Superlens device and method for realizing incoherent light edge detection through single photographing

Abstract

The invention discloses a superlens device and method for realizing incoherent light edge detection through single photographing, and belongs to the technical field of optical information processing and metasurface imaging. Obtaining an original image containing information of the left-hand circularly polarized light and the right-hand circularly polarized light at the same time; the optical system is configured to apply different phase modulations to the left-handed circularly polarized light and the right-handed circularly polarized light, so that the left-handed circularly polarized light and the right-handed circularly polarized light form two spatially separated point spread functions on an image plane; processing the original image, and extracting and outputting an edge detection image of the object; according to the super-lens device and method for realizing incoherent light edge detection through single-time photographing, optical edge detection can be realized through single-time photographing under the condition of incoherent light illumination, and the system structure is effectively simplified through the polarization multiplexing design of the super-lens.

Description

Technical Field

[0001] This invention relates to the fields of optical information processing and metasurface imaging technology, and in particular to a superlens device and method for detecting incoherent light edges in a single photograph. Background Technology

[0002] Edge detection is a key foundational technology in computer vision and image processing, with ever-increasing demands for real-time performance. Traditional digital image processing methods are limited by serial computation, resulting in low efficiency. Optical edge detection technology, utilizing the parallel processing capabilities of light, makes high-speed processing possible.

[0003] Currently, there are two main technical approaches to optical edge detection: First, coherent light methods based on spatial filtering (such as 4F systems) or interference. These methods are complex, have poor anti-interference capabilities, and are difficult to apply in incoherent lighting environments such as natural light. Second, incoherent light methods based on optical devices (such as polarizing elements and diffractive lenses) have paved the way for improved practicality. However, existing incoherent light edge detection schemes have inherent limitations in information acquisition efficiency: to extract phase or gradient information for edge detection from the incoherent light field, multiple exposures are usually required to obtain images under different conditions for synthesis calculations, or polarization-resolved imaging devices (such as polarization CCDs) are needed to simultaneously capture multiple polarization information. The former sacrifices temporal resolution and cannot be used in dynamic scenes; the latter significantly increases the cost, size, and complexity of the system. Therefore, how to directly acquire the edge image of an object in a single photograph under incoherent light illumination has become an urgent technical challenge, which is of great significance for promoting the development of high-speed, compact, and low-cost intelligent vision systems.

[0004] Metasurfaces, as an emerging type of planar photonic device, offer a powerful platform for constructing miniaturized and integrated optical systems by allowing for flexible design of subwavelength structural parameters and control over the optical field. In recent years, research has attempted to utilize metasurfaces for edge detection. However, existing metasurface-based solutions have yet to achieve a breakthrough in addressing the compatibility issues between single-shot imaging and incoherent light, still relying on multiple-shot or polarization-resolved imaging approaches. Summary of the Invention

[0005] The purpose of this invention is to provide a superlens device and method for achieving incoherent light edge detection in a single photograph, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for incoherent light edge detection in a single image capture, comprising the following steps: S1. Under incoherent illumination, an optical system is used to take a single photograph of the object to obtain an original image that simultaneously contains information of both left-circularly polarized light (LCP) and right-circularly polarized light (RCP); the optical system is configured to apply different phase modulations to the left-circularly polarized light and the right-circularly polarized light, so that they form two spatially separated point spread functions (PSFs) on the image plane. S2. Process the original image, extract and output the edge detection image of the object; the processing includes performing operations on the original image based on the difference result of the optical transfer function (OTF) corresponding to the two point spread functions; the difference result is equivalent to a filter function that is zero at zero frequency and proportional to the square of the spatial frequency.

[0007] Preferably, the specific steps for processing the original image in S2 include: a. Determine the first optical transfer function corresponding to left-handed circularly polarized light based on the point spread function. The second optical transfer function corresponding to right-hand circularly polarized light ; b. Calculate the difference between the first optical transfer function and the second optical transfer function to obtain the equivalent optical transfer function; c. Process the original image based on the equivalent optical transfer function to extract the edge detection image.

[0008] Preferably, the first optical transfer function in S2 Second optical transfer function Specifically: ; ; The equivalent optical transfer function is: ; Substitution and have to: ; in, For frequency, and It is a constant. This is the equivalent optical transfer function.

[0009] Preferably, the specific steps for processing the original image in S2 include: a. Separate and extract the first sub-image formed by left-handed circularly polarized light modulation and the second sub-image formed by right-handed circularly polarized light modulation from the original image; b. Spatially translate and align the first sub-image with the second sub-image; c. Perform intensity difference operation on the aligned first sub-image and the second sub-image. The resulting difference image is the edge detection image of the object.

[0010] A superlens device for incoherent light edge detection in a single photograph includes a glass substrate and a metasurface formed on the glass substrate. The metasurface is composed of a periodically arranged array of subwavelength nanopillars. The nanopillars are configured to apply a first composite phase to left-handed circularly polarized incident light and a second composite phase to right-handed circularly polarized incident light, resulting in two spatially separated point spread functions in the image plane. By differentiating the optical transfer functions corresponding to the two point spread functions, an equivalent optical transfer function that is zero at zero frequency and proportional to the square of the spatial frequency is obtained. This enables the imaging system equipped with the superlens device to directly acquire the edge detection image of an object in a single photograph under incoherent light illumination.

[0011] Preferably, the nanopillars introduce a propagation (PG) phase by controlling the geometric dimensions and a geometric (Pancharatnam-Berry, PB) phase by controlling the in-plane rotation angle. The first composite phase is the difference between the propagation phase and the geometric phase, and the second composite phase is the sum of the propagation phase and the geometric phase.

[0012] Preferably, the first composite phase and the second composite phase The following relationship must be satisfied: ; in, For propagation phase, For geometric phase, , , Let these represent the focusing phase, deflection phase, and vortex phase of the parabola, respectively, satisfying: , , , in, k 0 is the wave vector. Let be the polar coordinates of the lens plane. It is the deflection angle. f It is the focal length of the superlens.

[0013] Preferably, the nanopillars are made of hydrogenated amorphous silicon, with a period of 400 nm and a height of 800 nm; the superlens device operates at a wavelength of 633 nm.

[0014] Preferably, the method for fabricating the superlens device is as follows: a. Provide a glass substrate, and deposit a hydrogenated amorphous silicon layer on the glass substrate; b. Form a mask defining the nanopillar array pattern on a hydrogenated amorphous silicon layer; c. After etching the hydrogenated amorphous silicon layer with a mask, remove the mask.

[0015] Therefore, the superlens device and method for incoherent optical edge detection using a single image capture, as described above, have the following beneficial effects: (1) This invention encodes the calculation required for edge extraction into the optical imaging process through the principle of "dual-channel optical transfer function difference". This allows the imaging system to obtain the original information for directly generating edge detection images with only one photo and can work under incoherent light (such as natural light) illumination. This effectively avoids laser speckle noise and is closer to the actual imaging environment. This provides an effective technical solution to overcome the limitations of existing incoherent light edge detection technology, which usually requires multiple exposures or special imaging devices.

[0016] (2) The present invention designs a dedicated superlens device, which becomes a key component of the optical system through polarization multiplexing and phase modulation; the optical system does not need traditional laser light sources, mechanical scanning or complex 4F filter optical paths, greatly simplifying the system structure; at the same time, the superlens device itself has the characteristics of planarization, ultrathinness and integrability, and its fabrication process is compatible with mainstream semiconductor processes, laying the hardware foundation for developing low-cost, low-power and compact intelligent edge detection systems.

[0017] (3) This invention provides a feasible solution for developing a low-cost, low-power, compact intelligent edge detection system, which has broad application potential in biomedical imaging, machine vision and portable devices.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for non-coherent light edge detection using a single photograph according to the present invention. Figure 2 This invention provides a theoretical simulation of the spatial distribution and local spot pattern of the PSF in an embodiment of the invention. Figure 3The above are simulated OTF diagrams of the superlens under different circularly polarized output beams according to an embodiment of the present invention, where a is the simulated OTF sub-diagram of the superlens under LCP output beam and b is the simulated OTF sub-diagram of the superlens under RCP output beam. Figure 4 The figure shows the simulation results of the equivalent optical transfer function according to an embodiment of the present invention, where a is a two-dimensional simulation figure of the equivalent optical transfer function obtained by subtraction, and b is... hour A simulation diagram of the one-dimensional equivalent optical transfer function along the axis; Figure 5 A schematic diagram of a superlens device structure for achieving incoherent optical edge detection in a single photograph according to the present invention; Figure 6 The following are phase diagrams designed for LCP and RCP respectively according to embodiments of the present invention; Figure 7 This is a schematic diagram of the electron microscope structure of the superlens according to an embodiment of the present invention, where a is a top view and b is a side view; Figure 8 This is a schematic diagram of an integrated imaging system device according to an embodiment of the present invention; Figure 9 The images shown are experimental results from an embodiment of the present invention, where a is an image of right-handed light emission from the experimental object, b is an image of left-handed light emission from the experimental object, and c is an edge detection image of the experimental results. Detailed Implementation

[0020] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0021] Example In optical edge detection, incoherent optical systems process information through the linear transfer of light intensity. Therefore, in the frequency domain, the output intensity spectrum ( ) and input intensity spectrum ( The relationship can be represented as: ; in, , For frequency, It is the optical transfer function (OTF), which can be obtained by calculating the autocorrelation of the amplitude transfer function (ATF): ; Therefore, in order to achieve edge detection, it is necessary to satisfy the following conditions. and Proportional, of which This is the most common case, corresponding to a typical second-order edge operator.

[0022] The formula is in Taylor series expansion is used to numerically study the optical transfer function near zero frequency, and the expression is: ; in As a constant, it can be found that Only possible with They are directly proportional. However, according to the Cauchy-Schwarz inequality, Always The maximum value is reached at zero frequency, which contradicts the requirement that the transfer function be zero at zero frequency in image differentiation operations. Therefore, it is not possible to directly obtain the value of the transfer function at zero frequency. Proportional to OTF.

[0023] To address the above problems, this invention provides a method for incoherent light edge detection using a single image capture, such as... Figure 1 As shown, it includes the following steps: S1. Under incoherent illumination, an optical system is used to take a single photograph of the object to obtain an original image that simultaneously contains information of left-handed circularly polarized light (LCP) and right-handed circularly polarized light (RCP); the optical system is configured to apply different phase modulations to the left-handed and right-handed circularly polarized light, so that the two form two spatially separated point spread functions (PSF) on the image plane. In this embodiment, the optical system achieves phase modulation through a superlens device. The nanopillar units of the superlens device control the propagation phase (PG phase) through their geometric dimensions and the geometric phase (PB phase) through their in-plane rotation angle. By jointly designing these two types of phases, two independent and different composite phases are applied to left-handed circularly polarized light (LCP) and right-handed circularly polarized light (RCP). This design causes the LCP and RCP components of an unpolarized or linearly polarized incident beam to converge at two laterally separated positions on the focal plane (image plane) after passing through the superlens, forming two point spread functions (PSFs) with different shapes. The spatial distribution of the two PSFs obtained by theoretical calculation is as follows: Figure 2 As shown, the PSF of LCP light converges on the right to form an Airy disk, while the PSF of RCP light appears as a ring on the left.

[0024] S2. Process the original image, extract and output the edge detection image of the object; the processing includes performing operations on the original image based on the difference result of the optical transfer function (OTF) corresponding to the two point spread functions; the difference result is equivalent to a filter function that is zero at zero frequency and proportional to the square of the spatial frequency.

[0025] Specifically, S2 can be achieved through the following two equivalent paths: The frequency domain calculation path directly embodies the principle of "differential optical transfer function". The specific steps are as follows: a. Perform a Fourier transform on the two point spread functions to obtain the first optical transfer function corresponding to the left-handed circularly polarized light. The second optical transfer function corresponding to right-hand circularly polarized light ,like Figure 3 (a) and Figure 3 As shown in (b), it can be observed that both OTFs have a maximum value at zero frequency, which is consistent with the general characteristics of OTFs in optical systems. b. To achieve the high-pass filtering response required for edge detection, the optical transfer function needs to be differentially processed to obtain an equivalent optical transfer function that is zero at zero frequency and proportional to the square of the spatial frequency. ; in, This is the equivalent optical transfer function. and Let be the optical transfer functions for left-handed and right-handed circularly polarized light, respectively, specifically expressed as: ; ; Two-dimensional simulation diagram as follows Figure 4 As shown in (a), the one-dimensional cross-sectional view is as follows: Figure 4 As shown in (b), theoretical analysis and Figure 4 The results are all clearly marked. The response is zero at the spatial frequency zero point, and in its vicinity it is equal to the square of the frequency. Proportional, thus having the opportunity to obtain The equivalent OTF is proportional. At this point, the lateral shift effect of the image will be reflected in the phase of the OTF. Therefore, in the mathematical expression, taking the absolute value of the OTF of LCP and RCP will translate the image in actual space to align with the center of the origin.

[0026] Substitution and The equivalent optical transfer function is obtained as follows: ; in, and As a constant, this mathematical form represents an ideal isotropic high-pass filter that can effectively extract edge information from an image.

[0027] c. Process the original image based on the equivalent optical transfer function to extract the edge detection image.

[0028] Spatial computation path: From the perspective of light field distribution, incident light with different polarization states (LCP and RCP) can generate two different images within the same field of view, and their center positions are related to... x Axisymmetric. These two images are respectively along... x The axis is translated until it aligns with the center of the origin. Finally, by directly subtracting the intensities of these two images, the edge information of the image under incoherent illumination can be obtained, that is, image processing is performed directly in the spatial domain. The specific steps are as follows: a. Separate and extract the first sub-image formed by left-hand circularly polarized light modulation and the second sub-image formed by right-hand circularly polarized light modulation from the original image; based on the spatial separation characteristics of the two PSFs known in S1, decouple and extract the two sub-images corresponding to the LCP channel and RCP channel respectively from the single original image obtained in S3 by using digital image processing algorithms (such as image cropping or region segmentation based on calibration displacement).

[0029] b. Spatially translate and align the first and second sub-images; due to the gradient phase term in the phase design, the centers of the two sub-images are laterally offset. Based on the known offset (determined by the design parameters), translate one of the sub-images pixel-wise along the offset direction so that it completely overlaps with the object image in the other sub-image, thus completing spatial registration.

[0030] c. Perform intensity difference operation on the aligned first sub-image and the second sub-image. This spatial difference operation is mathematically equivalent to filtering the object's spectrum using the equivalent high-pass optical transfer function obtained in the frequency domain. This process directly suppresses low-frequency background information in the image while highlighting high-frequency edge information. The resulting difference image is the edge detection image of the object.

[0031] like Figure 5As shown, this invention provides a superlens device for incoherent light edge detection in a single image capture. The core of this device lies in applying two independent and different phase modulations to left-handed circularly polarized light (LCP) and right-handed circularly polarized light (RCP) through subwavelength structural units. The device includes a glass substrate and a metasurface formed on the glass substrate. The metasurface is composed of a periodically arranged array of subwavelength nanopillars. In a specific embodiment, the nanopillar material is hydrogenated amorphous silicon with a period of 400 nm and a height of 800 nm. The superlens device operates at a wavelength of 633 nm, has a diameter of 1000 μm, and a focal length of 1930 μm. The nanopillars are configured to apply a first composite phase to left-handed circularly polarized incident light (LCP) and a second composite phase to right-handed circularly polarized incident light (RCP). The design principle is as follows: Figure 6 As shown; The nanopillars introduce a propagation phase by controlling the geometric dimensions and a geometric phase by controlling the in-plane rotation angle. The first composite phase is the difference between the propagation phase and the geometric phase, and the second composite phase is the sum of the propagation phase and the geometric phase.

[0032] First composite phase Second composite phase The following relationship must be satisfied: ; in, For propagation phase, For geometric phase, , , Let these represent the focusing phase, deflection phase, and vortex phase of the parabola, respectively, satisfying: , , , in, k 0 is the wave vector. Let be the polar coordinates of the lens plane. It is the deflection angle. f It is the focal length of the superlens.

[0033] This design enables LCP and RCP light to generate two spatially separated point spread functions in the image plane. By differentiating the optical transfer functions corresponding to the two point spread functions, an equivalent optical transfer function that is zero at zero frequency and proportional to the square of the spatial frequency is obtained. This allows the imaging system equipped with the superlens device to directly acquire the edge detection image of the object by taking a single picture under incoherent light illumination.

[0034] The fabrication method of the superlens device is as follows: a. Provide a clean glass substrate and deposit a hydrogenated amorphous silicon thin film with a thickness of approximately 800 nm on the substrate using plasma-enhanced chemical vapor deposition.

[0035] b. Electron beam photoresist is spin-coated onto a hydrogenated amorphous silicon thin film, and the designed nanopillar array pattern is directly written into the photoresist using an electron beam exposure system. Development is then performed to form a photoresist mask with the nanopillar pattern voids. Next, a metal layer (such as chromium) is deposited on this structure by electron beam evaporation, and the photoresist and the metal above it are removed using a lift-off process, thereby forming a metal hard mask corresponding to the designed pattern on the surface of the hydrogenated amorphous silicon thin film.

[0036] c. Using the metal hard mask formed in step b as an etching barrier layer, reactive ion etching (RIE) with fluorine-based gas as the etchant is employed to precisely transfer the nanopillar pattern to the underlying hydrogenated amorphous silicon layer. After etching, the remaining metal hard mask is completely removed using a suitable metal etchant, ultimately yielding a superlens device composed of a hydrogenated amorphous silicon nanopillar array on a glass substrate. The structure of the fabricated device can be characterized using scanning electron microscopy, and its typical top and side views are shown below. Figure 7 (a) and Figure 7 As shown in (b), the regular arrangement of nanopillars is demonstrated.

[0037] To verify this method, a system was constructed as follows: Figure 8 The integrated imaging system shown integrates a superlens device with a CMOS image sensor in a single package, and incorporates a bandpass filter with a center wavelength of 633nm in the optical path. Under incoherent illumination, a single photograph of the object yields a raw image containing information from both LCP and RCP channels. The image processing steps include: firstly, separating and extracting the first sub-image formed by LCP modulation from the raw image (with a similar effect). Figure 9 (a) and the second sub-image formed by RCP modulation (similar effect) Figure 9 (b) Next, the two sub-images are spatially translated and aligned to cancel out the lateral displacement introduced by the design; finally, intensity difference is performed on the aligned two sub-images. This operation is mathematically equivalent to filtering using the aforementioned equivalent OTF. The final edge detection image is as follows: Figure 9 As shown in (c), the outline of the object (character) is clearly highlighted and the background is effectively suppressed, which fully demonstrates the effectiveness and feasibility of edge detection by a single photograph under incoherent light conditions in this embodiment.

[0038] Therefore, the present invention employs the above-mentioned superlens device and method for realizing incoherent light edge detection in a single photograph. Under incoherent light illumination conditions, optical edge detection can be achieved with only a single photograph. Through the polarization multiplexing design of the superlens, the system structure is effectively simplified, and the cost and integration difficulty are reduced, providing a new solution for the development of compact and efficient intelligent vision systems.

[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for incoherent light edge detection in a single image capture, characterized in that, Includes the following steps: S1. Under incoherent light illumination, an optical system is used to take a single photograph of an object to obtain an original image that simultaneously contains information about left-handed and right-handed circularly polarized light; the optical system is configured to apply different phase modulations to the left-handed and right-handed circularly polarized light, so that the two form two spatially separated point spread functions on the image plane. S2. Process the original image, extract and output the edge detection image of the object; the processing includes performing operations on the original image based on the difference result of the optical transfer function corresponding to the two point spread functions; the difference result is equivalent to a filter function that is zero at zero frequency and proportional to the square of the spatial frequency.

2. The method for incoherent optical edge detection in a single photograph according to claim 1, characterized in that, The specific steps for processing the original image in S2 include: a. Determine the first optical transfer function corresponding to left-handed circularly polarized light based on the point spread function. The second optical transfer function corresponding to right-hand circularly polarized light ; b. Calculate the difference between the first optical transfer function and the second optical transfer function to obtain the equivalent optical transfer function; c. Process the original image based on the equivalent optical transfer function to extract the edge detection image.

3. The method for incoherent optical edge detection in a single photograph according to claim 2, characterized in that, First optical transfer function in S2 Second optical transfer function Specifically: ； ； The equivalent optical transfer function is: ； Substitution and have to: ； in, For frequency, and It is a constant. This is the equivalent optical transfer function.

4. The method for incoherent optical edge detection in a single photograph according to claim 1, characterized in that, The specific steps for processing the original image in S2 include: a. Separate and extract the first sub-image formed by left-handed circularly polarized light modulation and the second sub-image formed by right-handed circularly polarized light modulation from the original image; b. Spatially translate and align the first sub-image with the second sub-image; c. Perform intensity difference operation on the aligned first sub-image and the second sub-image. The resulting difference image is the edge detection image of the object.

5. A superlens device for incoherent optical edge detection in a single photograph, used in the method for incoherent optical edge detection in a single photograph as described in any one of claims 1-4, characterized in that: The device includes a glass substrate and a metasurface formed on the glass substrate. The metasurface is composed of a periodically arranged array of subwavelength nanopillars. The nanopillars are configured to apply a first composite phase to left-handed circularly polarized incident light and a second composite phase to right-handed circularly polarized incident light, resulting in two spatially separated point spread functions in the image plane. By differentiating the optical transfer functions corresponding to the two point spread functions, an equivalent optical transfer function that is zero at zero frequency and proportional to the square of the spatial frequency is obtained. This enables the imaging system equipped with the superlens device to directly acquire the edge detection image of an object by taking a single photograph under incoherent light illumination.

6. The superlens device for incoherent optical edge detection in a single photograph according to claim 5, characterized in that: The nanopillars introduce a propagation phase by controlling their geometric dimensions and a geometric phase by controlling their inward rotation angle. The first composite phase is the difference between the propagation phase and the geometric phase, and the second composite phase is the sum of the propagation phase and the geometric phase.

7. A superlens device for incoherent optical edge detection in a single photograph according to claim 6, characterized in that, First composite phase and the second composite phase The following relationship must be satisfied: ； in, For propagation phase, For geometric phase, , , Let these represent the focusing phase, deflection phase, and vortex phase of the parabola, respectively, satisfying: ， ， ， in, k 0 is the wave vector. Let be the polar coordinates of the lens plane. It is the deflection angle. f It is the focal length of the superlens.

8. The superlens device for incoherent optical edge detection in a single photograph according to claim 5, characterized in that: The nanopillars are made of hydrogenated amorphous silicon, with a period of 400 nm and a height of 800 nm; the superlens device operates at a wavelength of 633 nm.

9. A superlens device for incoherent optical edge detection in a single photograph according to claim 5, characterized in that, The method for fabricating the superlens device is as follows: a. Provide a glass substrate, and deposit a hydrogenated amorphous silicon layer on the glass substrate; b. Form a mask defining the nanopillar array pattern on a hydrogenated amorphous silicon layer; c. After etching the hydrogenated amorphous silicon layer with a mask, remove the mask.

Images