A LED multi-spectral polarized parallel light source and a light spot detection method thereof

Abstract

The present invention discloses a LED multi-spectral polarized parallel light source and a light spot detection method thereof, wherein the parallel light source comprises: a light emitting module, a shaping modulation module, a collimation module and a fixing structure; a target light beam is emitted from an LED board of the light emitting module, and passes through a diffusion sheet, an aperture and a polarization film of the modulation module in sequence to form a plurality of polarized light beams, and then passes through a lens barrel and a collimation lens of the collimation module in sequence to output a plurality of parallel light beams containing polarization information. The present invention can reduce the volume of the light source, output a plurality of parallel lights in a light tube, and can solve all light spot information at one time, and is convenient for debugging and use.

Description

Technical Field

[0001] The invention belongs to the technical field of photoelectric detection, and in particular is an LED multi-spectral polarized parallel light source and a light spot detection method thereof. Background Art

[0002] The multi-spectral time-sharing polarization imaging system is a highly integrated, small-aperture polarization imaging remote sensor. In order to evaluate the reliability of the multi-spectral time-sharing polarization imaging system, various environmental simulation tests are required, such as mechanical, thermal, and vacuum tests. A dedicated test light source is used before and after the test to check the performance changes.

[0003] The stability of polarization imaging and response capability of the multi-spectral time-sharing polarization imaging system is the premise for its reliable application after it is launched into orbit with the satellite. The functional requirements for the test light source are:

[0004] 1) The spectrum covers all working spectrums of the corresponding probe;

[0005] 2) Form more than 3 parallel lights of different angles, with the light spots distributed at different positions of the image;

[0006] 3) Easy debugging;

[0007] 4) Good installation positioning repeatability;

[0008] 5) Good radiation energy stability and repeatability;

[0009] 6) Linear polarized light output function;

[0010] 7) Miniaturized design, easy to disassemble and assemble;

[0011] The Chinese invention patent with application number 201410141134.5 discloses an LED multi-point multi-spectrum collimated test light source. The use of several small collimators can realize the test of imaging and radiation response performance changes. The combination of multiple collimators with different wavelengths also meets the requirements of multi-band detection. A full-deflection channel is set in the optical path of the selected collimator to verify the polarization measurement function. This design satisfies the polarization imaging and response capability test functions as much as possible; however, this type of light source has major problems when applied to the multi-spectral time-sharing polarization imaging system:

[0012] (1) In the form of integrating multiple independent LED collimator light tubes, one light tube can only output one beam of parallel light spot, which makes the overall volume of the light source larger and is not suitable for the small-aperture multi-spectral time-sharing polarization imaging system. (2) Independent LED collimator light tubes do not have a focusing function, and the processing errors of actual optical components and structural components make it difficult to ensure that the focal length matches the design value, causing the output light beam to be out of focus, affecting the imaging quality and being unfavorable for subsequent data processing. (3) Another problem with the large number of independent LED collimator light tubes is that the optical debugging workload is large. If there are more than 40 independent LED collimator light tubes, then more than 40 debuggings are required, which is time-consuming, labor-intensive and economically costly. (4) The polarizer is selected as a polarizer, and the size must be designed in advance, which limits the flexibility of the structural design. (5) The existing spot centroid detection algorithm requires that there can only be one spot in a window. For the detection of multiple spots, it is necessary to know the approximate position of each spot in advance and divide it into multiple windows. Once the position of the spot changes significantly, it may be detected incorrectly or the window needs to be re-specified. Summary of the invention

[0013] The present invention aims to solve the deficiencies of the above-mentioned prior art and proposes an LED multi-spectral polarized parallel light source and a light spot detection method thereof, so as to reduce the volume of the light source, output a plurality of parallel lights in one light tube, and be able to calculate all the light spot information at one time, while being easy to debug and use.

[0014] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme:

[0015] The LED multi-spectrum polarized parallel light source of the present invention is characterized in that it comprises: a light emitting module, a shaping modulation module, a collimation module and a fixing structure;

[0016] The light emitting module comprises: a cooling fan, a cooling block, and an LED board;

[0017] The shaping and modulation module comprises: an aperture, a diffuser, and a second pressure ring;

[0018] The collimation module comprises: a lens barrel, a collimating lens, and a first pressure ring;

[0019] The fixing structure includes: a fastening screw, an LED board fixing seat, an adapter, and a locking nut;

[0020] A plurality of LED lamps are welded on the front of the LED board, and the LED lamps of the same wavelength are connected in series; each LED lamp of each wavelength after the series connection is driven by one of the multiple-channel LED light source constant current drivers;

[0021] The heat sink is provided on the back of the LED board and is bonded to the LED board via heat dissipation grease;

[0022] The LED board is fixed on the housing through the LED board fixing seat; a cooling fan facing the heat sink is provided above the housing through fastening screws; the lens barrel is inserted below the housing and fixed by a locking nut;

[0023] An aperture is provided on the housing above the interior of the lens barrel, a polarizing film is pasted on the surface of the aperture, a diffusion sheet is provided on the polarizing film and fixed by a second pressing ring;

[0024] The collimating lens is provided at the lower part of the lens barrel and is fixed to the bottom of the lens barrel through a first pressing ring. The bottom of the lens barrel is connected to the light shield through an adapter. The light shield is provided between the parallel light source and the probe.

[0025] The target light beam is emitted from the LED board of the light-emitting module, and passes through the diffuser, aperture and polarizing film of the shaping and modulation module in sequence to form several beams of polarized light, and then passes through the lens barrel and collimating lens of the collimating module in sequence to output several beams of parallel light containing polarization information.

[0026] The LED multi-spectral polarized parallel light source described in the present invention is characterized in that:

[0027] A pinhole array (61) is arranged on the aperture.

[0028] The number of LED lights of each band on the LED board is calculated according to the following steps:

[0029] Step 1: Calculate the response brightness of the LED light in each band;

[0030] Step 1.1: Use a spectrometer to collect the brightness response value of the LED light within the bandwidth of any specified band, and calculate the average brightness of the specified band within the bandwidth ;

[0031] Step 1.2: Calculate the sum of the brightness of all other bands that affect the specified band , thereby obtaining the total average brightness of the specified band within the bandwidth range + ;

[0032] Step 2: Calculate the actual number of LED lights in each band;

[0033] Step 2.1: Definition is the typical entrance pupil brightness of the probe in the specified band, and calculates the proportional value / ;

[0034] Step 2.2: According to the total number of LED lamps on the LED board (4), the reference number of LED lamps in each band is set to , calculate the theoretical number of LED lights in the specified band , thus After rounding up or down, the actual number of LED lights in the specified band is obtained. .

[0035] The feature of the light spot detection method of the LED multi-spectral polarized parallel light source of the present invention is that it utilizes the parallel light source and includes the following steps:

[0036] Step 1: Image preprocessing;

[0037] Step 1.1: using the probe to collect M×N frames of images of the parallel light source, wherein M represents the number of channels of the probe, and N represents the number of acquisitions; taking N frames of images of the same band and the same polarization direction in the M×N frames of images as a group of images, thereby obtaining M groups of images;

[0038] Take the grayscale average of each pixel of the M-1 groups of images except the background channel to obtain one frame of initial bright field image for each group, thereby obtaining M-1 frames of initial bright field images;

[0039] Step 1.2: Turn off the parallel light source, build a dark room, and use the probe to collect N frames of background images at the integration time corresponding to the M channel in a dark environment according to the process of step 1.1, and calculate M-1 frames of dark field images corresponding to M-1 groups of background images except the background channel;

[0040] Step 1.3: Subtract the grayscale value of the corresponding pixel of the dark field image of the e-th frame from the initial bright field image of the e-th frame one by one, and obtain the original image of the e-th frame, thereby obtaining the original image of the M-1 frame; ;

[0041] Step 1.4: Replace the gray value of the bad pixel on the original image of the e-th frame with the median of the gray values ​​of the surrounding pixels, thereby obtaining the standard image of the e-th frame;

[0042] Step 1.5: Traverse each pixel in the standard image of the e-th frame and set the pixel above the selected threshold Pixels below the selected threshold are retained The pixels of are replaced by "0", thereby obtaining the reference image after thresholding of the e-th frame;

[0043] Step 2: Spot area detection;

[0044] Step 2.1: Traverse each pixel in the thresholded reference image of the e-th frame from left to right and from top to bottom. When the first non-zero value is detected, replace the gray value of the corresponding pixel with the marker value "1";

[0045] Step 2.2: Continue to traverse each pixel in the reference image from left to right and from top to bottom, and let the coordinates of the currently traversed pixel be , take the coordinates as , , , The four adjacent pixels of

[0046] Define the label values ​​of the current pixel and the four adjacent pixels as , , , , ;

[0047] Define variable g and initialize g=2;

[0048] Step 2.3: Mark the value of the current pixel traversed in the e-th frame reference image Perform the following processing to obtain the intermediate image of the e-th frame after traversal:

[0049] (1) When and , , , When they are not all "0", the non-zero pixel label values ​​of the four adjacent pixels are assigned to ;

[0050] (2) When but , , , When all are zero, , and assign g+1 to g;

[0051] (3) When When, no processing;

[0052] Step 2.4: Traverse each pixel in the middle image of the e-th frame from right to left and from top to bottom, and perform the processing of step 2.3 to obtain the marked image of the e-th frame;

[0053] Step 2.5: Traverse each pixel in the e-th frame of the marked image, and store the coordinates of the pixels with the same marking value in the corresponding set. Let s be any marking value in the pixels of the marked image, then the coordinates of the pixels with marking value s are stored in the set Thus, we get a set containing all the coordinate information of the area covered by the sth light spot. ; Let t be The number of coordinates contained in Any i-th coordinate in is ( ), i∈[1, t];

[0054] Step 3: Calculate the centroid;

[0055] Step 3.1: Solve for the center of mass position:

[0056] definition The coordinates of the standard image in the e-th frame are ( ) pixel grayscale value, and use equation (4) to solve the centroid coordinates of the sth light spot ( , ):

[0057] ; (4)

[0058] Step 3.2: Calculate the centroid grayscale value:

[0059] The centroid coordinates of the sth light spot ( ) is rounded down to get the coordinates ( ), find the pixel coordinates ( )、( )、( )、( +1) and takes it as the centroid gray value ; Then, the centroid coordinates and centroid grayscales of the light spots corresponding to various marking values ​​in the e-th frame marking image constitute a set A;

[0060] Step 3.3: Front and rear centroid matching:

[0061] Step 3.3.1: After the probe is subjected to the required test, a set B consisting of centroid coordinates and centroid grayscale values ​​corresponding to various marking values ​​in the e-th frame marking image is obtained according to the process of steps 1.1 to 3.2;

[0062] Pick A, B; among them, is the pth element in the set A, is the centroid coordinate of the pth spot, is the gray value of the pth light spot; is the qth element in set B, is the centroid coordinate of the qth light spot, is the gray value of the qth light spot; 1≤p≤u, u is the number of elements in A; 1≤q≤v, v is the number of elements in B;

[0063] Step 3.3.2: Initialize p=1;

[0064] Step 3.3.3: Use formula (5) to calculate and Straight line distance , thus obtaining The straight-line distance set C to the coordinates of all elements in set B is:

[0065] (5)

[0066] Step 3.3.4: Find the minimum straight-line distance from the straight-line distance set C The subscript of the corresponding centroid coordinate in set B , then the first The gray value change rate of each spot is ( - ) / ×100%;

[0067] In step 3.3.5, after assigning p+1 to p, return to step 3.3.3 and execute sequentially until p>u, thereby obtaining the absolute displacement value and gray value change rate of all light spots before and after the probe test.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] 1. The aperture of the present invention is provided with a pinhole array, which enables the formation of several spot images with only one collimator, making the overall volume of the light source smaller and meeting the use scenario of a small-caliber probe. In addition, a polarizing film is provided on one side of the aperture, which can be arbitrarily cut according to the required size without limiting the structural design. At the same time, only one collimator is used, which greatly reduces the workload of optical debugging and saves debugging time.

[0070] 2. The collimator of the present invention has a real-time focusing function, which can not only compensate for the error in the actual structure, but also can be adjusted during the test process, which is convenient for debugging and use.

[0071] 3. The centroid solution method of the present invention combines the idea of ​​connected domain detection, and can simultaneously detect multiple light spots at different positions. There is no need to predict the number of light spots and their approximate positions in advance to divide the windows. It can flexibly deal with scenes where the positions of light spots change significantly, thereby improving the efficiency of data processing and reducing possible errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a structural diagram of a parallel light source of the present invention;

[0073] Figure 2 This is a target and pinhole array diagram of the present invention;

[0074] Numbers in the figure: 1-fan, 2-screw, 3-heat sink, 4-LED board, 5-LED board fixing seat, 6-aperture, 7-lens barrel, 8-collimating lens, 9-adapter, 10-light shield, 11-first pressure ring, 12-locking nut, 13-second pressure ring, 14-diffuser, 61-pinhole array. DETAILED DESCRIPTION

[0075] In this embodiment, a LED multi-spectral polarized parallel light source, such as Figure 1 As shown, it includes: a light emitting module, a shaping and modulation module, a collimation module and a fixing structure;

[0076] The light emitting module comprises: a cooling fan 1, a heat sink 3, and an LED board 4;

[0077] The shaping and modulation module includes: an aperture 6, a diffuser 14, and a second pressure ring 13;

[0078] The collimation module comprises: a lens barrel 7, a collimating lens 8, and a first pressure ring 11;

[0079] The fixing structure includes: a fastening screw 2, an LED board fixing seat 5, an adapter 9, and a locking nut 12;

[0080] A heat sink 3 is provided on the back of the LED board 4 and is bonded to the LED board 4 by heat dissipation grease;

[0081] The LED board 4 is fixed on the housing through the LED board fixing seat 5; a cooling fan 1 facing the heat sink 3 is provided above the housing through the fastening screws 2; a lens barrel 7 is inserted below the housing and fixed by a locking nut 12;

[0082] A diaphragm 6 is provided on the housing above the inner part of the lens barrel 7. Figure 2 As shown, the aperture 6 is provided with a pinhole array 61. In specific implementation, the arrangement of the pinhole array can be designed according to the actual debugging situation. A polarizing film is pasted on the surface of the aperture 6, and a diffuser 14 is provided on the polarizing film and fixed by a second pressing ring 13;

[0083] A collimating lens 8 is provided at the lower part of the lens barrel 7 and fixed to the bottom of the lens barrel 7 through a first pressing ring 11. The bottom of the lens barrel 7 is connected to a light shield 10 through an adapter 9. The light shield 10 is provided between the light source and the probe.

[0084] In a specific implementation, the number of LED lights of each band on the LED board 4 is calculated according to the following steps:

[0085] Step 1: Calculate the response brightness of the LED light in each band;

[0086] Step 1.1: Use a spectrometer to collect the brightness response value of the LED light within the bandwidth of any specified band, and calculate the average brightness of the specified band within the bandwidth ;

[0087] Step 1.2: Calculate the sum of the brightness of all other bands that affect the specified band , thus obtaining the total average brightness of the specified band within the bandwidth + ;

[0088] Step 2: Calculate the actual number of LED lights in each band;

[0089] Step 2.1: Definition The typical entrance pupil brightness of the probe in the specified band and calculate the ratio value / ;

[0090] Step 2.2: According to the total number of LED lights on the LED board 4, set the reference number of LED lights in each band to , calculate the theoretical number of LED lights in a specified band , thus After rounding up or down, the actual number of LED lights in the specified band is obtained. .

[0091] The selected LED lamps are welded to the front side of the LED board 4, and the LED lamps of the same wavelength band are connected in series; each LED lamp of each wavelength band after the series connection is driven by one of the multiple-channel LED light source constant current drivers.

[0092] Each time the light source and the light shield are connected, the thread structure should be tightened to ensure the repeatability of the position. After the connection, the LED of the required wavelength band is lit, and the target light beam is emitted from the LED board 4 of the light-emitting module, and passes through the diffuser 14, the aperture 6 and the polarizing film of the shaping and modulation module in sequence to form several beams of polarized light, and then passes through the lens barrel 7 and the collimating lens 8 of the collimating module in sequence to output several beams of parallel light containing polarization information, which are imaged as multiple light spots on the probe under test.

[0093] Adjust the LED current size and the corresponding integration time of the probe and other parameters so that the brightness of the imaging spot meets the probe test requirements. Record the parameters set at this time and keep them unchanged in subsequent experiments.

[0094] The locking nut 12 and the upper shell are rotated to adjust the focal length of the light source and the imaging position of the light spot. When the imaging is clear and the boundary between the light spot and the background is relatively distinct, the locking nut 12 is fixed and maintained at this state in subsequent use.

[0095] In a specific implementation, an image spot detection of an LED multi-spectral polarized parallel light source is to image multiple spots containing polarization information on the probe and solve the spot centroid and grayscale value, so as to detect the state change of the probe before and after the test. Specifically, it includes the following steps:

[0096] Step 1: Image preprocessing;

[0097] Step 1.1: Use the probe to collect M×N frames of images of the parallel light source, where M represents the number of channels of the probe and N represents the number of acquisitions; take N frames of images of the same band and the same polarization direction in the M×N frames of images as a group of images, thereby obtaining M groups of images;

[0098] Take the grayscale average of each pixel of the M-1 groups of images except the background channel to obtain one frame of initial bright field image for each group, thereby obtaining M-1 frames of initial bright field images;

[0099] Step 1.2: Turn off the parallel light source and build a dark room. In a dark environment, use the probe to collect N frames of background images at the integration time corresponding to the M channel according to the process of step 1.1, and calculate M-1 frames of dark field images corresponding to the M-1 groups of background images except the background channel;

[0100] Step 1.3: Subtract the grayscale value of the corresponding pixel of the dark field image of the e-th frame from the initial bright field image of the e-th frame one by one, and obtain the original image of the e-th frame, thereby obtaining the original image of the M-1 frame; ;

[0101] Step 1.4: Replace the gray value of the bad pixel on the original image of the e-th frame with the median of the gray values ​​of the surrounding pixels, thereby obtaining the standard image of the e-th frame;

[0102] Step 1.5: Traverse each pixel in the standard image of the e-th frame and set the pixel above the selected threshold Pixels below the selected threshold are retained The pixels of are replaced by "0", thereby obtaining the reference image after thresholding of the e-th frame;

[0103] Step 2: Spot area detection;

[0104] Step 2.1: Traverse each pixel in the thresholded reference image of the e-th frame from left to right and from top to bottom. When the first non-zero value is detected, replace the gray value of the corresponding pixel with the marker value "1";

[0105] Step 2.2: Continue to traverse each pixel in the reference image from left to right and from top to bottom, and let the coordinates of the currently traversed pixel be , take the coordinates as , , , The four adjacent pixels of

[0106] Define the label values ​​of the current pixel and the four adjacent pixels as , , , , ;

[0107] Define variable g and initialize g=2;

[0108] Step 2.3: Mark the value of the current pixel traversed in the e-th frame reference image Perform the following processing to obtain the intermediate image of the e-th frame after traversal:

[0109] (4) When and , , , When they are not all "0", the non-zero pixel label values ​​of the four adjacent pixels are assigned to ;

[0110] (5) When but , , , When all are zero, , and assign g+1 to g;

[0111] (6) When When, no processing;

[0112] Step 2.4: Traverse each pixel in the middle image of the e-th frame from right to left and from top to bottom, and perform the processing of step 2.3 to obtain the marked image of the e-th frame;

[0113] Step 2.5: Traverse each pixel in the e-th frame of the marked image, and store the coordinates of the pixels with the same marking value in the corresponding set. Let s be any marking value in the pixels of the marked image, then the coordinates of the pixels with marking value s are stored in the set Thus, we get a set containing all the coordinate information of the area covered by the sth light spot. ; Let t be The number of coordinates contained in Any i-th coordinate in is ( ), i∈[1, t];

[0114] Step 3: Calculate the centroid;

[0115] Step 3.1: Solve for the center of mass position:

[0116] definition The coordinates of the standard image in the e-th frame are ( ) pixel grayscale value, and use equation (4) to solve the centroid coordinates of the sth light spot ( , ):

[0117] ; (4)

[0118] Step 3.2: Calculate the centroid grayscale value:

[0119] The centroid coordinates of the sth light spot ( ) is rounded down to get the coordinates ( ), find the pixel coordinates ( )、( )、( )、( +1) and takes it as the centroid gray value ; Then, the centroid coordinates and centroid grayscales of the light spots corresponding to various marking values ​​in the e-th frame marking image constitute a set A;

[0120] Step 3.3: Front and rear centroid matching:

[0121] Step 3.3.1: After the probe is subjected to the required test, a set B consisting of the centroid coordinates and centroid grayscale values ​​corresponding to various marking values ​​in the e-th frame marking image is obtained according to the process of step 1.1 to step 3.2;

[0122] Pick A, B; among them, is the pth element in the set A, is the centroid coordinate of the pth spot, is the gray value of the pth light spot; is the qth element in set B, is the centroid coordinate of the qth light spot, is the gray value of the qth light spot; 1≤p≤u, u is the number of elements in A; 1≤q≤v, v is the number of elements in B;

[0123] Step 3.3.2: Initialize p=1;

[0124] Step 3.3.3: Use formula (5) to calculate and Straight line distance , thus obtaining The straight-line distance set C to the coordinates of all elements in set B is:

[0125] (5)

[0126] Step 3.3.4: Find the minimum straight-line distance from the straight-line distance set C The subscript of the corresponding centroid coordinate in set B , then the first The gray value change rate of each spot is ( - ) / ×100%;

[0127] In step 3.3.5, after assigning p+1 to p, return to step 3.3.3 and execute sequentially until p>u, thereby obtaining the absolute displacement value and gray value change rate of all light spots before and after the probe test.

Claims

1. An LED multi-spectrum polarized parallel light source, characterized in that: include: Light-emitting module, shaping and modulation module, collimation module and fixing structure; The light-emitting module comprises: a cooling fan (1), a cooling block (3), and an LED board (4); The shaping and modulation module comprises: an aperture (6), a diffuser (14), and a second pressure ring (13); The collimating module comprises: a lens barrel (7), a collimating lens (8), and a first pressure ring (11); The fixing structure comprises: a fastening screw (2), an LED board fixing seat (5), an adapter (9), and a locking nut (12); A plurality of LED lamps are welded on the front of the LED board (4), and the LED lamps of the same wavelength band are connected in series; each LED lamp of each wavelength band connected in series is driven by one channel of a multi-channel LED light source constant current driver; The heat sink (3) is provided on the back of the LED board (4) and is bonded to the LED board (4) via heat dissipation grease; The LED board (4) is fixed to the housing via the LED board fixing seat (5); a cooling fan (1) facing the cooling block (3) is provided above the housing via a fixing screw (2); the lens barrel (7) is inserted below the housing and fixed via a locking nut (12); An aperture (6) is provided on the housing above the interior of the lens barrel (7); a polarizing film is adhered to the surface of the aperture (6); a diffuser (14) is provided on the polarizing film and fixed by a second pressing ring (13); and a pinhole array (61) is provided on the aperture (6); The collimating lens (8) is provided at the lower part of the lens barrel (7) and is fixed to the bottom of the lens barrel (7) via a first pressing ring (11); the bottom of the lens barrel (7) is connected to a light shield (10) via an adapter (9); the light shield (10) is provided between the parallel light source and the probe; The target light beam is emitted from the LED board (4) of the light-emitting module, and passes through the diffuser (14), the aperture (6) and the polarization film of the shaping and modulation module in sequence to form a plurality of polarized light beams, and then passes through the lens barrel (7) and the collimating lens (8) of the collimating module in sequence to output a plurality of parallel light beams containing polarization information.

2. The LED multi-spectrum polarized parallel light source according to claim 1, characterized in that: The number of LED lights of each wavelength band on the LED board (4) is calculated according to the following steps: Step 1: Calculate the response brightness of the LED light in each band; Step 1.1: Use a spectrometer to collect the brightness response value of the LED light within the bandwidth of any specified band, and calculate the average brightness of the specified band within the bandwidth ; Step 1.2: Calculate the sum of the brightness of all other bands that affect the specified band , thereby obtaining the total average brightness of the specified band within the bandwidth range + ; Step 2: Calculate the actual number of LED lights in each band; Step 2.1: Definition is the typical entrance pupil brightness of the probe in the specified band, and calculates the proportional value / ; Step 2.2: According to the total number of LED lamps on the LED board (4), the reference number of LED lamps in each band is set to , calculate the theoretical number of LED lights in the specified band , thus After rounding up or down, the actual number of LED lights in the specified band is obtained. .

3. A method for detecting the light spot of a LED multi-spectral polarized parallel light source, characterized in that: The method comprises utilizing the parallel light source as claimed in claim 1 and comprising the following steps: Step 1: Image preprocessing; Step 1.1: using the probe to collect M×N frames of images of the parallel light source, wherein M represents the number of channels of the probe, and N represents the number of acquisitions; taking N frames of images of the same band and the same polarization direction in the M×N frames of images as a group of images, thereby obtaining M groups of images; Take the grayscale average of each pixel of the M-1 groups of images except the background channel to obtain one frame of initial bright field image for each group, thereby obtaining M-1 frames of initial bright field images; Step 1.2: Turn off the parallel light source, build a dark room, and use the probe to collect N frames of background images at the integration time corresponding to the M channel in a dark environment according to the process of step 1.1, and calculate M-1 frames of dark field images corresponding to M-1 groups of background images except the background channel; Step 1.3: Subtract the grayscale value of the corresponding pixel of the dark field image of the e-th frame from the initial bright field image of the e-th frame one by one, and obtain the original image of the e-th frame, thereby obtaining the original image of the M-1 frame; ; Step 1.4: Replace the gray value of the bad pixel on the original image of the e-th frame with the median of the gray values ​​of the surrounding pixels, thereby obtaining the standard image of the e-th frame; Step 1.5: Traverse each pixel in the standard image of the e-th frame and set the pixel above the selected threshold Pixels below the selected threshold are retained The pixels of are replaced by "0", thereby obtaining the reference image after thresholding of the e-th frame; Step 2: Spot area detection; Step 2.1: Traverse each pixel in the thresholded reference image of the e-th frame from left to right and from top to bottom. When the first non-zero value is detected, replace the gray value of the corresponding pixel with the marker value "1"; Step 2.2: Continue to traverse each pixel in the reference image from left to right and from top to bottom, and let the coordinates of the currently traversed pixel be , take the coordinates as , , , The four adjacent pixels of Define the label values ​​of the current pixel and the four adjacent pixels as , , , , ; Define variable g and initialize g=2; Step 2.3: Mark the value of the current pixel traversed in the reference image of the e-th frame Perform the following processing to obtain the intermediate image of the e-th frame after traversal: (1) When and , , , When they are not all "0", the non-zero pixel label values ​​of the four adjacent pixels are assigned to ; (2) When but , , , When all are zero, , and assign g+1 to g; (3) When When, no processing; Step 2.4: Traverse each pixel in the middle image of the e-th frame from right to left and from top to bottom, and perform the processing of step 2.3 to obtain the marked image of the e-th frame; Step 2.5: Traverse each pixel in the e-th frame of the marked image, and store the coordinates of the pixels with the same marking value in the corresponding set. Let s be any marking value in the pixels of the marked image, then the coordinates of the pixels with marking value s are stored in the set Thus, we get a set containing all the coordinate information of the area covered by the sth light spot. ; Let t be The number of coordinates contained in Any i-th coordinate in is ( ), i∈[1, t]; Step 3: Calculate the centroid; Step 3.1: Solve for the center of mass position: definition The coordinates of the standard image in the e-th frame are ( ) pixel grayscale value, and use equation (4) to solve the centroid coordinates of the sth light spot ( , ): ； (4) Step 3.2: Calculate the centroid grayscale value: The centroid coordinates of the sth light spot ( ) is rounded down to get the coordinates ( ), find the pixel coordinates ( )、( )、( )、( +1) and takes it as the centroid gray value ; Then, the centroid coordinates and centroid grayscales of the light spots corresponding to various marking values ​​in the e-th frame marking image constitute a set A; Step 3.3: Front and rear centroid matching: Step 3.3.1: After the probe is subjected to the required test, a set B consisting of centroid coordinates and centroid grayscale values ​​corresponding to various marking values ​​in the e-th frame marking image is obtained according to the process of steps 1.1 to 3.2; Pick A, B; among them, is the pth element in the set A, is the centroid coordinate of the pth spot, is the gray value of the pth light spot; is the qth element in set B, is the centroid coordinate of the qth light spot, is the gray value of the qth light spot; 1≤p≤u, u is the number of elements in A; 1≤q≤v, v is the number of elements in B; Step 3.3.2: Initialize p=1; Step 3.3.3: Use formula (5) to calculate and Straight line distance , thus obtaining The straight-line distance set C to the coordinates of all elements in set B is: (5) Step 3.3.4: Find the minimum straight-line distance from the straight-line distance set C The subscript of the corresponding centroid coordinate in set B , then the first The gray value change rate of each spot is ( - ) / ×100%; In step 3.3.5, after assigning p+1 to p, return to step 3.3.3 and execute sequentially until p>u, thereby obtaining the absolute displacement value and gray value change rate of all light spots before and after the probe test.

Images