Methods for analyzing the origin of interferometric rainbow problems in virtual reality imaging systems

Abstract

This invention discloses a method for analyzing the source of interference rainbow problems in a virtual reality imaging system, comprising the following steps: Step 1, obtaining an interference rainbow image Icr; Step 2, obtaining a filtered interference rainbow binary image Icrb2; Step 3, determining whether the white part in the middle of the filtered interference rainbow binary image Icrb2 is circular or elliptical; if the white part in the middle of the filtered interference rainbow binary image Icrb2 is circular, then the interference rainbow problem originates from the waveplate; if the white part in the middle of the filtered interference rainbow binary image Icrb2 is elliptical, and the Euclidean distance L(O1,O) between the center point O1 of the ellipse and the center point O of the image is greater than or equal to a threshold T3, then the rainbow interference is caused by light not being perpendicular to the incident aspherical lens; if it is less than the threshold T3, then the rainbow interference is caused by a quality problem with the aspherical lens.

Description

Technical Field

[0001] This invention relates to a quality analysis algorithm for VR and AR visual imaging systems. Specifically, it is an analysis algorithm based on multiple sources of interference rainbow problems, which can provide objective detection parameters for improving the quality of imaging systems. Background of the Invention

[0002] In the field of virtual reality, since the human eye receives 70% of all information from the five senses (sight, touch, hearing, taste, and smell), virtual vision is the most crucial element in realizing virtual reality. Currently, virtual vision primarily relies on virtual imaging systems. The most mainstream virtual imaging system is the PANCAKE scheme, which includes aspherical lenses, waveplates, polarizers, displays, and several other lenses used in different specific solutions.

[0003] Currently, one of the challenges facing PANCAKE-based solutions is the interference rainbow effect that ultimately appears to the user. The desired effect is a rainbow-colored display, resembling multiple outward-expanding wave patterns (or possibly irregular wave patterns) around the central area of ​​a virtual display, similar to the effect created when a stone is thrown into calm water.

[0004] There are two main reasons for this:

[0005] [1]. When light passes through a 1 / 4 wavelength glass plate, it is not incident perpendicularly, which to some extent forms a conical light, resulting in an optical path difference between the center and the edge, which in turn produces interference and forms patterns.

[0006] [2]. Uneven internal stress in a single aspherical lens causes light to form an irregular imaging effect within it.

[0007] Clearly, it is necessary to reduce the effect of polarization interference rainbows. However, before achieving this, an automated and objective analysis of this phenomenon is needed to provide objective evidence for subsequent problem-solving. The analysis of its sources mainly includes three aspects:

[0008] [1]. Quality issues with single aspherical lenses;

[0009] [2]. Quality problems caused by waveplate issues;

[0010] [3]. Quality problems caused by the tilting of the optical path.

[0011] Currently, no relevant research papers or patents have been found regarding this issue. Summary of the Invention

[0012] This invention proposes an analytical method for interference rainbow problems in virtual imaging systems, based on existing optical systems and image recognition technology. The main steps include: multicolor background statistics, spatial domain analysis, frequency domain analysis, internal stress analysis of a single aspherical lens, problem source analysis, and output. This algorithm can analyze three problem sources: waveplate generation, optical path tilting, and single aspherical lens quality issues, and provides reasonable analytical indicators for their changes, which can be used for product quality analysis and correction.

[0013] A method for analyzing the source of interference rainbow problems in a virtual reality imaging system, wherein the virtual reality imaging system is based on the PANCAKE system, includes the following steps:

[0014] Step 1: Form interference rainbow images on multiple monochromatic backgrounds, remove the monochromatic backgrounds and keep only the interference rainbow images, and take the average value of multiple interference rainbow images to obtain the interference rainbow image Icr;

[0015] Step 2: Perform grayscale processing on the interference rainbow image Icr to obtain the interference rainbow grayscale image Icrg. Then, filter the interference rainbow grayscale image Icrg to obtain the filtered interference rainbow grayscale image Icrg2. Finally, perform binarization to obtain the filtered interference rainbow binary image Icrb2.

[0016] Step 3: Determine whether the white part in the middle of the filtered interference rainbow binary image Icrb2 is circular or elliptical;

[0017] If the white part in the middle of the filtered binary interference rainbow Icrb2 is circular, and the light in the PANCAKE system is perpendicular to the single aspherical lens, the interference rainbow problem is only generated from the waveplate.

[0018] If the white part in the middle of the filtered interference rainbow binary image Icrb2 is elliptical, and the Euclidean distance L(O1,O) between the center point O1 of the ellipse and the center point O of the image is greater than or equal to the threshold T3, then the rainbow interference is caused by the light not being perpendicular to the incident aspherical lens in the PANCAKE system; the center point O1 of the ellipse refers to the center point of the minimum bounding rectangle Rec_Icrb2.

[0019] If the white part in the middle of the filtered binary interference rainbow image Icrb2 is elliptical, and the Euclidean distance L(O1,O) between the center point O1 of the ellipse and the center point O of the image is less than the threshold T3, then the rainbow interference is caused by the quality problem of the single aspherical lens.

[0020] Furthermore, in step 1, interference rainbow images are formed on multiple monochromatic backgrounds. The monochromatic backgrounds are removed, leaving only the interference rainbow images. The average value of multiple interference rainbow images is then taken to obtain the interference rainbow image Icr. This process specifically includes the following steps:

[0021] Step 1.1: Output single-color images of white, black, green and blue respectively, and capture them as white image Iw, black image Ik, green image Ig and blue image Ib respectively. Each of these images contains one or more interference rainbows.

[0022] Step 1.2: For the white image Iw, black image Ik, green image Ig, and blue image Ib, randomly sample the color average avg1 and its variance dev1 of N2 points at a Euclidean distance N1 from the image center of each image; the sampling points cannot fall on the interference rainbow image. Then remove all points with avg1±5*dev1 from the white image Iw, black image Ik, green image Ig, and blue image Ib to obtain the background-removed interference rainbow images Iwr, Irr, Igr, and Ibr, respectively.

[0023] Step 1.3: In the background-removed interference rainbow images Iwr, Irr, Igr, and Ibr, the channel values ​​at the same pixel position (i, j) are added together and averaged to obtain the composite interference rainbow image Icr.

[0024] Icr(i,j,ch)=(Iwr(i,j,ch)+Irr(i,j,ch)+Igr(i,j,ch)+Ibr(i,j,ch)) / 4

[0025] Where ch = 1 to 3, representing the red channel, green channel, and blue channel, respectively.

[0026] Furthermore, in step 2, the interference rainbow image Icr is processed into grayscale to obtain the interference rainbow grayscale image Icrg. Then, the interference rainbow grayscale image Icrg is filtered to obtain the filtered interference rainbow grayscale image Icrg2. Finally, it is binarized to obtain the filtered interference rainbow binary image Icrb2. Specifically, this includes the following:

[0027] Step 2.1: Convert the interference rainbow image Icr into a grayscale image to obtain the interference rainbow grayscale image Icrg. Perform a two-dimensional Fourier transform on the rainbow grayscale image Icrg and then center it to obtain the interference rainbow spectrum image Fcr.

[0028] Step 2.2. Use the Otsu method to binarize the rainbow spectrum Fcr to obtain the rainbow spectrum binary image Fcrb;

[0029] Step 2.3. In the largest white region of the rainbow spectrum binary image Fcrb, find the largest white circle inside it. Then, except for the region of this white circle, set all other parts of the entire Fcrb to 0 to obtain the filter template Ccrb.

[0030] Step 2.4. Multiply the filter template Ccrb by each pixel of the interferometric rainbow spectrum Fcr to obtain the filtered two-dimensional Fourier transform image Fcr2. The two-dimensional Fourier transform image Fcr2 retains the gray values ​​of the circular regions in the rainbow interferometric rainbow spectrum Fcr, while the gray values ​​of other regions are 0.

[0031] Step 2.5. Decenter the filtered two-dimensional Fourier transform graph Fcr2, and then perform an inverse two-dimensional Fourier transform to obtain the filtered graph Icrg2.

[0032] Step 2.6. Binarize Icrg2 to obtain the filtered binary interference rainbow image Icrb2.

[0033] Furthermore, if step 3 determines that the rainbow interference is caused by a quality problem with the single aspherical lens, it is necessary to further determine whether the rainbow interference is caused by excessive uneven stress on the single aspherical lens, including the following steps:

[0034] Step 4: Binarize the interference rainbow image Icr to obtain the binarized interference rainbow image Icrb. Calculate the distance of the first interference rainbow ring in the binarized interference rainbow image Icrb from the center point of the image and its own width.

[0035] Step 5: Calculate the distance of the first interference rainbow ring from the center point of the image and its width in the filtered binary interference rainbow image Icrb2;

[0036] Step 6: Subtract the distance between the first interference rainbow ring in Icrb2 and the distance between the first interference rainbow ring in Icrb and the center point of the image, and take the absolute value to obtain DiffR;

[0037] The difference between the width of the first interference rainbow ring in Icrb2 and the width of the first interference rainbow ring in Icrb, and the absolute value of the difference, is used to obtain DiffD.

[0038] When DiffR+DiffD is greater than the threshold T1, it is an interference rainbow caused by the uneven stress inside the single aspherical lens.

[0039] Furthermore, in step 4, the interference rainbow image Icr is binarized to obtain a binarized interference rainbow image Icrb. The distance and width of the first interference rainbow ring in the binarized interference rainbow image Icrb from the image center point are calculated, specifically including the following:

[0040] Step 4.1. Convert the interference rainbow image Icr into a binary image Icrb containing only 0s and 1s;

[0041] Step 4.2. In Icrb, starting from the center point O of the image, draw a ray L1 with a horizontal direction of 0 degrees to the right. The first point where L1 intersects with the part of Icrb with a value of 1 is P1L1. Then continue to trace along L1 to find the first point where it intersects with the part of Icrb with a value of 0, which is P2L1. Then calculate the Euclidean distance RL1 from P1L1 to point O and the Euclidean distance D1inL1 from P1L1 to point P2L1.

[0042] Step 4.3: Continuing in the image Icrb, starting from the center point O, extend horizontally to the right by 0 degrees and increase counterclockwise by 1 degree, drawing a ray L2. The first point where L2 intersects with the portion of Icrb with a value of 1 is P1L2. Then continue tracing along L2 to find the first point where it intersects with the portion of Icrb with a value of 0, which is P2L2. Then calculate the Euclidean distance RL2 from P1L2 to point O and the Euclidean distance D1inL2 from P1L2 to point P2L2.

[0043] Step 4.4: Rotate the image 360 ​​degrees sequentially as in Step 4.3, and find the array RL = [RL1, RL2, ..., RL360] to represent the distance of the first interference rainbow ring from the center point of the image, and D1inL = [D1inL1, D1inL2, ..., D1inL360] to represent the width of the first interference rainbow ring.

[0044] Furthermore, in step 5, the method for calculating the distance and width of the first interference rainbow ring in Figure Icrb2 from the center point of the image is the same as in step 4, resulting in the array RRL = [RRL1, RRL2, ..., RRL360] used to represent the distance of the first interference rainbow ring in Figure Icrb2 from the center point of the image, and RD1inL = [RD1inL1, RD1inL2, ..., RD1inL360] used to represent the width of the first interference rainbow ring in Figure Icrb2.

[0045] Furthermore, step 6 includes the following:

[0046] Step 6.1: Subtract the corresponding data in array RL from the data in array RRL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffR;

[0047] Step 6.2: Subtract the corresponding data in array D1inL from the data in array RD1inL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffD;

[0048] Step 6.3. DiffY = DiffR + DiffD

[0049] Step 6.4. Set the threshold T1. If DiffY>T1, it is because the internal stress of the single aspherical lens is too large and causes interference rainbow.

[0050] Furthermore, the method for determining whether the white portion in the middle of the filtered interference rainbow binary image Icrb2 is circular or elliptical in step 3 is as follows:

[0051] Step 3.1: Calculate the minimum outer rectangle Rec_Icrb2 of the filtered interference rainbow binary image Icrb2 to find the minimum outer rectangle Rec_Icrb2 of the first interference rainbow ring.

[0052] Step 3.2: Measure the minimum bounding rectangle Rec_Icrb2. If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is less than the threshold T2, it is considered to be circular.

[0053] If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is greater than or equal to the threshold T2, it is considered to be elliptical.

[0054] Beneficial Effects: This invention provides a method for analyzing interference rainbow problems in virtual imaging systems, identifying three sources of the problem: waveplate generation, optical path tilting, and quality issues with the aspherical lens. Furthermore, when the source of the interference rainbow is a quality issue with the aspherical lens, this invention also indicates whether the interference rainbow is caused by excessive uneven stress in the aspherical lens, for use in product quality analysis and correction. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the optical path according to an embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram illustrating the effect of rainbow interference.

[0057] Figure 3 This is a schematic diagram showing the positions of several points in step 2 of this embodiment of the invention;

[0058] Figure 4 This is a schematic diagram showing the positions of several points in step 5 of this embodiment of the invention;

[0059] Figure 5 This is an overall flowchart of an embodiment of the present invention; Specific Implementation

[0060] The following is in conjunction with the appendix Figure 1 ~Attached Figure 5 To further illustrate the embodiments of this solution.

[0061] This invention proposes an analytical method for interference rainbow problems in virtual imaging systems, based on existing optical systems and image recognition technology. The main steps include: multicolor background statistics, spatial domain analysis, frequency domain analysis, internal stress analysis of a single aspherical lens, problem source analysis, and output. This algorithm can analyze three problem sources: waveplate generation, optical path tilting, and single aspherical lens quality issues, and provides reasonable analytical indicators for their changes, which can be used for product quality analysis and correction.

[0062] like Figure 1 As shown, the PANCAKE scheme is an imaging system that includes a single aspherical lens, waveplate, polarizer, display, and several other lenses in different specific schemes. Figure 1 The leftmost image shows the human eye, followed by, in order, a single aspherical lens, a polarizing filter, a quarter-wave plate, and a monitor. Between the quarter-wave plate and the monitor, there are various lens options under different PANCAKE schemes.

[0063] A method for analyzing the source of interference rainbow problems in virtual reality imaging systems includes the following steps:

[0064] Step 1: The interference rainbow image is formed on multiple monochromatic images. The monochromatic colors are removed and only the interference rainbow image is retained. The average value of multiple interference rainbow images is then taken to obtain the interference rainbow image Icr.

[0065] Step 1.1: Control the display output to output single-color images of white, black, green, and blue respectively. During display, capture images as white image Iw, black image Ik, green image Ig, and blue image Ib respectively. These images contain one or more interference rainbows. Figure 2 This is to demonstrate the extreme effects of this problem, with multiple interference rainbows, but in most cases, it is as follows: Figure 3 , Figure 4 The diagram shows the case with only one interference rainbow. For a convex aspherical lens, the interference rainbow consists of red, orange, yellow, green, blue, indigo, and violet colors from the outside in; for a concave aspherical lens, the interference rainbow consists of red, orange, yellow, green, blue, indigo, and violet colors from the inside out. Both colors have red edges, so for ease of preprocessing, the red color was not controlled to be displayed during statistical analysis.

[0066] Step 1.2: For the white image Iw, black image Ik, green image Ig, and blue image Ib, randomly sample the color average avg1 and its variance dev1 of N2 points at a Euclidean distance N1 from the image center of each image. The sampling points cannot fall on the interference rainbow image. Then, remove all points with avg1±5*dev1 from the white image Iw, black image Ik, green image Ig, and blue image Ib to obtain the interference rainbow images Iwr, Irr, Igr, and Ibr with the image background removed, respectively.

[0067] In this embodiment, N1 is 10 pixels and N2 is 10 points.

[0068] Step 1.3: Add the values ​​of each channel in each position of the interference rainbow images Iwr, Irr, Igr, and Ibr after removing the background image. If there is only one overlap of points in Iwr, Irr, Igr, and Ibr at the same pixel position, keep that one. If there is more than one overlap of points in Iwr, Irr, Igr, and Ibr, average the three channel values ​​(red, green, and blue) of each image at that point to obtain the composite interference rainbow image Icr.

[0069] For example, for position (i,j), if only image Iwr has a value at position (i,j) among images Iwr, Irr, Igr, and Ibr, then the three channel values ​​at position (i,j) in image Iwr are assigned to the three channels at position (i,j) in the new image Icr, respectively.

[0070] Icr(i,j,ch)=Iwr(i,j,ch), where ch=1~3;

[0071] The three channels refer to the red channel, green channel, and blue channel respectively; ch=1 represents the red channel, ch=2 represents the green channel, and ch=3 represents the blue channel.

[0072] For example, at position (i,j), if images Iwr, Irr, Igr, and Ibr all have values ​​at position (i,j), then the average of the three channel values ​​from the four images is assigned to the three channels at position (i,j) in the new image Icr, i.e.:

[0073] Icr(i,j,ch)=(Iwr(i,j,ch)+Irr(i,j,ch)+Igr(i,j,ch)+Ibr(i,j,ch)) / 4, where ch=1~3;

[0074] The remaining positions are points that are not covered by any Iwr, Irr, Igr, or Ibr. This algorithm sets the values ​​of all these points to 0.

[0075] This composite image Icr is the average interference rainbow under the main different lighting conditions.

[0076] Step 2: Perform spatial domain analysis; statistically analyze the distance and width of the first interference rainbow ring in the interferometric rainbow image Icr from the image center point. This data contains noise, which is caused by the inconsistent internal stress of the single aspherical lens.

[0077] Step 2.1. Convert the interference rainbow image Icr into a binary image containing only 0 and 1, that is, set all points with non-zero values ​​to 1. In other words, set all the points covered by Iwr, Irr, Igr, and Ibr in step 1.3 to 1, and obtain the interference rainbow binary image Icrb.

[0078] Step 2.2. Figure 3 As shown, in Icrb, starting from the center point O of the displayed image, a ray L1 is drawn horizontally to the right at 0 degrees. The first point where L1 intersects with the part of Icrb with a value of 1 is P1L1. Then, continue tracing along L1 to find the first point where it intersects with the part of Icrb with a value of 0, which is P2L1. Then, calculate the distances RL1 and D1inL1. RL1 is the Euclidean distance between P1L1 and point O, and D1inL1 is the Euclidean distance between P1L1 and point P2L1.

[0079] Then, continuing in Icrb, starting from the center point O of the displayed image, extend horizontally to the right by 0 degrees and increase counterclockwise by 1 degree, drawing a ray L2. The first point where L2 intersects with the portion of Icrb with a value of 1 is P1L2. Then, continue tracing along L2 to find the first point where it intersects with the portion of Icrb with a value of 0, which is P2L2. Then calculate the distances RL2 and D1inL2. RL2 is the Euclidean distance between P1L2 and point O, and D1inL2 is the Euclidean distance between P1L2 and point P2L2.

[0080] Rotate in this manner 360 times to find the array RL = [RL1, RL2, ..., RL360] which represents the distance of the first interference rainbow ring from the center point of the image, and D1inL = [D1inL1, D1inL2, ..., D1inL360] which represents the width of the first interference rainbow ring; these data will be used in the irregularity measurement step.

[0081] Step 3, Frequency Domain Filtering

[0082] Step 3.1. Convert the interference rainbow image Icr to a grayscale image to obtain the interference rainbow grayscale image Icrg. Perform a two-dimensional Fourier transform on the rainbow grayscale image Icrg, and then center it to obtain the interference rainbow spectrum Fcr. The brighter areas in the grayscale image represent higher energy, and their positions are coordinates of two two-dimensional frequencies. This step is well-known and will not be elaborated upon in this invention.

[0083] Step 3.2. Use the Otsu method to binarize the rainbow spectrum Fcr to obtain the rainbow spectrum binary image Fcrb;

[0084] Step 3.3. In the largest white region (i.e., the largest area of ​​1) in the binary spectrum Fcrb, find the largest white circle (with a value of 1) inside it. Then, set all other parts of the Fcrb except for this white circle to 0, thus obtaining the filter template Ccrb. In the filter template Ccrb, only the circle region has a value of 1, and the other regions have a value of 0.

[0085] Step 3.4. Multiply the filter template Ccrb by each pixel of the interferometric rainbow spectrum Fcr to obtain the filtered two-dimensional Fourier transform image Fcr2. The two-dimensional Fourier transform image Fcr2 retains the gray values ​​of the circular regions in the rainbow interferometric rainbow spectrum Fcr, while the gray values ​​of other regions are 0.

[0086] Step 3.5. Decentralize Fcr2 and then perform an inverse two-dimensional Fourier transform to obtain the filtered image Icrg2.

[0087] Step 3.6. Binarize Icrg2 to obtain the filtered binary interference rainbow image Icrb2. This will result in a filtered and shaped binary interference rainbow image, which may appear as a circle or an ellipse. Let's assume the center of the circle or ellipse is O1.

[0088] Step 4, Internal stress analysis of a single aspherical lens

[0089] Icrb2 is the filtered interferometric rainbow binary image, reflecting the overall structure of the image; Icrb is the unfiltered, noisy image, the noise of which originates from the inconsistency of internal stress in the single aspherical lens. Therefore, calculating the difference in their detection parameters can reflect the effect of the inconsistency of internal stress in the single aspherical lens in the spatial domain.

[0090] Step 4.1. Following the method in Step 2, calculate the distance and width of the first interference rainbow ring from the center point of the image in the filtered binary interference rainbow image Icrb2; also find the array RRL = [RRL1, RRL2, ..., RRL360] to represent the distance of the first interference rainbow ring from the center point of the image, and RD1inL = [RD1inL1, RD1inL2, ..., RD1inL360] to represent the width of the first interference rainbow ring.

[0091] Step 4.2. Subtract the corresponding data in array RL from the data in array RRL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffR.

[0092] Step 4.3. Subtract the corresponding data in array D1inL from the data in array RD1inL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffD.

[0093] Step 4.4. DiffY = DiffR + DiffD

[0094] Step 4.5. Set a threshold T1. If DiffY > T1, it indicates that there is an excessively large inconsistency in the internal stress of the single aspherical lens. In this embodiment, T1 is set to 20.

[0095] Step 4.6. This step is optional. You can refer to Zhang Zhengyou's chessboard calibration method. Output a standard chessboard grid from the display screen, obtain the displayed image, analyze the deviation Ds between each feature point and its correct position, and then take the average value Dsav of all feature point deviations Ds as the reference output. Zhang Zhengyou's chessboard calibration method is publicly available, and will not be elaborated upon here.

[0096] Step 5: Determine the source of the interference rainbow problem in the virtual reality imaging system.

[0097] If the white part in the middle of the filtered binary interference rainbow image Icrb2 is circular, then the center O1 of the circle generally coincides with or deviates very little from the center point O of the image. This indicates that the light in the PANCAKE system is perpendicular to the single aspherical lens, and the interference rainbow problem is only generated from the waveplate.

[0098] If the white portion in the middle of the filtered binary interference rainbow image Icrb2 is elliptical, and this white portion is formed by the outer ring of the first interference rainbow, then the aspherical lens itself may not be circular, indicating a potential quality issue with the aspherical lens. Further testing is needed at this point to check if the distance between the center O1 of the ellipse and the center point O of the image is significant. If it is significant, it indicates a tilt in the optical path, meaning the light entering the PANCAKE system is not perpendicular to the aspherical lens. This is an optical path problem requiring correction of the mechanical structure. If the distance between O1 and O is small, it indicates no tilt in the optical path, and the issue is still with the aspherical lens.

[0099] Step 5.1. As Figure 4 As shown, the minimum bounding rectangle Rec_Icrb2 of the filtered interference rainbow binary image is calculated. Finding the minimum bounding rectangle is a common topic in computer vision and will not be elaborated here.

[0100] Step 5.2. Measure the smallest outer rectangle Rec_Icrb2. If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is less than the threshold T2, it indicates that it is a qualified circle, the light path is perpendicular to the single aspherical mirror, and the interference rainbow problem only originates from the waveplate. Proceed to Step 5.6.

[0101] In this embodiment, T2 is set to 4.

[0102] Step 5.3. If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is greater than or equal to the threshold T2, it indicates an ellipse.

[0103] Step 5.4. Measure the minimum bounding rectangle Rec_Icrb2, find its center point O1, and calculate the Euclidean distance L(O1, O) between O1 and the center point O of the image.

[0104] Step 5.5. If L(O1, O) < T3, it means the optical path is not tilted, which is a problem with the quality of the single aspherical lens; if not, it means the optical path is tilted, that is, whether the entire light is incident perpendicular to the single aspherical lens in the PANCAKE system. This is a problem with the optical path and the mechanical structure needs to be corrected. In this embodiment, T3 is set to 4.

[0105] Step 6, output

[0106] Combining the content of Steps 4 and 5, output the parameter indicators, and at the same time, the pictures during the process are also provided for user investigation.

[0107] If it is only generated from the wave plate, then output the parameters RL, RD1inL;

[0108] If the optical path is tilted, then output the parameters center point O1, L(O1, O);

[0109] If the internal stress of the single aspherical lens is inconsistent and偏大, then output the parameters DiffR, DiffD, DiffY, and the optional Dsav.

Claims

1. A method for analyzing the source of interference rainbow problems in a virtual reality imaging system, wherein the virtual reality imaging system is based on the PANCAKE system, characterized in that, Includes the following steps: Step 1: Form interference rainbow images on multiple monochromatic backgrounds, remove the monochromatic backgrounds and keep only the interference rainbow images, and take the average value of multiple interference rainbow images to obtain the interference rainbow image Icr; Step 2: Perform grayscale processing on the interference rainbow image Icr to obtain the interference rainbow grayscale image Icrg. Then, filter the interference rainbow grayscale image Icrg to obtain the filtered interference rainbow grayscale image Icrg2. Finally, perform binarization to obtain the filtered interference rainbow binary image Icrb2. Step 3: Determine whether the white part in the middle of the filtered binary interference rainbow Icrb2 is circular or elliptical. If the white part in the middle of the filtered binary interference rainbow Icrb2 is circular, the light in the PANCAKE system is perpendicular to the single aspherical lens, and the interference rainbow problem is only generated from the waveplate. If the white part in the middle of the filtered interference rainbow binary image Icrb2 is elliptical, and the Euclidean distance L(O1,O) between the center point O1 of the ellipse and the center point O of the image is greater than or equal to the threshold T3, then the rainbow interference is caused by the light not being perpendicular to the incident aspherical lens in the PANCAKE system. If the white part in the middle of the filtered binary interference rainbow image Icrb2 is elliptical, and the Euclidean distance L(O1,O) between the center point O1 of the ellipse and the center point O of the image is less than the threshold T3, then the rainbow interference is caused by the quality problem of the single aspherical lens.

2. The method for analyzing the source of interference rainbow problems in the virtual reality imaging system according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 1.1: Output single-color images of white, black, green and blue respectively, and capture them as white image Iw, black image Ik, green image Ig and blue image Ib respectively. Each of these images contains one or more interference rainbows. Step 1.2: For the white image Iw, black image Ik, green image Ig, and blue image Ib, randomly sample the color average avg1 and its variance dev1 of N2 points at a Euclidean distance N1 from the center of each image. The sampling points cannot fall on the interference rainbow image. Then, all avg1±5*dev1 points in the white image Iw, black image Ik, green image Ig, and blue image Ib are removed to obtain the background-removed interference rainbow images Iwr, Irr, Igr, and Ibr, respectively. Step 1.3: In the background-removed interference rainbow images Iwr, Irr, Igr, and Ibr, the channel values ​​at the same pixel position (i, j) are added together and averaged to obtain the composite interference rainbow image Icr. Icr(i,j,ch)=(Iwr(i,j,ch)+Irr(i,j,ch)+Igr(i,j,ch)+Ibr(i,j,ch)) / 4 Where ch = 1 to 3, representing the red channel, green channel, and blue channel, respectively.

3. The method for analyzing the source of interference rainbow problems in the virtual reality imaging system according to claim 1, characterized in that, Step 2 specifically includes the following: Step 2.1: Convert the interference rainbow image Icr into a grayscale image to obtain the interference rainbow grayscale image Icrg. Perform a two-dimensional Fourier transform on the rainbow grayscale image Icrg and then center it to obtain the interference rainbow spectrum image Fcr. Step 2.

2. Use the Otsu method to binarize the rainbow spectrum Fcr to obtain the rainbow spectrum binary image Fcrb; Step 2.

3. In the largest white area of ​​the rainbow spectrum binary image Fcrb, find the largest white circle inside, and then set all other parts of the entire Fcrb except for this white circle to 0 to obtain the filter template Ccrb. Step 2.

4. Multiply the filter template Ccrb by each pixel of the interferometric rainbow spectrum Fcr to obtain the filtered two-dimensional Fourier transform image Fcr2. The two-dimensional Fourier transform image Fcr2 retains the gray values ​​of the circular regions in the rainbow interferometric rainbow spectrum Fcr, while the gray values ​​of other regions are 0. Step 2.

5. Decenter the filtered two-dimensional Fourier transform graph Fcr2, and then perform an inverse two-dimensional Fourier transform to obtain the filtered graph Icrg2. Step 2.

6. Binarize Icrg2 to obtain the filtered binary interference rainbow image Icrb2.

4. The method for analyzing the source of interference rainbow problems in the virtual reality imaging system according to claim 1, characterized in that, When rainbow interference is caused by a quality problem with a single aspherical lens, the following steps are also included: Step 4: Binarize the interference rainbow image Icr to obtain the binarized interference rainbow image Icrb. Calculate the distance of the first interference rainbow ring in the binarized interference rainbow image Icrb from the center point of the image and its own width. Step 5: Calculate the distance of the first interference rainbow ring from the center point of the image and its width in the filtered binary interference rainbow image Icrb2; Step 6: Subtract the distance between the first interference rainbow ring in Icrb2 and the distance between the first interference rainbow ring in Icrb and the center point of the image, and take the absolute value to obtain DiffR; The difference between the width of the first interference rainbow ring in Icrb2 and the width of the first interference rainbow ring in Icrb, and the absolute value of the difference, is used to obtain DiffD. When DiffR+DiffD is greater than the threshold T1, it is an interference rainbow caused by the uneven stress inside the single aspherical lens.

5. The method for analyzing the source of interference rainbow problems in a virtual reality imaging system according to claim 4, characterized in that, Step 4 specifically includes the following: Step 4.

1. Convert the interference rainbow image Icr into a binary image Icrb containing only 0s and 1s; Step 4.

2. In Icrb, starting from the center point O of the image, draw a ray L1 with a horizontal direction of 0 degrees to the right. The first point where L1 intersects with the part of Icrb with a value of 1 is P1L1. Then continue to trace along L1 to find the first point where it intersects with the part of Icrb with a value of 0, which is P2L1. Then calculate the Euclidean distance RL1 from P1L1 to point O and the Euclidean distance D1inL1 from P1L1 to point P2L1. Step 4.3: Continuing in the image Icrb, starting from the center point O, extend horizontally to the right by 0 degrees and increase counterclockwise by 1 degree, drawing a ray L2. The first point where L2 intersects with the portion of Icrb with a value of 1 is P1L2. Then continue tracing along L2 to find the first point where it intersects with the portion of Icrb with a value of 0, which is P2L2. Then calculate the Euclidean distance RL2 from P1L2 to point O and the Euclidean distance D1inL2 from P1L2 to point P2L2. Step 4.4: Rotate the image 360 ​​degrees sequentially as in Step 4.3, and find the array RL = [RL1, RL2, ..., RL360] to represent the distance of the first interference rainbow ring from the center point of the image, and D1inL = [D1inL1, D1inL2, ..., D1inL360] to represent the width of the first interference rainbow ring.

6. The method for analyzing the source of interference rainbow problems in a virtual reality imaging system according to claim 5, characterized in that, In step 5, the method for calculating the distance and width of the first interference rainbow ring in Figure Icrb2 from the center point of the image is the same as in step 4, resulting in the array RRL = [RRL1, RRL2, ..., RRL360] to represent the distance of the first interference rainbow ring in Figure Icrb2 from the center point of the image, and RD1inL = [RD1inL1, RD1inL2, ..., RD1inL360] to represent the width of the first interference rainbow ring in Figure Icrb2.

7. The method for analyzing the source of interference rainbow problems in a virtual reality imaging system according to claim 6, characterized in that, Step 6 includes the following: Step 6.1: Subtract the corresponding data in array RL from the data in array RRL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffR; Step 6.2: Subtract the corresponding data in array D1inL from the data in array RD1inL, take the absolute value of the result, and then take the average of the 360 ​​absolute values ​​to obtain DiffD; Step 6.

3. DiffY = DiffR + DiffD Step 6.

4. Set the threshold T1. If DiffY>T1, it is because the internal stress of the single aspherical lens is too large and causes interference rainbow.

8. The method for analyzing the source of interference rainbow problems in a virtual reality imaging system according to claim 1, characterized in that, The method for determining whether the white part in the middle of the filtered interference rainbow binary image Icrb2 is circular or elliptical in step 3 is as follows: Step 3.1: Calculate the minimum outer rectangle Rec_Icrb2 of the filtered interference rainbow binary image Icrb2 to find the minimum outer rectangle Rec_Icrb2 of the first interference rainbow ring. Step 3.2: Measure the minimum bounding rectangle Rec_Icrb2. If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is less than the threshold T2, it is considered to be circular. If the length difference between its two adjacent sides Rec_Icrb2_L1 and Rec_Icrb2_L2 is greater than or equal to the threshold T2, it is considered to be elliptical.

9. The method for analyzing the source of interference rainbow problems in a virtual reality imaging system according to claim 1, characterized in that, The center point O1 of the ellipse mentioned in step 3 refers to the center point of the smallest outer rectangle Rec_Icrb2.

Images