**2. Principles**

Figure 1a,b schematically represent the measurements for the reflection spherical mirrors and transmission spherical lenses (bi-convex or bi-concave), respectively. To facilitate understanding, an *x–y–z* coordinate system is introduced in the figures. A collimated and expanded laser beam with a wavelength of *λ* passes through the Ronchi grating (G), located at (*<sup>x</sup>*0, *y*0); the transmittance of the grating can be written as

$$t\_0(\mathbf{x}\_0, y\_0) = \frac{1}{2}(1 + \cos \frac{2\pi \mathbf{x}\_0}{p\_1}),\tag{1}$$

where *p*1 is the period of the grating in the *x*–direction. The electric field behind the grating is denoted as *U*0 (*<sup>x</sup>*0, *y*0, 0) (at *z* = 0) and the intensity can be expressed as

$$I\_0 = \left| \mathcal{U}\_0(\mathbf{x}\_0, y\_0, 0) \right|^2 = \frac{1}{4} (1 + 2 \cos \frac{2 \pi \mathbf{x}\_0}{p\_1} + \cos^2 \frac{2 \pi \mathbf{x}\_0}{p\_1}),\tag{2}$$

**Figure 1.** Schematic of the measurement setup for (**a**) reflection and (**b**) transmission samples. SF, spatial light filter; G, Ronchi grating; TS, test sample; OS, observation screen; DC, digital camera.

According to Fresnel diffraction, the electric field *U*1 (*<sup>x</sup>*1, *y*1, *d*1) of the test light reaching the test sample (TS) after propagating a distance of *d*1 can be expressed as

$$dI\_1(\mathbf{x}\_1, \mathbf{y}\_1, d\_1) = \frac{e^{j k d\_1}}{j \lambda d\_1} \int \int\_{-\infty}^{\infty} t\_0(\mathbf{x}\_0, \mathbf{y}\_0) e^{j \frac{k}{2d\_1} \left[ \left( \mathbf{x}\_1 - \mathbf{x}\_0 \right)^2 + \left( y\_1 - y\_0 \right)^2 \right]} d\mathbf{x}\_0 d\mathbf{y}\_0. \tag{3}$$

Meanwhile, the transmittance of the test sample, *t*1 (*<sup>x</sup>*1, *y*1), can be expressed as [13]

$$t\_1(x\_1, y\_1) = \left. e^{-j\frac{k}{2f}(x\_1^2 + y\_1^2)} \right|,\tag{4}$$

where *k =* 2*πn/λ* is the propagation number, *n* is the refractive index of the environmental medium, and *f* is the focal length of the test sample. For converging lenses or concave mirrors, the focal length has a positive sign. By contrast, for diverging lenses or convex mirrors, the focal length has a negative sign.

As shown in Figure 1a,b, respectively, the reflected and transmitted electric field *U* 1 can be written as

$$\mathcal{U}\_{1}^{\prime}(\mathbf{x}\_{1},y\_{1},d\_{1}) = \frac{e^{j k d\_{1}}}{j \lambda d\_{1}} \int \int\_{-\infty}^{\infty} t\_{0}(\mathbf{x}\_{0},y\_{0}) e^{j \frac{k}{2T\_{1}} \left[ (x\_{1}-x\_{0})^{2} + (y\_{1}-y\_{0})^{2} \right]} e^{-j \frac{k}{2f} (x\_{1}^{2}+y\_{1}^{2})} d\mathbf{x}\_{0} dy\_{0}. \tag{5}$$

Again, the test light travels an additional distance of propagation *d*2 and reaches the observation screen (OS), as shown in Figure 1a,b, respectively. According to Fresnel diffraction, the electric field of the test light on the observation screen can be written as

$$\text{Ll2}(\mathbf{x}\_2, y\_2, d\_2) \;= \frac{e^{\bar{\lambda}d\_2}}{\bar{\lambda}d\_2} \int \int\_{-\infty}^{\infty} \text{Ll}'\_1(\mathbf{x}\_1, y\_1, d\_1) e^{\frac{\bar{\lambda}}{2d\_2} \left[ (\mathbf{x}\_2 - \mathbf{x}\_1)^2 + (y\_2 - y\_1)^2 \right]} d\mathbf{x}\_1 dy\_1. \tag{6}$$

With extensive substitution, Equation (6) can be expressed as [13]

$$\mathcal{U}l\_2(x\_2, y\_2, d\_2) = \frac{1}{2} \left\{ 1 + \cos \left[ 2\pi \frac{x\_2}{p\_1 \left( 1 - \frac{d\_2}{f} \right)} \right] \times \exp \left[ -j \frac{\pi \lambda d\_1}{p\_1^2} \left( \frac{\frac{1}{d\_1} - \frac{1}{f} + \frac{1}{d\_2}}{\frac{1}{d\_2} - \frac{1}{f}} \right) \right] \right\}. \tag{7}$$

Accordingly, the intensity on the observation screen, recorded using a digital camera (DC), can be expressed as follows:

$$\left|\mathbb{I}(\mathbf{x}\_{2},y\_{2},d\_{2})\right| = \left|\mathbb{I}l\_{2}(\mathbf{x}\_{2},y\_{2},d\_{2})\right|^{2} = \frac{1}{4}\left\{1 + 2\cos\frac{2\pi x\_{2}}{p\_{1}(1-\frac{d\_{2}}{f})}\cos\frac{\pi\lambda d\_{1}}{p\_{1}}\frac{\frac{1}{\overline{d\_{1}}} - \frac{1}{f} + \frac{1}{\overline{d\_{2}}}}{\frac{1}{\overline{d\_{2}}} - \frac{1}{f}}\right) + \cos^{2}\frac{2\pi x\_{2}}{p\_{1}(1-\frac{d\_{2}}{f})}\right\}.\tag{8}$$

The captured intensity with a period of *mp*1 and a grating vector *K***1** = (2 *π*/*mp*1)<sup>ˆ</sup>*<sup>i</sup>* is superimposed on a digital grating with a period of *p*2 and a grating vector *K***2** = (2 *π* cos*θ*/*p*2)<sup>ˆ</sup>*i*−(2 *π* sin*θ*/*p*2)<sup>ˆ</sup>*j*, as shown in Figure 2. The value of *p*2 is determined by the captured pattern described in Equation (8) and is set as *p*2 ∼= *mp*1. Therefore, a moiré pattern with a slant angle *α* can be obtained. The slant angle *α* is defined as the angle between the grating vector *K***3** of moiré pattern and the *y*-axis which can be written as [14]

$$\tan^{-1}\left(\frac{p\_2 - mp\_1 \cos\theta}{mp\_1 \sin\theta}\right) = \tan^{-1}\left(\frac{p\_2 - \left|1 - \frac{d\_2}{\hbar}\right| p\_1 \cos\theta}{\left|1 - \frac{d\_2}{\hbar}\right| p\_1 \sin\theta}\right) = \tan^{-1}\left(\frac{1 - M\cos\theta}{M\sin\theta}\right),\tag{9}$$

where *m* = |1 − *d*2*/f*| is the imaging magnification for grating G, *θ* is the angle between the practical and digital gratings, and *M* is the period ratio of the two period patterns (captured intensity and digital grating) which equals *mp1/p*2. Therefore, the observation plane is in the grid image given by the test sample under study. It results from the setup of Figure 1 and the magnification formula *m* = |1 − *d*2*/f*|.

**Figure 2.** Schematic illustration for the generation of moiré pattern and related orientations.

> Accordingly, the focal length of the test samples can be expressed as

$$f = \frac{p\_1(\tan a \sin \theta + \cos \theta) d\_2}{p\_1(\tan a \sin \theta + \cos \theta) - p\_2},\tag{10}$$

and the radius of curvature of the test samples can be written as

$$R = -2f = -\frac{2p\_1(\tan a \sin \theta + \cos \theta)d\_2}{p\_1(\tan a \sin \theta + \cos \theta) - p\_2} \text{ (for spherical mirrors)}\tag{11}$$

and

$$R\_1 = -R\_2 = 2(n\_l - 1)f = \frac{2(n\_l - 1)p\_1(\tan a \sin \theta + \cos \theta)d\_2}{p\_1(\tan a \sin \theta + \cos \theta) - p\_2}, \text{ (for spherical lenses)}\tag{12}$$

where *R*1 and *R*2 are the radius of curvature of the bi-convex or bi-concave lens's first and second surfaces, and *nl* is the refractive index of the lens. Thus, the focal length and radius of curvature of the test samples can be determined using the following experimental parameters: the periods of the practical and digital gratings, *p*1 and *p*2, respectively; the relative angle between these gratings, *θ*; the slant angle of the moiré pattern, *α*; the distance, *d*2; and the refractive index of the lens, *nl*.

#### **3. Experimental Results and Discussion**

To demonstrate the feasibility of the proposed method, three convex mirrors (Edmund Optics) labeled TS1, TS2, and TS3, with focal lengths of −7.751, −25.80, −52.08 mm, respectively, and radius of curvatures of −15.50, −51.60, and −104.2 mm, respectively, and two BK7 concave lenses (Edmund Optics) (*nl* = 1.5151 @ *λ* = 0.6328 μm) labeled TS4 and TS5, with focal lengths −50.00 and −200.0 mm, respectively, and radius of curvatures of −51.51 and −206.0 mm, respectively, were measured. The photos of experiment setup are shown in Figure 3a,b. A Ronchi grating of laser direct writing photomask (MLA150, HEIDELBERG INSTRUMENTS) with a period of 0.2822 mm was used. A 0.6328 μm helium–neon laser was used as the test light source and a digital camera (Canon EOS 650D) was used to capture the intensity of the test light field. The vertical pixel array of the camera was aligned to the pattern of light field along *y*-direction. For convenience, the distances *d*1 and *d*2 were both set at 251.7 mm. The captured intensities of TS1–5

(on the observation screen) are illustrated in Figure 4a–e, respectively. The gray-level transformed pictures are illustrated in Figure 4f–j, respectively. The corresponding digital gratings with periods of 9.480, 3.072, 1.659, 1.720, and 0.6410 mm which were set similar to that of the gray-level transformed patterns are shown in Figure 4k–o. The angle *θ* between the practical and digital grating was set as 30◦. Accordingly, the superimposed moiré patterns were obtained as illustrated in Figure 4p–t, respectively. With these moiré patterns, the obtained slant angles, *α*, of TS1–5 were 15.46◦, 15.35◦, 15.79◦, 15.83◦, and 15.50◦, respectively. After substituting the relevant parameters into Equations (10)–(12), as listed in Table 1, the focal lengths of TS1–5 were −7.757, −25.67, −52.05, −49.85, and −199.6 mm, respectively, and the radius of curvatures were −15.51, −51.34, −104.1, −51.36, and −205.6 mm, respectively. Compared with commercial reference values, the results exhibited a high accuracy of measurement of the focal lengths and radius of curvatures, and the percent errors were less than 0.5%.

**Figure 3.** Experiment setup for measuring (**a**) reflection and (**b**) transmission samples. SF, spatial light filter; G, Ronchi grating; TS, test sample; OS, observation screen.

**Figure 4.** *Cont*.

**Figure 4.** Captured intensities on the screen for test samples (TSs): (**a**) TS1, (**b**) TS2, (**c**) TS3, (**d**) TS4, and (**e**) TS5. Gray-level transformed images of the captured intensity for the TSs: (**f**) TS1, (**g**) TS2, (**h**) TS3, (**i**) TS4, and (**j**) TS5. Digital gratings with a slant angle of 30◦ for TSs: (**k**) TS1, (**l**) TS2, (**m**) TS3, (**n**) TS4, and (**o**) TS5. Moiré patterns obtained with digital image post processing of the TSs: (**p**) TS1, (**q**) TS2, (**r**) TS3, (**s**) TS4, and (**t**) TS5.



#### *3.1. Determination of the Slant Angle for the Digital Gratings*

The relationships between the slant angle α of the moiré pattern and the angle *θ* between the practical and digital gratings were simulated according to Equation (9). These relationships are associated with the value of *M* (the adapted period ratio for two period patterns), as shown in Figure 5. When *M* = 1, the relationship between *α* and *θ* is linear. In practice, the period of the digital grating typically is slightly different from that of the captured pattern. Bending curves are generated, which gradually deviate from the linear (*M* = 1) according to the deviation of the value of *M* from 1. When *M* < 1, the curves lie above the linear line. By contrast, when *M* > 1, the curves lie below the linear line. The values of *M* for TS1–TS5 were 0.9964, 0.9881, 0.9922, 0.9900, and 0.9943, respectively (Figure 5). The value of the slant angle α varies significantly with small variations in the

angle *θ* when *θ* < 2.5◦. By contrast, *α* has a gentle slope when *θ* > 10◦. In addition, a large value of *α*, approaching 90◦ (with a small angle *θ*), should be avoided to avert an extreme value of the tangent function of *α* in Equations (10)–(12). Therefore, the ideal angle *θ* was set as 30◦, as illustrated in Figure 5 and Table 1.

**Figure 5.** The relationships between the slant angle *α* of moiré pattern and the angle *θ* between practical and digital gratings.

#### *3.2. Analysis of Measurement Uncertainty*

With Equation (10), the measurement uncertainty of the focal length can be expressed as

$$
\Delta f = \sqrt{\left(\frac{\partial f}{\partial p\_1} \Delta p\_1\right)^2 + \left(\frac{\partial f}{\partial p\_2} \Delta p\_2\right)^2 + \left(\frac{\partial f}{\partial d\_2} \Delta d\_2\right)^2 + \left(\frac{\partial f}{\partial a} \Delta a\right)^2 + \left(\frac{\partial f}{\partial \theta} \Delta \theta\right)^2}, \tag{13}
$$

with

$$\frac{\partial f}{\partial p\_1} = \frac{-p\_2 d\_2 (\tan a \sin \theta + \cos \theta)}{\left[p\_1 (\tan a \sin \theta + \cos \theta) - p\_2\right]^2} \tag{14}$$

$$\frac{\partial f}{\partial p\_2} = \frac{p\_1 d\_2 (\tan \alpha \sin \theta + \cos \theta)}{\left[p\_1 (\tan \alpha \sin \theta + \cos \theta) - p\_2\right]^2} \tag{15}$$

$$\frac{\partial f}{\partial d\_2} = \frac{p\_1(\tan \alpha \sin \theta + \cos \theta)}{p\_1(\tan \alpha \sin \theta + \cos \theta) - p\_2},\tag{16}$$

$$\frac{\partial f}{\partial \alpha} = \frac{-p\_1 p\_2 d\_2 \sec^2 \alpha \sin \theta}{\left[p\_1(\tan \alpha \sin \theta + \cos \theta) - p\_2\right]^2} \tag{17}$$

$$\frac{\partial f}{\partial \theta} = \frac{-p\_1 p\_2 d\_2 (\tan \alpha \cos \theta - \sin \theta)}{\left[p\_1 (\tan \alpha \sin \theta + \cos \theta) - p\_2\right]^2} \tag{18}$$

where Δ*p*1, Δ*p*2, Δ*d*2, Δ*α*, andΔ*θ* are errors of the periods of the practical and digital gratings, the distance *d*2, the slant angle of the moiré pattern, and the relative angle between these gratings, respectively. The value of Δ*p*1 equals to 0.005 mm which corresponds to the

fabrication line width of photomask (MLA150, HEIDELBERG INSTRUMENTS). With the assistance of a laser rangefinder (GCL 25, Bosch), the distance from the grating to the vertex of test samples was measured which was used to determine the value of *d*2 and therefore Δ*d*2 equals to 0.3 mm. For spherical mirrors, *d*2 equals the measured value. For spherical lenses, *d*2 equals the distance from the grating to the lens principal point. Therefore, the distance from the vertex to the principal point which can be estimated as one-third the lens center thickness must be added into the measured value. Because the period *p*2, the angles *α* and *θ* were measured on digital images, the measurement values and deviations are related to the number of pixels, and the physical scale corresponding to one pixel. Shown in Figure 6 is a digital image of captured pattern with 15 megapixel (5184 pixel × 2912 pixel) in which an area with 1200 pixel × 1200 pixel and a scale bar of 50 mm with 1761 pixels are marked. Therefore, the value of Δ*p*2 equals to 0.028 mm. Figure 7a,b illustrate how to estimate an angle Θ and an angle deviation ΔΘ on a digital image with *K* × *J* pixels. The angle Θ can be expressed as

$$\Theta = \tan^{-1} \left[ \frac{K(\text{pixels})}{J(\text{pixels})} \right],\tag{19}$$

and the angle deviation ΔΘ can be written as

$$\Delta\Theta = \tan^{-1}\left[\frac{1(\text{pixel})}{\sqrt{K^2 + \bar{f}^2}(\text{pixels})}\right]. \tag{20}$$

**Figure 6.** A digital image (5184 pixel × 2912 pixel) of captured pattern marked with scale bar and number of pixels.

**Figure 7.** Schematic of the evaluation of (**a**) angle and (**b**) angle deviation on a digital image.

Therefore, with Equation (20), considering the digital image with 1200 × 1200 pixels (*K* = *J* = 1200) in Figure 6, the angle deviations of Δ*α* and Δ*θ* equal to 0.03◦. According to Equation (13), the measurement uncertainty Δ*f* of the focal length for TS1–5 can be evaluated with values of 0.2, 0.8, 2.0, 1.9, 18.3 mm, respectively. The values of relevant terms, parameters, and measurement uncertainty Δ*f* are summarized in Table 2. In order to clarify the large measurement uncertainty of TS5, from Equations (14)–(18), we can find that when *p*1 and *p*2 have close values, the term of [*p*1(tan*α*sin*θ* + cos*θ*) − *p*2]<sup>2</sup> in the denominator will magnify introduced errors of Δ*p*1, Δ*p*2, Δ*α*, and Δ*θ* (due to tan*α*sin*θ* + cos*θ* ∼= 1, when *α* ∼= 15◦ and *θ* = 30◦). By contrast, in the cases of TS1 and TS2, *p*2 has a value much larger than *p*1 (>>1), therefore the measurement uncertainty can reach sub-millimeter.

**Table 2.** Values for relevant terms and parameters for the measurement uncertainty of focal length.


rewritten as

Furthermore, with Equation (9) and condition of *M* ≈ 1, Equations (14)–(18) can be

$$\frac{\partial f}{\partial p\_1} = \frac{-\left(p\_2/M\right)d\_2}{\left[\left(p\_1/M\right) - p\_2\right]^2} \approx \frac{-p\_2d\_2}{\left(p\_1 - p\_2\right)^2} = \frac{f}{p\_1} \cdot \mathcal{S}\_{1\prime} \tag{21}$$

$$\frac{\partial f}{\partial p\_2} = \frac{(p\_1/M)d\_2}{\left[\left(p\_1/M\right) - p\_2\right]^2} \approx \frac{p\_1d\_2}{\left(p\_1 - p\_2\right)^2} = \frac{f}{p\_1} \cdot \text{S}\_{2\prime} \tag{22}$$

$$\frac{\partial f}{\partial d\_2} = \frac{p\_1/M}{(p\_1/M) - p\_2} \approx \frac{p\_1}{p\_1 - p\_2} = S\_{3\prime} \tag{23}$$

$$\frac{\partial f}{\partial a} = \frac{-p\_1 p\_2 d\_2 \sec^2 a \sin \theta}{\left[\left(p\_1/M\right) - p\_2\right]^2} \approx \frac{-p\_1 p\_2 d\_2 \sec^2 a \sin \theta}{\left(p\_1 - p\_2\right)^2} = f(\sec^2 a \cdot \sin \theta) \cdot \text{S}\_1 \tag{24}$$

$$\frac{\partial f}{\partial \theta} = \frac{-p\_1 p\_2 d\_2 (\tan a \cos \theta - \sin \theta)}{\left[ (p\_1/M) - p\_2 \right]^2} \approx \frac{-p\_1 p\_2 d\_2 (\tan a \cos \theta - \sin \theta)}{\left( p\_1 - p\_2 \right)^2} = -f(\tan a \cos \theta - \sin \theta) \cdot S\_{2\prime} \tag{25}$$

with

$$S\_1 = \frac{-d\_r \left| 1 - d\_r \right|}{\left[ 1 - \left| 1 - d\_r \right| \right]^2},\tag{26}$$

$$S\_2 = \frac{d\_r}{\left[1 - |1 - d\_{l'}|\right]^2} \,\tag{27}$$

$$\text{and}$$

$$S\_3 = \frac{1}{1 - |1 - d\_{\bar{l}}|},\tag{28}$$

where *dr* is defined as relative distance which equals *d*2/*f*. In Equations (21)–(25) (with sec<sup>2</sup> *α* sin *θ* < 1 and (tan *α* cos *θ* − sin *θ*) < 1) it is obvious that smaller absolute values of *S*1, *S*2, and *S*3 ensure relatively small introduced errors which are directly related to the relative distance in Equations (26)–(28). Therefore, the relationships between *S*1, *S*2, and *S*3 to the relative distance *dr* are plotted in Figure 8a–c. The simulation results reveal that a smaller measurement uncertainty can be achieved with increasing the distance of *d*2 for cases of *f* > 0 and *f* < 0. In the case of *f* > 0, a smaller measurement uncertainty can also be obtained when *d*2 is near the focal length; meanwhile, the distance of *d*2 = 2*f* must be avoided. In addition, from Equations (21) and (22), properly increasing the period *p*1 of the grating can also reduce the measurement uncertainty. Because a small measurement uncertainty ultimately relies on a high-precision control of experimental conditions. Considering errors of Δ*p*1 = 0.001 mm, Δ*p*2 = 0.01 mm, Δ*d*2 = 0.01 mm, Δ*α* = Δ*θ* = 0.01◦ under accurate controls, theoretical evaluations for the measurement uncertainty of focal length under different focal lengths, relative distances, and periods of grating are summarized in Table A1, in Appendix A. Accordingly, this method could be applied to regular commercial spherical lenses and mirrors ( −200 mm< *f* < +200 mm) with Δf < 1.6 mm within a measurement distance of 2 m. Considering the limited length of optical benches, there must be a trade-off between measurement resolution and measurement distance. In addition, distortion of the grid image may occur on the observation screen during measurement which is a function of the off-axis image distance. Therefore, it is necessary to capture the image of the paraxial area to avoid introducing additional errors on the moiré fringes.

**Figure 8.** *Cont*.

#### *3.3. Simulation of Light Field Distributions*

The intensity behind the grating (with a period *p*1 = 0.2822 mm) and its intensity profile were simulated according to Equation (2), as shown in Figure 9. The intensities on the observation screen, and their intensity profiles, the corresponding digital grating with *θ* = 30◦, and their moiré pattern for TS1–5 were simulated (*p*1 = 0.2822 mm, *d*1 = *d*2 = 251.7 mm, and *θ* = 30◦) according to Equation (8), as shown in Figure 10. The simulation results corresponded suitably with the experimental results shown in Figure 4, thus confirming the correctness of the derived equations.

**Figure 9.** Simulated (**a**) intensities behind the grating (with a period of *p*1 = 0.2822 mm) and (**b**) the intensity profile.

**Figure 10.** *Cont*.

**Figure 10.** Simulated (**a**) intensities on the observation screen and (**b**) intensity profiles, (**c**) corresponding digital gratings, and (**d**) the moiré patterns for samples: TS1–TS5.

Compared to Equation (2), Equation (8) contains a cosine modulation term, which varies the intensity distribution on the observation screen. Therefore, the principal maximum of the intensity profile decreased and a subsidiary maximum occurred (Figure 10). The energy splitting reduced the intensity contrast. The period of the principal maximum of the intensity profile on the observation screen was *p*1 *(*1 − *d*2*/f)*. When the Gaussian law was followed, (i.e., 1/*f =* 1*/d*1 *+* 1*/d2*), Equation (8) degenerated to Equation (2) with an imaging magnification of *m* = |1 − *d*2*/f* |. Because of advancements in the technology used in personal computers, the field intensity distribution of light propagation can be visualized. Considering a special case of *d*1 *= zT +* 2*f* and *d*2 *=* 2*f*, where *zT =* <sup>2</sup>*p*1*2/<sup>λ</sup>* is the Talbot distance, Equation (8) degenerates to Equation (2) with an imaging magnification of *m* = 1, and the simulation of the intensity carpet (*d*2 versus *x*2) is shown in Figure 11a (with *λ* = 0.6328 μm, *p*1 = 0.2822 mm, *f* = 100 mm). As shown in Figure 11b, at a distance *d*2 *= 2f = 200* mm, the intensity profile is exactly equal to that in Figure 9b, which is the self-image of grating G at *zT* = 251.7 mm with *m =* 1. In the case of TS5, a sub-moiré pattern was observed (Figure 10 TS5: (d)) because of the subsidiary maximum of the intensity profile. The sub-moiré pattern was parallel to the principal moiré pattern, which did not affect the determination of the slant angle α of the moiré pattern. Because of the introduction of digital-grating, measurements taken using this method require only one practical grating and a single photograph with digital image post-processing. The proposed method could be applied to regular commercial spherical mirrors and lenses (bi-convex or bi-concave).

**Figure 11.** Simulated (**a**) intensity carpet (with *λ* = 0.6328 μm, *p*1 = 0.2822 mm, *f* = 100 mm) and (**b**) the intensity profiles at *d*2=200mm.
