrel detector value of each Hadamard speckle;
And step four, accumulating all the barrel detector values of the Hadamard speckle to obtain total barrel detector values, performing second-order correlation on the total barrel detector values and the unshifted Hadamard speckle pattern, and recovering the image of the detected object.
As a further improvement of the present invention, the hadamard speckle pattern in the first step is S k(xi,yj), where k represents the kth hadamard speckle and (x i,yj) represents the coordinate position.
As a further improvement of the present invention, the average filtered template is 3*3 average filtered templates, for example, 9 moving speckle patterns are generated, and the relationship of the 9 moving speckle patterns is:
Wherein the method comprises the steps of An mth movement pattern which is a kth hadamard speckle pattern.
As a further improvement of the present invention, the bucket detector values in step three are:
Wherein R (x i,yj) represents information distribution of the object to be detected.
As a further improvement of the present invention, in step four, the total bin detector values are:
Wherein S k(xi,yj) is the kth Hadamard speckle pattern, R (x i,yj) is the object to be measured, R (x i,yj+1) is the object to be measured which moves down by one bit in the y direction, R (x i,yj-1) is the object to be measured which moves up by one bit in the y direction, R (x i-1,yj) is the object to be measured which moves right by one bit in the x direction, R (x i+1,yj) is the object to be measured which moves left by one bit in the x direction, R (x i+1,yj+1) is the object to be measured which moves by one bit in the southwest direction, R (x i-1,yj-1) is the object to be measured which moves by one bit in the northeast direction, R (x i+1,yj-1) is the object to be measured which moves by one bit in the northwest direction, and R (x i-1,yj+1) is the object to be measured which moves by one bit in the southeast direction.
As a further improvement of the present invention, step four further includes: the optical signals generated after transmitting or reflecting the object to be measured are converged by the lens and then received by the barrel detector.
As a further improvement of the invention, the barrel detector is a barrel detector without spatial resolution capability.
As a further improvement of the present invention, in the fourth step, the image of the object to be measured is: Wherein B k is the total bin detector value of all Hadamard speckle moving speckles irradiated onto the object to be measured, S k(xi,yj) is the kth Hadamard speckle pattern, and </cndot > is the statistical average of the set.
The beneficial effects of the invention are as follows: the method is different from the traditional associated imaging method, and the concept of mean filtering is introduced into the associated imaging process, so that the influence of Gaussian white noise on imaging can be effectively reduced, and meanwhile, noise and imaging are removed.
Drawings
FIG. 1 is a flow chart of a denoising method of the present invention based on mean filtering correlation imaging.
Fig. 2 is a graph of simulation results of pictures "double slit" and "Ghost" in accordance with the present invention using a second order correlation function and mean filtering as a recovery algorithm.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1, the invention discloses a denoising method of correlated imaging based on mean filtering, which mainly comprises the following steps:
step one, generating a required Hadamard speckle pattern by a computer according to the size of an object;
selecting a template with mean filtering, moving the coordinate position of the generated speckle to obtain a speckle pattern containing mean filtering information, namely a moving speckle pattern, and loading the moving speckle pattern onto a digital micromirror array;
Step three, the light source irradiates the digital micro-mirror array, and the mobile Hadamard speckle required by the associated imaging can be obtained through the modulation of the digital micro-mirror array; sequentially irradiating the moving Hadamard speckles on an object to be detected, and receiving an optical signal acquired from the object to be detected by utilizing a barrel detector through transmission or reflection to obtain the barrel detector value of each Hadamard speckle;
And step four, accumulating all the barrel detector values of the Hadamard speckle to obtain total barrel detector values, performing second-order correlation on the total barrel detector values and the unshifted Hadamard speckle pattern, and recovering the image of the detected object.
Specifically, the computer generates the desired speckle without movement, and then moves the speckle to generate the hadamard speckle S k(xi,yj required for simulation, where k represents the kth hadamard speckle, (x i,yj) represents the coordinate position.
Then the object is irradiated by the generated moving speckles, optical signals are generated after the object is transmitted and/or reflected, and then the optical signals are received by a barrel detector without space resolution after being focused by a lens, so that the values of the barrel detector are as follows:
Wherein R (x i,yj) represents the distribution of the object to be measured.
The total bucket detector values are:
the object and the speckle are in relative motion, so that movement of the object can be seen as movement of the speckle.
The template of the mean filtering in the invention is 3*3 template mean filtering, then 9 moving speckles (including speckles unknown to the center of movement) are generated, and the relationship is as follows:
After all randomly moving speckle is illuminated on the object and the bucket detector receives all the intensity values, we can formulate the above process as:
wherein B k is the total bucket detector value of all the k-th Hadamard speckle shift speckles irradiated onto the object to be measured.
The second-order correlation function can be used to directly obtain a restored image of the object, and the theoretical process is as follows:
Wherein B k is the total bin detector value of all k-th Hadamard speckle shift speckles irradiated onto the object to be measured, S k(xi,yj) is the k-th Hadamard speckle pattern, and <. Is the statistical average of the set.
Through the theoretical process, the average filtering is applied to the correlated imaging, the image of the object can be obtained according to the second-order correlated function, and through theoretical deduction, the influence of noise on the imaging can be reduced to a certain extent by using the method.
Fig. 2 is a simulation result diagram of a denoising method of correlated imaging based on mean filtering. The invention verifies images of double slit and Ghost. The specific simulation process is as follows: nine sets of moving hadamard speckles are generated by a computer, then the unknown object is irradiated by the nine sets of hadamard speckles, light transmitted by the object is converged by a lens and received by a barrel detector without any spatial resolution, and finally an image of the unknown object is restored according to the sum of the hadamard speckles and the barrel detector by a second-order correlation function.
The simulation results for "double slit" and "Ghost" are shown in fig. 2, which includes simulation results using the mean filtering method and without using the mean filtering method. Meanwhile, in order to compare the recovery results when different recovery algorithms are used, peak signal-to-noise ratio (PSNR) is adopted as an evaluation standard, and the larger the peak signal-to-noise ratio and the structural similarity is, the better the effect of the recovery algorithm on image recovery is, and the closer the recovery result is to a real image. The simulation result is calculated, and for a 'double slit' image, PSNR of a restored image obtained by mean value-free filtering is 13.09dB. The PSNR of the restored image obtained using the mean value filtering is 15.8dB; for "Ghost", the PSNR of the restored image obtained by the non-average filtering was 13.03dB, and the PSNR of the restored image obtained by the average filtering was 15.44dB. As can be seen from the simulation result, compared with the result without using the mean value filtering, the recovery result of the mean value filtering algorithm is clearer, namely the recovery result has better quality.
In conclusion, after the correlated imaging denoising method based on the mean filtering is adopted, the peak signal-to-noise ratio (PSNR) of the restored image is greatly improved compared with the restored result without the mean filtering, and the restored result is clearer. Combining the above analysis shows that applying the mean filter to the correlated imaging can effectively reduce the effect of gaussian white noise on the imaging.
The above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention.
Claims (5)
1. The denoising method for correlated imaging based on mean filtering is characterized by mainly comprising the following steps of:
step one, generating a required Hadamard speckle pattern by a computer according to the size of an object;
selecting a template with mean filtering, moving the coordinate position of the generated speckle to obtain a speckle pattern containing mean filtering information, namely a moving speckle pattern, and loading the moving speckle pattern onto a digital micromirror array;
Step three, the light source irradiates the digital micro-mirror array, and the mobile Hadamard speckle required by the associated imaging can be obtained through the modulation of the digital micro-mirror array; sequentially irradiating the moving Hadamard speckles on an object to be detected, and receiving an optical signal acquired from the object to be detected by utilizing a barrel detector through transmission or reflection to obtain the barrel detector value of each Hadamard speckle;
Accumulating all the barrel detector values of the Hadamard speckle to obtain total barrel detector values, performing second-order correlation on the total barrel detector values and the unshifted Hadamard speckle pattern, and recovering the image of the detected object;
Wherein, in the first step, the hadamard speckle pattern is S k(xi,yj), k represents the kth hadamard speckle, and (x i,yj) represents the coordinate position;
the average filtered template is 3*3 average filtered templates, 9 moving speckle patterns are generated, and the relation of the 9 moving speckle patterns is as follows:
Wherein the method comprises the steps of An mth movement pattern which is a kth hadamard speckle pattern;
In the fourth step, the image of the object to be measured is:
wherein B k is the total bin detector value of all Hadamard speckle moving speckles irradiated onto the object to be measured, S k(xi,yj) is the kth Hadamard speckle pattern, and </cndot > is the statistical average of the set.
2. The denoising method of correlated imaging based on mean value filtering according to claim 1, wherein the bin detector values in step three are:
Wherein R (x i,yj) represents information distribution of the object to be detected.
3. The method of denoising of correlated imaging based on mean filtering according to claim 1, wherein in step four the total bin detector value is:
Wherein S k(xi,yj) is the kth Hadamard speckle pattern, R (x i,yj) is the object to be measured, R (x i,yj+1) is the object to be measured which moves down by one bit in the y direction, R (x i,yj-1) is the object to be measured which moves up by one bit in the y direction, R (x i-1,yj) is the object to be measured which moves right by one bit in the x direction, R (x i+1,yj) is the object to be measured which moves left by one bit in the x direction, R (x i+1,yj+1) is the object to be measured which moves by one bit in the southwest direction, R (x i-1,yj-1) is the object to be measured which moves by one bit in the northwest direction, R (x i+1,yj-1) is the object to be measured which moves by one bit in the southwest direction, and R (x i-1,yj+1) is the object to be measured which moves by one bit in the southwest direction.
4. The denoising method of correlated imaging based on mean value filtering according to claim 1, wherein step four further comprises: the optical signals generated after transmitting or reflecting the object to be measured are converged by the lens and then received by the barrel detector.
5. The denoising method of correlated imaging based on mean value filtering according to claim 4, wherein: the barrel detector is a barrel detector without spatial resolution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111134437.0A CN114049263B (en) | 2021-09-27 | 2021-09-27 | Correlated imaging denoising method based on mean value filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111134437.0A CN114049263B (en) | 2021-09-27 | 2021-09-27 | Correlated imaging denoising method based on mean value filtering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114049263A CN114049263A (en) | 2022-02-15 |
| CN114049263B true CN114049263B (en) | 2024-05-03 |
Family
ID=80204844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111134437.0A Active CN114049263B (en) | 2021-09-27 | 2021-09-27 | Correlated imaging denoising method based on mean value filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114049263B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109901190A (en) * | 2019-03-07 | 2019-06-18 | 南京邮电大学 | Relevance imaging method based on linear regression |
| CN110243398A (en) * | 2019-06-27 | 2019-09-17 | 南京邮电大学 | A kind of relevance imaging method of the phase object based on relevant detection |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI485361B (en) * | 2013-09-11 | 2015-05-21 | Univ Nat Taiwan | Measuring apparatus for three-dimensional profilometry and method thereof |
-
2021
- 2021-09-27 CN CN202111134437.0A patent/CN114049263B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109901190A (en) * | 2019-03-07 | 2019-06-18 | 南京邮电大学 | Relevance imaging method based on linear regression |
| CN110243398A (en) * | 2019-06-27 | 2019-09-17 | 南京邮电大学 | A kind of relevance imaging method of the phase object based on relevant detection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114049263A (en) | 2022-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11368608B2 (en) | Compressed sensing based object imaging system and imaging method therefor | |
| CN109242808B (en) | Associated imaging method based on fractional Fourier transform | |
| Wu et al. | Resolution enhancement of long-range imaging with sparse apertures | |
| CN109901190B (en) | Correlation imaging method based on linear regression | |
| CN110243398B (en) | Coherent detection-based phase object correlation imaging method | |
| CN111986118B (en) | Underwater calculation ghost imaging image denoising method and system with minimized weighted kernel norm | |
| Yang et al. | Tracking compensation in computational ghost imaging of moving objects | |
| Chen et al. | Computational ghost imaging with uncertain imaging distance | |
| Cheng et al. | First-photon imaging with independent depth reconstruction | |
| CN114049263B (en) | Correlated imaging denoising method based on mean value filtering | |
| Li et al. | Far‐Field Synthetic Aperture Imaging via Fourier Ptychography with Quasi‐Plane Wave Illumination | |
| Guan et al. | Source separation and noise reduction in single-pixel imaging | |
| Sun et al. | Double‐flow convolutional neural network for rapid large field of view Fourier ptychographic reconstruction | |
| Wang et al. | A deep learning-based target recognition method for entangled optical quantum imaging system | |
| Ma et al. | Direct noise-resistant edge detection with edge-sensitive single-pixel imaging modulation | |
| EP3499201B1 (en) | Electromagnetic wave phase/amplitude generation device, electromagnetic wave phase/amplitude generation method, and electromagnetic wave phase/amplitude generation program | |
| CN110716211B (en) | Iterative filtering associated imaging method | |
| Rahmat et al. | 3D shape from focus using LULU operators and discrete pulse transform in the presence of noise | |
| Zhang et al. | Fourier single-pixel imaging based on lateral inhibition for low-contrast scenes | |
| Proppe et al. | 3D-2D Neural Nets for Phase Retrieval in Noisy Interferometric Imaging | |
| Krawczyk et al. | Artifact generation when using perceptual loss for image deblurring | |
| Boehrer et al. | Turbulence mitigation in imagery including moving objects from a static event camera | |
| Hu et al. | Fourier ptychographic layer-based imaging of hazy environments | |
| Boehrer et al. | Turbulence mitigation in neuromorphic camera imagery | |
| Zhu et al. | Convenient noncooperative speckle-correlation imaging method |
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 |