Phase recovery method and system fusing intensity transmission equation and angular spectrum iteration

Abstract

The invention provides a phase recovery method and a phase recovery system fusing an intensity transmission equation and angular spectrum iteration, which are applied to a data acquisition stage of visibility detection of a monitoring camera and are used for acquiring the visibility of the monitoring camera by utilizing an image acquired at a front end and using a novel iteration phase recovery algorithm fusing the intensity transmission equation. The algorithm comprises zero filling and gradient descent angular spectrum iteration for phase recovery, and phase information can be provided for a visibility detection algorithm; a monitoring camera expansion piece is designed at the front end, and zooming operation is achieved through an electric control zoom lens. According to the phase recovery method, the problems of iteration uncertainty, low convergence speed, diffraction distortion, easiness in falling into local minimum and the like of a traditional algorithm are solved; meanwhile, the front end has the advantages of being convenient and fast to install, low in cost and convenient to upgrade and maintain on the basis of not influencing the normal work of the original monitoring camera.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring camera visibility, and in particular to a phase recovery method and system integrating intensity transmission equation and angular spectrum iteration. Background Art

[0002] The main reason for serious traffic accidents caused by bad weather on highways is the decrease in visibility. The driving sight distance in low visibility environment is shortened, and the identification effect of variable information boards, signs and markings and other traffic safety facilities is poor. Especially in dense fog and fog belts, it is very easy to cause chain rear-end collisions and other traffic accidents. Most of the information of objects, including depth, surface contour and refractive index, is contained in the phase. However, since the vibration frequency of light waves is as high as 3.1×10 15 rad / s, existing CCD sensors and human eyes cannot directly capture the phase information of an object, but can only obtain the intensity information of the object. Therefore, it is of great significance to directly recover the phase information from the intensity information. The current quantitative phase recovery methods are mainly divided into two types: interferometry and non-interferometry.

[0003] The interferometry method uses light interference to achieve quantitative phase reconstruction. It uses reference light and modulated object light to interfere to obtain interference fringes, and then demodulates the phase of the object through the corresponding post-processing algorithm.

[0004] The non-interference method only uses the nonlinear relationship between the easily collected intensity information and the phase to be determined to invert the phase from the intensity, which is a classic inverse problem in mathematics. Non-interference phase recovery algorithms can be divided into two categories: one is a phase recovery algorithm based on the intensity transmission equation, and the other is an iterative phase recovery algorithm. The intensity transmission equation describes the quantitative relationship between the axial intensity change and the phase distribution perpendicular to the optical axis, which has the characteristics of non-interference, non-iteration, and easy implementation. The iterative phase recovery algorithm has an angular spectrum iterative phase recovery algorithm, which is an algorithm that uses the diffraction angle spectrum propagation theory to construct the forward and reverse diffraction process between the focus plane and the defocus plane light field. Invention patent CN 118348668 A discloses a phase recovery method and system based on phase perturbation and gradient compensation. The invention is applied to the field of microscopic imaging, using a motor structure and a beam splitter to realize the acquisition of images with different focal lengths, which may cause slow acquisition speed, and it only uses a randomly guessed phase as the initial phase of the algorithm, which may cause slow algorithm iteration speed and iterative errors and misunderstandings.

[0005] In summary, most of the existing visibility detection methods use neural networks to analyze and extract features from the images of vehicle-mounted surveillance cameras, and only use intensity image information, and there is still a problem of low visibility detection accuracy. Summary of the invention

[0006] The object of the present invention is to provide a phase recovery method integrating intensity transmission equation and angular spectrum iteration to solve the above-mentioned deficiencies of the prior art.

[0007] To achieve the above object, the present invention adopts the following technical solution:

[0008] A phase retrieval method integrating intensity transfer equation and angular spectrum iteration comprises the following steps:

[0009] S1, acquiring a focused intensity image and a defocused intensity image of a sample, and solving a phase of the sample by an intensity transfer equation based on the focused intensity image and the defocused intensity image;

[0010] S2, setting the total number of iterations, performing zero-filling processing on the acquired focus intensity image, defocus intensity image and phase, and combining the phase after zero-filling processing as the initial estimated phase with the amplitude of the focus intensity image after zero-filling processing to form the complex amplitude of the initial estimated focus plane of the sample;

[0011] S3, determining the complex amplitude of the current focus plane, using the amplitude of the focus intensity image after zero filling processing to constrain the complex amplitude of the current focus plane, and obtaining an updated complex amplitude of the focus plane;

[0012] S4, forward angular spectrum propagation is performed on the updated complex amplitude of the focused surface to the defocused surface to obtain the complex amplitude of the defocused surface, random perturbation is added to the phase of the complex amplitude of the defocused surface, and the amplitude of the defocused intensity image after zero filling is used to constrain the complex amplitude of the defocused surface with the random phase perturbation, so as to obtain the updated complex amplitude of the defocused surface;

[0013] S5, performing reverse angular spectrum propagation on the updated complex amplitude of the defocused plane to propagate to the focusing plane to obtain the complex amplitude of the focusing plane again, performing gradient compensation on the phase of the complex amplitude of the focusing plane obtained again, and obtaining the complex amplitude of the focusing plane of the next iteration;

[0014] S6. Iteratively repeat steps S3-S5. When the number of iterations is less than the set total number of iterations, enter the next iteration; when the number of iterations reaches the set total number of iterations, end the iteration and the phase recovery is completed.

[0015] Further, solving the phase by using the intensity transfer equation based on the focused intensity image and the defocused intensity image specifically includes the following steps:

[0016] S11, substituting the light field intensity at the focused intensity image and the defocused intensity image into the intensity transfer equation:

[0017]

[0018] In formula (1), I(x, y, z0) represents the light field intensity distribution of the focused intensity image at the z0 plane; I(x, y, z0+Δz) represents the light field intensity distribution of the defocused intensity image at the z0+Δz plane; ψ(x, y, z0) represents the sample phase; Δz represents the defocus distance; represents the gradient operator; k represents the wave number of light;

[0019] S12. Use the fast Fourier transform method to solve the phase.

[0020] Furthermore, the phase after zero filling is used as the initial estimated phase and combined with the amplitude of the focus intensity image after zero filling to form the complex amplitude of the initial estimated focus plane of the sample, which is achieved by the following formula (2):

[0021] U1=|U|exp(jψ0) (2)

[0022] In formula (2), U1 represents the complex amplitude of the initial estimated focus plane; |U| represents the amplitude of the focus intensity image obtained after zero filling; ψ0 represents the initial phase obtained after zero filling; exp represents the exponential function with e as the base, and j represents the imaginary unit i.

[0023] Furthermore, the determination of the complex amplitude of the current focus plane is achieved by the following formula (3):

[0024] U k =|U k |exp(jψ k ) (3)

[0025] In formula (3), k represents the current number of iterations, U k represents the complex amplitude of the focus plane returned in the kth iteration, |U k | represents the amplitude of the complex amplitude of the focus plane returned in the kth iteration, ψ k Represents the phase of the complex amplitude of the focus plane returned at the kth iteration.

[0026] Furthermore, the use of the zero-filled focus intensity image amplitude to constrain the complex amplitude of the current focus plane is specifically achieved by the following steps:

[0027] Based on formula (3), the amplitude |U| of the focus intensity image obtained after zero filling is used to replace |U k |, and the new complex amplitude is obtained by the following formula (4);

[0028] U k1 =|U|exp(jψ k ) (4)

[0029] In formula (4): Uk1 Represents the complex amplitude of the updated focus plane.

[0030] Furthermore, the updated complex amplitude of the focus plane is forwardly propagated in the angular spectrum, and the complex amplitude of the defocus plane is obtained by propagating to the defocus plane through the following formula (5):

[0031] U′ k =|U′ k |exp(jφ k ) (5)

[0032] In formula (5): U′ k is the complex amplitude of the defocused surface returned in the kth iteration, |U′ k | is the amplitude of the complex amplitude of the defocused surface returned in the kth iteration, φ k The phase of the complex amplitude of the defocused surface returned for the kth iteration.

[0033] Furthermore, the adding of random perturbations to the phase of the complex amplitude of the defocused surface, and using the amplitude of the defocus intensity image after zero filling to constrain the complex amplitude of the defocused surface with the random phase perturbation added, to obtain the updated complex amplitude of the defocused surface, specifically comprises the following steps:

[0034] S41, adding a random perturbation to the phase of the complex amplitude of the defocused surface in the kth iteration of the sample;

[0035] This step is achieved by the following formulas (6) and (7):

[0036]

[0037] θ k =φ k +o k (7)

[0038] In formulas (6) and (7), k represents the current number of iterations, K represents the set total number of iterations; β represents the calibration factor, which determines the initial size of the disturbance; k represents random disturbance, φ k represents the phase of the complex amplitude of the defocused surface returned by the kth iteration, θ k represents the phase after adding random perturbations, (M, N) represents the image size;

[0039] S42, using the defocus amplitude constraint, using the amplitude of the defocus intensity image |U′| after zero filling to replace the amplitude |U′ of the complex amplitude of the defocus surface returned in the kth iteration k |, and with phase θ k Combine to generate complex amplitude;

[0040] This step is achieved by the following formula (8):

[0041] U′ k2 =|U′|exp(jθ k ) (8)

[0042] In formula (8): U′ k2 is the complex amplitude of the updated defocused surface.

[0043] Further, the updated complex amplitude of the defocused plane is subjected to reverse angular spectrum propagation to the focused plane to obtain the complex amplitude of the focused plane again, and the phase of the complex amplitude of the focused plane obtained again is subjected to gradient compensation to obtain the complex amplitude of the focused plane of the next iteration, which specifically includes the following steps:

[0044] S51, the complex amplitude U' k2 Perform reverse angular spectrum propagation to the focusing plane to obtain the complex amplitude of the focusing plane again;

[0045] This step is achieved by the following formula (9):

[0046]

[0047] In formula (9): U k+1 represents the complex amplitude of the defocused surface after propagating through the reverse angular spectrum to the complex amplitude of the focused surface; |U k+1 | represents the amplitude of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface;

[0048] S52, phase of complex amplitude on the focal plane Perform gradient compensation, in phase Based on this, add an iterative additional value α k h k , and we get the phase ψ k+1 , and the phase ψ k+1 as the phase for the next iteration;

[0049] This step is achieved by the following formulas (10) to (13):

[0050]

[0051]

[0052] In formulas (10) to (13): h k Represents the gradient direction, which is proportional to the phase difference between the current iteration and the most recent iteration; α k represents the acceleration factor; t k represents the phase difference of the complex amplitude of the focal plane in the kth iteration, ψ krepresents the phase of the complex amplitude of the focus plane returned at the kth iteration, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface;

[0053] S53, the phase of the complex amplitude of the focal plane The phase ψ after gradient compensation k+1 Substitution is performed to obtain the complex amplitude of the focus plane of the next iteration;

[0054] This step is achieved by the following formula (14):

[0055] U k+1 =|U k+1 |exp(jψ k+1 ) (14)

[0056] Among them, U k+1 represents the complex amplitude of the focus plane for the next iteration; |U k+1 | represents the amplitude of the complex amplitude of the focal plane for the next iteration, ψ k+1 is the phase of the complex amplitude of the focus plane for the next iteration.

[0057] A phase recovery system integrating intensity transfer equation and angular spectrum iteration, the system is constructed based on the phase recovery method integrating intensity transfer equation and angular spectrum iteration, and is used to improve the accuracy of visibility detection of surveillance cameras. The system comprises a surveillance camera for collecting focused intensity images and defocused intensity images of samples, a zoom imaging unit and a central control unit; the zoom imaging unit comprises an electric-controlled zoom module and a magnetic base; the central control unit comprises a processor, and the central control unit can process the sample images collected by the surveillance camera and realize phase recovery based on the phase recovery method integrating intensity transfer equation and angular spectrum iteration; the electric-controlled zoom module is connected to the magnetic base by magnetic attraction, and is controlled and communicated through metal contacts; the magnetic base is fixed on the lens of the surveillance camera through a suction cup, and the magnetic base is connected to the central control unit through a data cable; the processor can send a zoom control signal and transmit it to the magnetic base through the data cable, and then control the focal length of the lens in the electric-controlled zoom module through the metal contacts to realize zoom.

[0058] It can be seen from the above technical solutions that compared with the prior art, the present invention has the following advantages:

[0059] 1. In order to solve the problem of slow convergence speed and inaccurate results of the traditional angular spectrum iteration algorithm, the method described in the present invention uses the intensity transfer equation to preliminarily solve the phase before the angular spectrum iteration, and then uses the phase to replace the random phase of the traditional angular spectrum iteration as the initial guess. Compared with the traditional angular spectrum iteration algorithm that uses a random phase to guess the initial phase of the sample, the phase solved by the intensity transfer equation in the method described in the present invention contains the prior knowledge of the sample phase, has more effective information, can accelerate the convergence speed of the angular spectrum iteration algorithm, and improve the accuracy of the angular spectrum iteration algorithm in recovering the sample phase results;

[0060] 2. In order to solve the problem that when the size of the incident light field and the diffracted light field is comparable to the calculation window, insufficient sampling will cause diffraction distortion and affect the accuracy of phase recovery in the traditional angular spectrum iterative phase recovery method, the method of the present invention avoids the above problem by adding spectral components through zero padding. The target image is filled and expanded along the x-axis and y-axis with zero pixels, and the size of the window is changed to improve the accuracy of the calculation;

[0061] 3. In order to solve the problem that the traditional angular spectrum iterative phase recovery method is prone to fall into the local minimum, resulting in the algorithm being unable to converge to the correct result, the method described in the present invention adds phase perturbation during the angular spectrum iteration process. The phase perturbation can break the stable state of the local minimum, allowing the algorithm to converge toward the global minimum and obtain accurate phase recovery results. The traditional angular spectrum iterative algorithm has many iterations and a slow phase recovery speed. The method described in the present invention performs phase compensation during the iteration process, which increases the step size of the algorithm moving toward the converged result, reduces the number of iterations, and improves the speed and efficiency of phase recovery. The present invention uses a phase recovery method of an angular spectrum iterative algorithm based on phase perturbation and gradient compensation to perform phase recovery, which is implemented using focused and defocused intensity images. The angular spectrum iterative algorithm based on phase perturbation and gradient compensation can avoid the algorithm from falling into the local minimum problem by adding phase perturbation during the iteration process. By using gradient compensation, the compensation can be expanded toward the converged result, accelerating the convergence speed of the algorithm, thereby achieving accurate and rapid phase recovery;

[0062] 4. The magnetic base of the zoom imaging unit proposed in the present invention is connected to the lens of the surveillance camera by using a suction cup, without changing the original hardware structure of the surveillance camera; at the same time, the electric-controlled zoom module and the magnetic base are connected by magnetic attraction, which is convenient and quick to install. The zoom imaging unit does not affect the acquisition performance of the original surveillance camera. Only when the algorithm needs to collect defocused images, the zoom imaging unit works, and the purpose of obtaining defocused images is achieved by changing the focal length of the electric-controlled zoom lens. The electric-controlled zoom unit and the magnetic base are controlled and communicated through metal contacts, and the magnetic base and the main control unit communicate through a USB data cable. The main control unit is connected to the network, and the staff can remotely debug and maintain the equipment through the background, and the entire system is easy to maintain and upgrade. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a flow chart of a phase recovery method integrating intensity transmission equation and angular spectrum iteration according to the present invention;

[0064] Figure 2 It is a schematic diagram of a surveillance camera extension for acquiring a defocused image according to the present invention;

[0065] Figure 3 is a result diagram of a phase recovery experiment conducted using the phase recovery system and method of the present invention;

[0066] In the picture: 1. Electric zoom module; 2. Magnetic base; 3. Metal contacts; 4. Data cable; 5. Surveillance camera; 6. Processor. DETAILED DESCRIPTION

[0067] A preferred embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0068] like Figure 1 The phase retrieval method of the fusion intensity transfer equation and angular spectrum iteration shown in the figure comprises the following steps:

[0069] S1. Acquire a focused intensity image and a defocused intensity image of a sample, and solve the phase of the sample through an intensity transfer equation based on the focused intensity image and the defocused intensity image.

[0070] Specifically, the focus intensity image and the defocus intensity image are obtained by shooting with a monitoring camera such as a vehicle-mounted camera, and solving the phase by using the intensity transfer equation based on the focus intensity image and the defocus intensity image specifically includes the following steps:

[0071] S11, substituting the light field intensity at the focused intensity image and the defocused intensity image into the intensity transfer equation:

[0072]

[0073] In formula (1), I(x, y, z0) represents the light field intensity distribution of the focused intensity image at the z0 plane; I(x, y, z0+Δz) represents the light field intensity distribution of the defocused intensity image at the z0+Δz plane; ψ(x, y, z0) represents the sample phase; Δz represents the defocus distance; represents the gradient operator; k represents the wave number of light;

[0074] S12. Use the fast Fourier transform method to solve the phase.

[0075] S2. Set the total number of iterations, and perform zero-filling processing on the acquired focus intensity image, defocus intensity image and phase at the same time, and use the phase after zero-filling processing as the initial estimated phase and combine it with the amplitude of the focus intensity image after zero-filling processing to form the complex amplitude of the initial estimated focal plane of the sample.

[0076] In this step, the phase after zero filling is used as the initial estimated phase and combined with the amplitude of the focus intensity image after zero filling to form the complex amplitude of the initial estimated focus plane of the sample, which is achieved by the following formula (2):

[0077] U1=|U|exp(jψ0) (2)

[0078] In formula (2), U1 represents the complex amplitude of the initial estimated focus plane; |U| represents the amplitude of the focus intensity image obtained after zero filling; ψ0 represents the initial phase obtained after zero filling; exp represents the exponential function with e as the base, and j represents the imaginary unit i.

[0079] S3, determining the complex amplitude of the current focus plane, using the amplitude of the focus intensity image after zero filling processing to constrain the complex amplitude of the current focus plane, and obtaining an updated complex amplitude of the focus plane;

[0080] Specifically, the complex amplitude of the current focus plane is determined by the following formula (3):

[0081] U k =|U k |exp(jψ k ) (3)

[0082] In formula (3), k represents the current number of iterations, U k represents the complex amplitude of the focus plane returned in the kth iteration, |U k | represents the amplitude of the complex amplitude of the focus plane returned in the kth iteration, ψ k Represents the phase of the complex amplitude of the focus plane returned at the kth iteration.

[0083] Furthermore, the amplitude constraint of the complex amplitude of the current focus plane using the focus intensity image amplitude after zero filling is specifically implemented by the following steps:

[0084] Based on formula (3), the amplitude |U| of the focus intensity image obtained after zero filling is used to replace |U k |, and the new complex amplitude is obtained by the following formula (4);

[0085] U k1 =|U|exp(jψ k ) (4)

[0086] In formula (4): U k1 Represents the complex amplitude of the updated focus plane.

[0087] S4. Perform forward angular spectrum propagation on the updated complex amplitude of the focused surface to the defocused surface to obtain the complex amplitude of the defocused surface. Meanwhile, add random perturbations to the phase of the complex amplitude of the defocused surface, and use the amplitude of the defocus intensity image after zero filling to perform amplitude constraints on the complex amplitude of the defocused surface with the random phase perturbation added, so as to obtain the updated complex amplitude of the defocused surface.

[0088] Specifically, the updated complex amplitude of the focused plane is forwardly propagated in the angular spectrum, and the complex amplitude of the defocused plane is obtained by propagating to the defocused plane through the following formula (5):

[0089] U′ k =|U′ k |exp(jφ k ) (5)

[0090] In formula (5): U′ k is the complex amplitude of the defocused surface returned in the kth iteration, |U′ k | is the amplitude of the complex amplitude of the defocused surface returned in the kth iteration, φ k The phase of the complex amplitude of the defocused surface returned for the kth iteration.

[0091] Furthermore, the adding of random perturbations to the phase of the complex amplitude of the defocused surface, and using the amplitude of the defocus intensity image after zero filling to constrain the complex amplitude of the defocused surface with the random phase perturbation added, to obtain the updated complex amplitude of the defocused surface, specifically comprises the following steps:

[0092] S41, adding a random perturbation to the phase of the complex amplitude of the defocused surface in the kth iteration of the sample;

[0093] This step is achieved by the following formulas (6) and (7):

[0094]

[0095] θ k =φ k +o k (7)

[0096] In formulas (6) and (7), k represents the current number of iterations, K represents the set total number of iterations; β represents the calibration factor, which determines the initial size of the disturbance and is generally set to π / 4; k represents random disturbance, φ k represents the phase of the complex amplitude of the defocused surface returned by the kth iteration, θ k represents the phase after adding random perturbations, (M, N) represents the image size;

[0097] S42, using the defocus amplitude constraint, using the amplitude of the defocus intensity image |U′| after zero filling to replace the amplitude |U′ of the complex amplitude of the defocus surface returned in the kth iteration k |, and with phase θ k Combine to generate complex amplitude;

[0098] This step is achieved by the following formula (8):

[0099] U′ k2 =|U′|exp(jθ k ) (8)

[0100] In formula (8): U′ k2 is the complex amplitude of the updated defocused surface.

[0101] S5. Perform reverse angular spectrum propagation on the updated complex amplitude of the defocused plane to propagate it to the focusing plane to regain the complex amplitude of the focusing plane, perform gradient compensation on the phase of the regained complex amplitude of the focusing plane to obtain the complex amplitude of the focusing plane for the next iteration.

[0102] Specifically, the updated complex amplitude of the defocused plane is propagated in reverse angular spectrum to the focused plane to obtain the complex amplitude of the focused plane again, and the phase of the complex amplitude of the focused plane obtained again is gradient compensated to obtain the complex amplitude of the focused plane of the next iteration, which specifically includes the following steps:

[0103] S51, the complex amplitude U' k2 Perform reverse angular spectrum propagation to the focusing plane to obtain the complex amplitude of the focusing plane again;

[0104] This step is achieved by the following formula (9):

[0105]

[0106] In formula (9): U k+1 represents the complex amplitude of the defocused surface after propagating through the reverse angular spectrum to the complex amplitude of the focused surface; |Uk+1 | represents the amplitude of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface;

[0107] S52, phase of complex amplitude on the focal plane Gradient compensation is performed by introducing a phase gradient decrease in the focusing plane, that is, Based on this, add an iterative additional value α k h k , and we get the phase ψ k+1 , and the phase ψ k+1 as the phase for the next iteration;

[0108] This step is achieved by the following formulas (10) to (13):

[0109]

[0110] In formulas (10) to (13): h k Represents the gradient direction, which is proportional to the phase difference between the current iteration and the most recent iteration; α k represents the acceleration factor; t k represents the phase difference of the complex amplitude of the focal plane in the kth iteration, ψ k represents the phase of the complex amplitude of the focus plane returned at the kth iteration, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface;

[0111] S53, the phase of the complex amplitude of the focal plane The phase ψ after gradient compensation k+1 Perform substitution to obtain the complex amplitude of the focus plane of the next iteration;

[0112] This step is achieved by the following formula (14):

[0113] U k+1 =|U k+1 |exp(jψ k+1 ) (14)

[0114] Among them, U k+1 represents the complex amplitude of the focus plane for the next iteration; |U k+1 | represents the amplitude of the complex amplitude of the focal plane for the next iteration, ψ k+1 is the phase of the complex amplitude of the focus plane for the next iteration.

[0115] S6. Iteratively repeat steps S3-S5. When the number of iterations is less than the set total number of iterations, enter the next iteration; when the number of iterations reaches the set total number of iterations, end the iteration and the phase recovery is completed.

[0116] In summary, the novel iterative phase recovery algorithm integrating the intensity transfer equation used in the present invention optimizes and improves the problems of the random initial phase value of the traditional algorithm, which leads to iterative uncertainty, insufficient sampling will cause diffraction distortion, and slow convergence speed. Compared with the disclosed invention patent CN 118348668 A, the present invention uses the intensity transfer equation to preliminarily recover the sample phase as the initial phase of the iteration before the iterative algorithm, so that the phase of the iterative algorithm contains certain prior information, which can optimize the iterative uncertainty of the traditional angular spectrum iterative algorithm and accelerate the convergence speed; by zero-filling the image before the start of the iteration to increase the spectral component, the traditional angular spectrum iterative phase recovery method is avoided. When the size of the incident light field and the diffracted light field is comparable to the calculation window, insufficient sampling will cause diffraction distortion, affecting the accuracy of phase recovery; by adding phase perturbation during the iteration process, the algorithm can be prevented from falling into the local minimum problem, and by using gradient compensation, the compensation can be expanded toward the convergence result, accelerating the convergence speed of the algorithm, thereby achieving accurate and fast phase recovery.

[0117] like Figure 2 A phase recovery system that integrates the intensity transfer equation and angular spectrum iteration is shown. The system is constructed based on the above-mentioned phase recovery method that integrates the intensity transfer equation and angular spectrum iteration, and is used to improve the accuracy of visibility detection of surveillance cameras. Compared with the disclosed invention patent CN 118348668 A, which is applied in the field of microscopic imaging and uses a motor structure and a beam splitter to realize the acquisition of images with different focal lengths, the present invention uses an electrically controlled zoom lens and a surveillance camera to complete image acquisition, which has the advantages of fast speed and convenient control. The system includes a surveillance camera for collecting focus intensity images and defocus intensity images of samples, a zoom imaging unit and a central control unit; the zoom imaging unit includes an electric-controlled zoom module 1 and a magnetic base 2; the electric-controlled zoom module 1 is connected to the magnetic base 2 by magnetic attraction, and is controlled and communicated through a metal contact 3; the magnetic base 2 is fixed on the lens of the surveillance camera through a suction cup, and the magnetic base 2 is connected to the central control unit through a data cable; the central control unit includes a processor 6, and the central control unit can process the sample images collected by the surveillance camera and realize phase recovery based on the phase recovery method of the fusion intensity transmission equation and angular spectrum iteration; the processor 6 can send a zoom control signal and transmit it to the magnetic base 2 through a data cable 4, and then control the focal length of the lens in the electric-controlled zoom module 1 through the metal contact 3 to realize zoom.

[0118] like Figure 3The result diagram of the phase recovery test performed by the phase recovery system and method of the present invention is shown. Phase recovery is performed based on the images captured by the surveillance camera. Through the phase recovery method of the present invention, the phase information is successfully recovered, thereby providing more information for visibility detection.

[0119] Specifically, the magnetic base 2 of the zoom imaging unit described in this preferred embodiment is connected to the lens of the surveillance camera by using a suction cup, without changing the original hardware structure of the surveillance camera; at the same time, the electric zoom module 1 and the magnetic base 2 are connected by magnetic attraction, which is convenient and quick to install. The zoom imaging unit does not affect the acquisition performance of the original surveillance camera. Only when the algorithm needs to collect defocused images, the zoom imaging unit works, and the purpose of obtaining defocused images is achieved by changing the focal length of the lens in the electric zoom module 1. The electric zoom unit and the magnetic base 2 are controlled and communicated through the metal contact 3, and the magnetic base 2 and the main control unit communicate through the USB data cable. The main control unit is connected to the network, and the staff can remotely debug and maintain the equipment through the background, and the entire system is easy to maintain and upgrade. Specifically, the electric zoom module 1 does not work when it does not receive a zoom command. At this time, the image captured by the surveillance camera is a focus intensity image; when the electric zoom module 1 receives a zoom command, the focal length of the lens in the electric zoom module 1 is changed, and the image captured by the surveillance camera is a defocus intensity image. The acquired focused intensity image and defocused intensity image are substituted into the phase recovery method of the fused intensity transmission equation and angular spectrum iteration described in the present invention to perform phase recovery.

[0120] The above-described embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A phase retrieval method integrating intensity transfer equation and angular spectrum iteration, characterized in that: The method comprises the following steps: S1, acquiring a focused intensity image and a defocused intensity image of a sample, and solving a phase of the sample by an intensity transfer equation based on the focused intensity image and the defocused intensity image; S2, setting the total number of iterations, performing zero-filling processing on the acquired focus intensity image, defocus intensity image and phase, and combining the phase after zero-filling processing as the initial estimated phase with the amplitude of the focus intensity image after zero-filling processing to form the complex amplitude of the initial estimated focus plane of the sample; S3, determining the complex amplitude of the current focus plane, using the amplitude of the focus intensity image after zero filling processing to constrain the complex amplitude of the current focus plane, and obtaining an updated complex amplitude of the focus plane; S4, forward angular spectrum propagation is performed on the updated complex amplitude of the focused surface to the defocused surface to obtain the complex amplitude of the defocused surface, random perturbation is added to the phase of the complex amplitude of the defocused surface, and the amplitude of the defocused intensity image after zero filling is used to constrain the complex amplitude of the defocused surface with the random phase perturbation, so as to obtain the updated complex amplitude of the defocused surface; S5, performing reverse angular spectrum propagation on the updated complex amplitude of the defocused plane to propagate to the focusing plane to obtain the complex amplitude of the focusing plane again, performing gradient compensation on the phase of the complex amplitude of the focusing plane obtained again, and obtaining the complex amplitude of the focusing plane of the next iteration; S6. Iteratively repeat steps S3-S5. When the number of iterations is less than the set total number of iterations, enter the next iteration; when the number of iterations reaches the set total number of iterations, end the iteration and the phase recovery is completed.

2. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 1 is characterized in that: Solving the phase by using the intensity transfer equation based on the focused intensity image and the defocused intensity image specifically includes the following steps: S11, substituting the light field intensity at the focused intensity image and the defocused intensity image into the intensity transfer equation: In formula (1), I(x, y, z0) represents the light field intensity distribution of the focused intensity image at the z0 plane; I(x, y, z0+Δz) represents the light field intensity distribution of the defocused intensity image at the z0+Δz plane; ψ(x, y, z0) represents the sample phase; Δz represents the defocus distance; represents the gradient operator; k represents the wave number of light; S12. Use the fast Fourier transform method to solve the phase.

3. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 2 is characterized in that: The phase after zero filling is used as the initial estimated phase and combined with the amplitude of the focus intensity image after zero filling to form the complex amplitude of the initial estimated focus plane of the sample, which is achieved by the following formula (2): U1=|U|exp(jψ0) (2) In formula (2), U1 represents the complex amplitude of the initial estimated focus plane; |U| represents the amplitude of the focus intensity image obtained after zero filling; ψ0 represents the initial phase obtained after zero filling; exp represents the exponential function with e as the base, and j represents the imaginary unit i.

4. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 3 is characterized in that: The determination of the complex amplitude of the current focus plane is achieved by the following formula (3): IN k =|U k |exp(jψ k ) (3) In formula (3), k represents the current number of iterations, U k represents the complex amplitude of the focus plane returned in the kth iteration, |U k | represents the amplitude of the complex amplitude of the focus plane returned in the kth iteration, ψ k Represents the phase of the complex amplitude of the focus plane returned at the kth iteration.

5. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 4 is characterized in that: The use of the zero-filled focus intensity image amplitude to constrain the complex amplitude of the current focus plane is specifically achieved by the following steps: Based on formula (3), the amplitude |U| of the focus intensity image obtained after zero filling is used to replace |U k |, and the new complex amplitude is obtained by the following formula (4); IN k1 =|U|exp(jψ k ) (4) In formula (4): U k1 Represents the complex amplitude of the updated focus plane.

6. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 5 is characterized in that: The forward angular spectrum propagation of the updated complex amplitude of the focus plane to the defocus plane to obtain the complex amplitude of the defocus plane is realized by the following formula (5): U′ k =|U′ k |exp(jφ k ) (5) In formula (5): U′ k is the complex amplitude of the defocused surface returned in the kth iteration, |U′ k | is the amplitude of the complex amplitude of the defocused surface returned in the kth iteration, φ k The phase of the complex amplitude of the defocused surface returned for the kth iteration.

7. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 6 is characterized in that: The method of adding random perturbation to the phase of the complex amplitude of the defocused surface, and using the amplitude of the defocus intensity image after zero filling to constrain the complex amplitude of the defocused surface to which the random phase perturbation is added, to obtain the updated complex amplitude of the defocused surface, specifically includes the following steps: S41, adding a random perturbation to the phase of the complex amplitude of the defocused surface in the kth iteration of the sample; This step is achieved by the following formulas (6) and (7): i k =φ k +o k (7) In formulas (6) and (7), k represents the current number of iterations, K represents the set total number of iterations; β represents the calibration factor, which determines the initial size of the disturbance; k represents random disturbance, φ k represents the phase of the complex amplitude of the defocused surface returned by the kth iteration, θ k represents the phase after adding random perturbations, (M, N) represents the image size; S42, using the defocus amplitude constraint, using the amplitude of the defocus intensity image |U′| after zero filling to replace the amplitude |U′ of the complex amplitude of the defocus surface returned in the kth iteration k |, and with phase θ k Combine to generate complex amplitude; This step is achieved by the following formula (8): IN' k2 =|U′|exp(jθ k ) (8) In formula (8): U′ k2 is the complex amplitude of the updated defocused surface.

8. The phase recovery method integrating the intensity transfer equation and the angular spectrum iteration according to claim 7 is characterized in that: The method of performing reverse angular spectrum propagation on the updated complex amplitude of the defocused plane to propagate to the focusing plane to obtain the complex amplitude of the focusing plane again, and performing gradient compensation on the phase of the complex amplitude of the focusing plane to obtain the complex amplitude of the focusing plane of the next iteration specifically includes the following steps: S51, the complex amplitude U' k2 Perform reverse angular spectrum propagation to the focusing plane to obtain the complex amplitude of the focusing plane again; This step is achieved by the following formula (9): In formula (9): U k+1 represents the complex amplitude of the defocused surface after propagating through the reverse angular spectrum to the complex amplitude of the focused surface; |U k+1 | represents the amplitude of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface; S52, phase of complex amplitude on the focal plane Gradient compensation is performed in phase Based on this, add an iterative additional value α k h k , and we get the phase ψ k+1 , and the phase ψ k+1 as the phase for the next iteration; This step is achieved by the following formulas (10) to (13): In formulas (10) to (13): h k Represents the gradient direction, which is proportional to the phase difference between the current iteration and the most recent iteration; α k represents the acceleration factor; t k represents the phase difference of the complex amplitude of the focal plane in the kth iteration, ψ k represents the phase of the complex amplitude of the focus plane returned at the kth iteration, It represents the phase of the complex amplitude of the defocused surface after propagation through the reverse angular spectrum to the focused surface; S53, the phase of the complex amplitude of the focal plane The phase ψ after gradient compensation k+1 Substitution is performed to obtain the complex amplitude of the focus plane of the next iteration; This step is achieved by the following formula (14): IN k+1 =|U k+1 |exp(jψ k+1 ) (14) Among them, U k+1 represents the complex amplitude of the focus plane for the next iteration; |U k+1 | represents the amplitude of the complex amplitude of the focal plane for the next iteration, ψ k+1 is the phase of the complex amplitude of the focus plane for the next iteration.

9. A phase recovery system integrating intensity transfer equation and angular spectrum iteration, which is constructed based on the phase recovery method integrating intensity transfer equation and angular spectrum iteration as described in any one of claims 1 to 8, and is used to improve the accuracy of visibility detection of surveillance cameras, and is characterized in that: The system includes a monitoring camera for collecting a focused intensity image and a defocused intensity image of a sample, a zoom imaging unit, and a central control unit; The zoom imaging unit includes an electric-controlled zoom module and a magnetic base; The central control unit includes a processor, and the central control unit can process the sample images collected by the monitoring camera and realize phase recovery based on the phase recovery method of the fusion intensity transmission equation and angular spectrum iteration; The electric zoom module is connected to the magnetic base by magnetic attraction, and is controlled and communicated through metal contacts; The magnetic base is fixed on the lens of the surveillance camera through a suction cup, and the magnetic base is connected to the central control unit through a data cable; The processor can send out a zoom control signal and transmit it to the magnetic base through a data line, and then control the focal length of the lens in the electric control zoom module through metal contacts to achieve zoom.

Images