Method and system for measuring multiband spatial variation point spread function of simple lens for computational imaging

Abstract

The invention discloses a simple lens multiband spatial variation point spread function measurement method and system for computational imaging, and belongs to the technical field of optics. The invention provides a measurement system of a spatial variation point spread function, and provides a specific method for compensating and correcting a measurement result. By measuring a proper number of horizontal view field point diffusion images of a simple lens and performing interpolation, rotation and other operations on the horizontal view field point diffusion images, a full-view-field two-dimensional space multiband point diffusion function array is generated, the measurement efficiency and the measurement accuracy are greatly improved, and improvement of the quality of a restored image is facilitated. The method can accurately and quickly measure the point spread function of the simple lens, can be directly used for inputting deconvolution operation of the simple lens, and can also be used for evaluating the imaging quality of the lens.

Description

Technical Field

[0001] The invention belongs to the field of optical technology, and in particular relates to a method and system for measuring a multi-band spatial variation point spread function of a simple lens for computational imaging. Background Art

[0002] Computational imaging technology has shown great potential in improving the resolution of optical systems and reducing the weight of optical systems. Simple lens imaging is an important research direction of computational imaging. It does not require the elimination of optical aberrations in the optical design stage to reduce the difficulty of optical design, so that the optical lens forms a blurred image, and finally obtains a clear imaging result through the deconvolution algorithm in the image processing stage. To obtain a clear restored image, it not only depends on advanced deconvolution algorithms, but also depends on the accurate estimation of the point spread function of the optical lens. Since the point spread function of a simple imaging lens depends on the wavelength of the system and changes nonlinearly with the field of view, it will take a lot of time and energy to fully obtain the point spread function of each field of view and each band.

[0003] At present, the point spread function is obtained by placing the light-emitting LED array on the object plane, which improves the measurement efficiency to a certain extent. However, this method cannot simulate an infinitely far target and cannot compensate for the influence of the LED's own size on the measurement result, resulting in a large measurement error. Therefore, it is urgent to develop an accurate and fast point spread function measurement method. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention proposes a method and system for measuring multi-band spatially varying point spread function of a simple lens for computational imaging, which can achieve accurate and fast point spread function measurement.

[0005] To achieve the above object, according to one aspect of the present invention, a simple lens multi-band spatial variation point spread function measurement system for computational imaging is provided, comprising: a light source, a filter, an off-axis collimator, a simple lens to be measured, a microscopic imaging component, an electrically controlled turntable and an image acquisition module;

[0006] The light emitted by the light source is limited to the measurement band range through the filter to achieve measurement of multiple bands. The light after the filter is collimated by an off-axis collimator to simulate an infinite star point target. The infinite star point target is imaged by the simple lens under test, amplified by the microscopic imaging component, and the point spread function image is saved by the image acquisition module.

[0007] The electrically controlled turntable drives the simple lens and microscopic imaging components to be measured to rotate, so as to achieve measurement of different fields of view.

[0008] According to another aspect of the present invention, a method for measuring a multi-band spatially varying point spread function of a simple lens for computational imaging is provided, comprising:

[0009] After the simple lens and the microscopic imaging assembly to be tested are driven to rotate horizontally to the target measurement field of view by the electric control turntable, the point spread function image is saved by the image acquisition module and the point spread function image is corrected;

[0010] After completing the point spread function image measurement and correction of all bands and the measurement field corresponding to each band, a multi-band point spread function image is obtained;

[0011] The corresponding image height and the rotation angle of the corresponding point spread function in the two-dimensional space are obtained from the two-dimensional plane field of view position;

[0012] The multi-band point spread function image is interpolated according to the corresponding image height in the two-dimensional space, and the interpolated multi-band point spread function image is rotated according to the rotation angle of the corresponding point spread function. Then, each point spread function image is spliced ​​to obtain a point spread function image with full-field multi-band spatial variation.

[0013] In some optional embodiments, the correcting the point spread function image comprises:

[0014] The point spread function image is compensated for the first time in the frequency domain by the frequency domain distribution of the pinhole diffraction, so as to compensate for the error of the pinhole size in the point spread function measurement;

[0015] The frequency domain representation sampled by the detector of the microscopic imaging assembly in the frequency domain performs a second compensation on the point spread image after the first compensation to compensate for the error caused by the detector sampling to the point spread function measurement.

[0016] In some optional embodiments, The first compensation is performed on the point spread function image, where U w represents the point spread function image after the first compensation, represents the Fourier transform, represents the inverse Fourier transform, P w Represents the point spread function image, H w represents the frequency domain distribution of pinhole diffraction, D is the diameter of the small hole, f is the spatial frequency, and J1(x) is the first-order Bessel function.

[0017] In some optional embodiments, The point spread image after the first compensation is compensated for the second time, where V w It represents the point spread function image after the second compensation. represents the Fourier transform, represents the inverse Fourier transform, K w represents the detector sampling in the frequency domain, S is the size of the detector pixel of the microscopic imaging component, and f is the spatial frequency.

[0018] In some optional embodiments, Get the two-dimensional plane field of view position, where h m is the maximum image height of the simple lens under test, and the imaging field of view of the simple lens under test is divided into N*N blocks, i∈[1,N], j∈[1,N].

[0019] In some optional embodiments, Get the corresponding image height L in two-dimensional space rc and the corresponding point spread function rotation angle θ rc .

[0020] In some optional embodiments, based on cubic spline interpolation, the point spread functions of different fields of view are interpolated to obtain the value of each image height L rc The point spread function image G at rc .

[0021] In some optional embodiments, For the interpolated multi-band point spread function image G rc Rotate.

[0022] In some optional embodiments, The point spread function is downsampled, where the convolution kernel E is an all-one matrix, the convolution step size is t, and t is the downsampling rate. n is G' rc The resolution of the microscope, S is the pixel size of the detector of the microscope imaging component, M is the magnification of the microscope imaging component, d is the pixel size of the imaging detector, Q' rc It is the point spread function image at the position (r,c) after downsampling according to the size of the imaging detector.

[0023] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0024] (1) The measurement band range is limited by filters, and an off-axis collimator is used to simulate an infinitely distant star target. The simple lens and microscopic imaging components under test are driven to rotate by an electrically controlled turntable to achieve measurement of different fields of view. It can accurately and quickly perform multi-band, spatially varying point spread function measurements.

[0025] (2) By measuring an appropriate number of point spread images of the horizontal field of view of a simple lens and performing operations such as interpolation and rotation on the point spread images, a multi-band point spread function array in two-dimensional space with a full field of view is generated. This method can accurately and quickly measure the point spread function of a simple lens, which can be directly used as the input of the deconvolution operation of a simple lens and can also be used to evaluate the imaging quality of the lens. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a diagram of a point spread function measurement system provided by an embodiment of the present invention;

[0027] Figure 2 It is a flow chart of a method for measuring multi-band spatially varying point spread function of a simple lens for computational imaging provided by an embodiment of the present invention;

[0028] Figure 3 It is a flow chart of another method for measuring multi-band spatially varying point spread function of a simple lens for computational imaging provided by an embodiment of the present invention;

[0029] Figure 4 is a comparison diagram before and after compensation of an on-axis point spread function provided by an embodiment of the present invention;

[0030] Figure 5 It is a schematic diagram of rotation of a point spread function provided by an embodiment of the present invention;

[0031] Figure 6 This is a measurement result picture of a multi-band spatially varying point spread function provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0033] In this embodiment, a Figure 1 A multi-band, spatially varying point spread function measurement system shown includes a light source 101, a filter 102, an off-axis collimator 103, a simple lens to be tested 104, a microscopic imaging component 105, an electrically controlled turntable 106 and an image acquisition module 107.

[0034] Specifically, the filter 102 is used to limit the measured band range to achieve the measurement of the point spread function of the 0.486um, 0.546um, and 0.656um imaging bands. The off-axis collimator 103 collimates the pinhole target emitted by the light source to simulate the infinite star point target. After the infinite star point target is imaged by the simple lens 104 to be measured, it is magnified by the microscopic imaging component 105 and finally collected by the image acquisition center 107. The electric control turntable 106 drives the simple lens 104 to be measured and the microscopic imaging component 105 to rotate to achieve the measurement of different fields of view.

[0035] Further, in this embodiment, based on Figure 2 The flowchart shown in the figure completes the measurement of the full-band spatial variation point spread function of a simple lens, including the following steps:

[0036] After the simple lens and the microscopic imaging assembly to be tested are driven to rotate horizontally to the target measurement field of view by the electric control turntable, the point spread function image is saved by the image acquisition module and the point spread function image is corrected;

[0037] After completing the point spread function image measurement and correction of all bands and the measurement field corresponding to each band, a multi-band point spread function image is obtained;

[0038] The corresponding image height and the rotation angle of the corresponding point spread function in the two-dimensional space are obtained from the two-dimensional plane field of view position;

[0039] The multi-band point spread function image is interpolated according to the corresponding image height in the two-dimensional space, and the interpolated multi-band point spread function image is rotated according to the rotation angle of the corresponding point spread function. Then, each point spread function image is spliced ​​to obtain a point spread function image with full-field multi-band spatial variation.

[0040] Further, in this embodiment, based on Figure 3 As shown in the flowchart, firstly, the electric control turntable 106 drives the simple lens 104 to rotate horizontally to the starting measurement field position w1, w∈[w1 w2…w n ], n is the number of horizontal measurement fields. In the specific implementation of the present invention, the electric control turntable 106 drives the simple lens 104 to rotate horizontally to the starting measurement field position w=0, w∈[0,-0.5h m , 0.5h m , -0.707h m ,0.707h m ,-h m ,h m ],h m =3.0 mm, and the point spread function image P is saved by the image acquisition module 107 w , the point spread function is saved as Figure 4As shown in (a).

[0041] Furthermore, in this embodiment, the size of the star point cannot be infinitesimal, and the point spread function lens is the response of an infinitesimal pulse. Therefore, it is necessary to compensate for the error caused by the pinhole size in the point spread function measurement. Specifically, it is compensated in the frequency domain, and the calculation method is as follows:

[0042]

[0043] Among them, U w represents the point spread function image after the first compensation, represents the Fourier transform, represents the inverse Fourier transform, H w It represents the frequency domain distribution of the pinhole diffraction. The calculation method is as follows: D = 0.310 mm is the diameter of the pinhole, f is the spatial frequency, J1(x) is the first-order Bessel function,

[0044]

[0045] Furthermore, in this embodiment, the point spread function image is collected by the detector of the microscopic imaging component 105. The sampling process of the detector is equivalent to convolving the energy distribution with the rectangular window function, which will cause image blur. Therefore, it is necessary to compensate for the error caused by the detector sampling to the point spread function measurement. Specifically, it is compensated in the frequency domain, and the calculation method is as follows:

[0046]

[0047] Among them, V w It represents the point spread function image after the second compensation. represents the Fourier transform, represents the inverse Fourier transform, K w It represents the detector sampling in the frequency domain. The calculation method is as follows: S = 4.54um is the size of the detector pixel of the microscopic imaging component, f is the spatial frequency,

[0048] After the aperture size compensation and detector sampling compensation, the point spread function is as follows: Figure 4 As shown in (b), it can be clearly observed that the point spread function size before and after compensation is smaller than the original point spread function.

[0049] Furthermore, the electric-controlled turntable 106 is controlled to drive the simple lens 104 to rotate horizontally to the next field of view position, and the above steps are repeated to complete the acquisition of point spread function images of all fields of view.

[0050] Furthermore, in this embodiment, the filter 102 at the light source is replaced, and the above steps are repeated until the measurement of the three bands of 0.486um, 0.546um, and 0.656um is completed, and the measured multi-band point spread function image is saved, which is expressed as K = [K1…K n ].

[0051] Furthermore, in this embodiment, the maximum image height h of the simple lens 104 under test is m =3.0mm, divide the lens imaging field of view into 9*9 blocks, and calculate the two-dimensional image height position (r, c) of each measurement field of view. The specific method is as follows, where i∈[1,9], j∈[1,9]:

[0052] r=i·h m / N

[0053] c=j·h m / N

[0054] Furthermore, in this embodiment, since the simple lens 104 to be tested based on computational imaging is composed of one or two lenses, its center deviation can be easily eliminated by adjustment. Assuming that the point spread function has spatial rotation invariance, the point spread function at the same image height at different spatial positions can be obtained by rotating the point spread function in the horizontal direction. Therefore, by measuring the point spread function in the horizontal direction, the point spread function of the full field of view space variation can be obtained by rotating it. Figure 5 As shown, the point spread function of the A1 field of view can be obtained by rotating the point spread function of the horizontal field of view A1', and the point spread function of the B1 field of view can be obtained by rotating the point spread function of the horizontal field of view B1'. The same is true for the remaining non-axis points. The image height L corresponding to a certain field of view (r, c) in two-dimensional space rc and the corresponding point spread function rotation angle θ rc They can all be calculated from the field of view position coordinates (r, c), the specific method is as follows:

[0055]

[0056] Furthermore, in this embodiment, based on cubic spline interpolation, the point spread function K of different fields of view is interpolated to obtain each image height L rc The point spread function image G at rc .

[0057] Furthermore, in this embodiment, according to the rotation angle θ corresponding to each view point position (r, c), rc , for each field of view after interpolation, the point spread function G rc The specific calculation method is as follows:

[0058]

[0059] Furthermore, in this embodiment, since the microscopic imaging component 105 in the measurement system magnifies the point spread function image by M=10 times, it is equivalent to upsampling the point spread function. The detector of the microscopic imaging component 105 samples it, and the detector pixel size d adapted to the simple lens 104 under test is 2.75um. In order to match the size of the actual imaging detector pixel, the point spread function needs to be downsampled. The calculation method of downsampling the point spread function image is as follows:

[0060]

[0061] Among them, the convolution kernel E is an all-one matrix, and the convolution step is t. t is the downsampling rate, which is calculated as follows, and n is G' rc The resolution of the microscope, S is the pixel size of the detector of the microscope imaging component, Q' rc The point spread function image at the position (r, c) after downsampling according to the size of the imaging detector, the magnification of the microscope imaging component is M=10, and the pixel size of the imaging detector is d=2.75um.

[0062]

[0063] Furthermore, in this embodiment, the point spread functions at all positions are spliced ​​to obtain a point spread function image with full field of view multi-band spatial variation, as shown in FIG. Figure 6 As shown, (a) is the point spread function image of the 0.486 band, (b) is the point spread function image of the 0.546 band, (c) is the point spread function image of the 0.656 band, and (d) is the merged point spread function image.

[0064] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, and two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0065] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A simple lens multi-band spatial variation point spread function measurement system for computational imaging, characterized in that: include: Light source, filter, off-axis collimator, simple lens to be tested, microscopic imaging components, electric control turntable and image acquisition module; The light emitted by the light source is limited to the measurement band range through the filter to achieve measurement of multiple bands. The light after the filter is collimated by an off-axis collimator to simulate an infinite star point target. The infinite star point target is imaged by the simple lens under test, amplified by the microscopic imaging component, and the point spread function image is saved by the image acquisition module. The electrically controlled turntable drives the simple lens and microscopic imaging components to be tested to rotate, so as to achieve measurement of different fields of view.

2. A method for measuring a simple lens multi-band spatially varying point spread function for computational imaging based on claim 1, comprising: After the simple lens and the microscopic imaging assembly to be tested are driven to rotate horizontally to the target measurement field of view by the electric control turntable, the point spread function image is saved by the image acquisition module and the point spread function image is corrected; After completing the point spread function image measurement and correction of all bands and the measurement field corresponding to each band, a multi-band point spread function image is obtained; The corresponding image height and the rotation angle of the corresponding point spread function in the two-dimensional space are obtained from the two-dimensional plane field of view position; The multi-band point spread function image is interpolated according to the corresponding image height in the two-dimensional space, and the interpolated multi-band point spread function image is rotated according to the rotation angle of the corresponding point spread function. Then, each point spread function image is spliced ​​to obtain a point spread function image with full-field multi-band spatial variation.

3. The method according to claim 2, characterized in that The step of correcting the point spread function image comprises: The point spread function image is compensated for the first time in the frequency domain by the frequency domain distribution of the pinhole diffraction, so as to compensate for the error of the pinhole size in the point spread function measurement; The frequency domain representation sampled by the detector of the microscopic imaging assembly in the frequency domain performs a second compensation on the point spread image after the first compensation to compensate for the error caused by the detector sampling to the point spread function measurement.

4. The method according to claim 3, characterized in that: Depend on The first compensation is performed on the point spread function image, where U w represents the point spread function image after the first compensation, represents the Fourier transform, represents the inverse Fourier transform, P w Represents the point spread function image, H w represents the frequency domain distribution of pinhole diffraction, D is the diameter of the small hole, f is the spatial frequency, and J1(x) is the first-order Bessel function.

5. The method according to claim 4, characterized in that Depend on The point spread image after the first compensation is compensated for the second time, where V w It represents the point spread function image after the second compensation. represents the Fourier transform, represents the inverse Fourier transform, K w represents the detector sampling in the frequency domain, S is the size of the detector pixel of the microscopic imaging component, and f is the spatial frequency.

6. The method according to any one of claims 2 to 5, characterized in that: Depend on Get the two-dimensional plane field of view position, where h m is the maximum image height of the simple lens under test, and the imaging field of view of the simple lens under test is divided into N*N blocks, i∈[1,N], j∈[1,N].

7. The method according to claim 6, characterized in that Depend on Get the corresponding image height L in two-dimensional space rc and the corresponding point spread function rotation angle θ rc .

8. The method according to claim 7, characterized in that Based on cubic spline interpolation, the point spread functions of different fields of view are interpolated to obtain the value of each image height L. rc The point spread function image G at rc .

9. The method according to claim 8, characterized in that Depend on For the interpolated multi-band point spread function image G rc Rotate.

10. The method according to claim 9, characterized in that Depend on The point spread function is downsampled, where the convolution kernel E is an all-one matrix, the convolution step size is t, and t is the downsampling rate. n is G' rc The resolution of the microscope, S is the pixel size of the detector of the microscope imaging component, M is the magnification of the microscope imaging component, d is the pixel size of the imaging detector, Q' rc It is the point spread function image at the position (r,c) after downsampling according to the size of the imaging detector.

Images