A dynamic four-quadrant underwater high-rate guidance method

Abstract

The application discloses a dynamic four-quadrant underwater high-speed guidance method, solves the problem that existing binocular guidance method and four-quadrant guidance method cannot simultaneously meet the requirement of large angle and high frequency, and cannot effectively realize reliable guidance to high-speed running targets. N The application determines the working time sequence by stroboscopic setting of one light source, obtains +1 frame image based on the working time sequence, determines the light source sequence number corresponding to other frame images through the dark image sequence number in the frame image, calculates the row sum and column sum of each frame bright image in the frame bright image, selects the rectangular spot area according to the preset threshold value, carries out automatic exposure according to the row sum in the selected area, divides the selected area into four quadrants, merges the pixel gray scale values in the four quadrants, and calculates the angle of each light source relative to the detector. N N N N According to the angle, sequence number and relative installation position of the multiple light sources, the angle and position of the light source center point relative to the detector are obtained through geometric relationship.​​​

Description

TECHNICAL FIELD

[0001] The present application relates to underwater guidance methods, in particular to a dynamic four-quadrant underwater high-speed guidance method. BACKGROUND

[0002] In the field of underwater target guidance, the existing technical means have significant limitations, which are difficult to meet the growing complex application requirements. Currently, binocular guidance method and four-quadrant guidance method are relatively common underwater guidance methods.

[0003] The binocular guidance method uses a binocular camera and at least three groups of continuous light sources for underwater target guidance. The binocular camera is used to shoot the cooperative light source, and then the angle and position of the binocular camera relative to the light source are calculated according to the position of the light source in the image. The camera pixel usually has high resolution, and a large field of view lens can achieve a large angle measurement of more than 60°, but it will cause the image recognition algorithm to have a large amount of calculation and a relatively low output frequency, generally about 10Hz.

[0004] The four-quadrant guidance method uses a four-quadrant detector and at least three groups of stroboscopic light sources. The cooperative light source is received by the four-quadrant detector, and then the angle of each light source is calculated according to the intensity ratio of the light spot in the four quadrants, and the angle and position of the light source relative to the detector are calculated in combination with the angles of multiple light sources. Since the four-quadrant detector has a high working frequency, and the four-quadrant algorithm has a small amount of calculation, the output frequency can reach a high frequency of KHz. However, in the four-quadrant detection, the light spot covers four quadrants at the same time and does not exceed the boundary, so the angle measurement cannot be performed when the angle of the light source is too large, and the light spot cannot cover the four quadrants, so the angle measurement cannot be performed. Therefore, the effective detection angle of the four-quadrant detection is very small, only about 10°, and it is difficult to achieve large angle measurement.

[0005] With the increasing complexity of underwater operating environment and the increasing demand for reliable guidance of high-speed running targets, such as precise guidance and recovery of underwater high-speed vehicles, rapid tracking and interception of high-speed running underwater targets, and other practical application scenarios, the existing binocular guidance method and four-quadrant guidance method cannot simultaneously meet the requirements of large angle and high frequency (KHz) work, and it is difficult to effectively realize reliable guidance of high-speed running targets. A new technical solution is urgently needed to solve these problems to meet the actual needs of underwater guidance applications. SUMMARY

[0006] In order to solve the technical problems that the existing binocular guidance method and four-quadrant guidance method cannot simultaneously meet the requirements of large angle and high frequency (KHz) work, resulting in difficulty in effectively realizing reliable guidance of high-speed running targets, the present application provides a dynamic four-quadrant underwater high-speed guidance method.

[0007] In order to achieve the above purpose, the present application adopts the following technical scheme:

[0008] A dynamic four-quadrant underwater high-speed guidance method, characterized in that it comprises the following steps:

[0009] Step 1, setting light source serial number and flashing period for N light sources on the recovery platform, determining working time sequence and exposure time of the detector arranged on the mobile platform by controlling the flashing of the N light sources;

[0010] Step 2, the detector performs image acquisition based on the working time sequence and exposure time, so that N+1 frames of images can be acquired within the corresponding working time sequence period, wherein N frames are bright images when the N light sources are on, and one frame is a dark image when all light sources are off, each frame of image is numbered, and the corresponding relationship between the image number and the light source serial number is determined according to the position of the dark image;

[0011] Step 3, performing the following steps on the N bright images:

[0012] 3.1, for each frame of image, calculating the sum of the gray values of all pixels in each row and the sum of the gray values of all pixels in each column, and screening to obtain sum of a rows and sum of a columns, so that and ;

[0013] 3.2, dividing the region in each frame of image into four quadrants, accumulating the row sums of the pixel gray values in the first and second quadrants and the third and fourth quadrants, and accumulating the column sums of the pixel gray values in the first and fourth quadrants and the second and third quadrants;

[0014] 3.3, calculating the centroid coordinates of each frame of image according to the accumulated row sums and column sums in step 3.2;

[0015] 3.4, obtaining the deflection angle and pitch angle corresponding to the centroid coordinates of each frame of image, and obtaining the angle of each light source relative to the detector based on the corresponding relationship between the image number and the light source serial number;

[0016] 3.5, calculating the included angle between the straight line OO' and the detection surface and the coordinates of point O' in the detection surface coordinate system within the working time sequence period according to the angle of each light source relative to the detector and the relative positions of the N light sources; wherein point O in the straight line OO' is the center point of the detection surface, and point O' is the geometric center point of the N light sources;

[0017] Step 4, summing the a row sums determined in step 3.1, and averaging the sum of the a row sums in the N bright images within the corresponding working time sequence period to obtain the average value L m ;

[0018] Step 5, based on the average value L mDetermine the exposure time of the next working timing cycle, and return to step 2 to continuously acquire the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the detection surface coordinate system until the dynamic four-quadrant underwater high-speed guidance is completed.

[0019] Further, step 1 comprises:

[0020] Step 1.1, set the light source serial number for the N light sources on the recovery platform, the light source serial number is 1, 2…i…N in turn, N≥3, and set the flashing period of each light source as (2N+2)T, control N light sources to periodically flash respectively, wherein, in the first (2N)T, N light sources flash according to their serial number in turn, only one light source is on each time, and the duration of the on and off is T, in the last 2T, N light sources are all off; the range of T matches the frame frequency of the detector arranged on the moving platform for receiving the light signal;

[0021] Step 1.2, the detector receives the light signal emitted by the light source, and acquires images in three time sequences, the acquisition timing cycle of each time sequence is (2N+2)T, the delay time of the first time sequence is 0, the delay time of the second time sequence is T / 2, and the delay time of the third time sequence is-T / 2, N+1 frames of images are acquired in each time sequence, and the acquisition period of each frame of image is 2T, and the exposure time is E t , wherein the initial exposure time is T;

[0022] Step 1.3, compare all the images acquired in the three time sequences, and select the time sequence corresponding to the image with the maximum brightness as the working time sequence.

[0023] Further, step 2 specifically comprises:

[0024] The detector acquires the light signal based on the working time sequence, the acquisition timing cycle is (2N+2)T, the frame frequency is 1 / (2T), and the exposure time is E t , again continuously acquire the light signal, then N+1 frames of images can be acquired in one working timing cycle, wherein N frames are bright images when N light sources are on, and one frame is a dark image when all light sources are off, and the serial number of N+1 frames of images is P1, P2…P N+1 , through the circular ring sorting method, the serial number of each frame of image corresponding to the light source serial number is determined according to the position of the dark image.

[0025] Further, after step 2 and before step 3, it further comprises the step of threshold filtering N frames of bright images:

[0026] Obtain the maximum pixel gray value L from N frames of bright images, and assign the pixel gray value less than LxG1 in N frames of bright images to 0, wherein G1 is the first preset proportion.

[0027] Further, step 3.1 comprises:

[0028] 3.1.1, record m rows x n columns pixel gray value of each frame image, for each frame image, calculate the sum of all pixel gray value of each row and the sum of all pixel gray value of each column, get m row sums and n column sums;

[0029] 3.1.2, select a row sums greater than the row threshold from the m row sums, and select b column sums greater than the column threshold from the n column sums; the row threshold is H x G2, and the column threshold is Q x G3, wherein H is the maximum value in the m row sums, Q is the maximum value in the n column sums, G2 is the second preset proportion, and G3 is the third preset proportion;

[0030] 3.1.3, if and , directly execute step 3.2;

[0031] if and / or , re-execute step 3.1.2, and correspondingly reduce the row threshold and / or the column threshold by half, re-screen, until and , get new row sums and new column sums, and then execute step 3.2.

[0032] Further, step 3.2 specifically comprises:

[0033] divide the region in each frame image into four quadrants, accumulate the row sums of the pixel gray values in the first quadrant and the second quadrant , accumulate the row sums of the pixel gray values in the third quadrant and the fourth quadrant , accumulate the column sums of the pixel gray values in the first quadrant and the fourth quadrant , and accumulate the column sums of the pixel gray values in the second quadrant and the third quadrant .

[0034] Further, step 3.3 specifically comprises:

[0035] according to the four-quadrant measurement principle, calculate the centroid coordinates (X i , Y i ) of each frame image:

[0036]

[0037] wherein, .

[0038] Further, step 3.4 specifically comprises:

[0039] From the calibrated database of correspondences between the deflection angle and elevation angle of the independent light source relative to the detector and the centroid coordinates of the light spot detected by the detector, the centroid coordinates (X, Y) of each frame of image are obtained. i Y i The corresponding deflection angle ψ i and pitch angle θ i Thus, the angle (ψ) of each light source relative to the detector is obtained. i θ i ).

[0040] Furthermore, step 3.5 specifically includes:

[0041] According to the angle of each light source (ψ) i θ i The relative positions of N light sources and the light source serial numbers corresponding to the corresponding image numbers are used to calculate the angle between the straight line OO′ and the detection surface and the coordinates of point O′ in the detection surface coordinate system through geometric relationships. This allows us to obtain the angle between the straight line OO′ and the detection surface and the coordinates of point O′ in the detection surface coordinate system within the working time period.

[0042] Furthermore, step 5 specifically includes:

[0043] When L min ≤L m ≤L max If the exposure time remains unchanged, return to step 2 to collect data for the next working time cycle, thereby continuously acquiring the angle between the line OO′ and the detection surface, as well as the coordinates of point O′ in the detection surface coordinate system;

[0044] When L m >L max or L m <L min When that happens, the exposure time will be updated to... Return to step 2 to perform the next working time cycle acquisition, thereby continuously acquiring the angle between the line OO′ and the detection surface, as well as the coordinates of point O′ in the detection surface coordinate system, where L max L is the preset upper limit value for the average. min The preset lower limit value;

[0045] Until dynamic four-quadrant underwater high-speed guidance is completed.

[0046] The beneficial effects of this invention are:

[0047] 1. The dynamic four-quadrant underwater high-speed guidance method provided by the application, through flash setting of N light sources to determine the working timing, then based on the working timing, N+1 frames of images are acquired, the light source serial number corresponding to each of the other N frames of images is determined through the dark image serial number, in the N frames of bright images, the row sum and column sum of each frame of bright image are calculated, the light spot area is selected according to the preset threshold, automatic exposure is performed according to the row sum in the selected light spot area, the selected light spot area is divided into four quadrants, the pixel gray value in the four quadrants is further merged to calculate the angle of each light source relative to the detector, and then the angle and position of the center point of the N light sources relative to the detector are obtained from the geometric relationship according to the angle, serial number and relative installation position of the multiple light sources, the method meets the working requirements of large angle and high frequency (KHz) at the same time, and reliable guidance of the high-speed running target is effectively realized.

[0048] 2. The dynamic four-quadrant underwater high-speed guidance method provided by the application can adopt an image detector with more pixels than a four-quadrant detector, and then dynamically select a four-quadrant region according to the light spot size, so that the light spot covers the four quadrants at the same time and does not exceed the boundary, thereby realizing large-angle measurement of more than 60 degrees.

[0049] 3. The dynamic four-quadrant underwater high-speed guidance method provided by the application greatly reduces the data amount by using row sum and column sum signals as original data, and adopts four-quadrant measurement principle when calculating the angle, so that the calculation amount is small, and the output frequency can reach KHz while realizing large-angle measurement. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a structural schematic view of the recovery platform A, the light source, the mobile platform B and the detector in step 1.1 of the embodiment of the dynamic four-quadrant underwater high-speed guidance method of the application;

[0051] Figure 2 is a schematic view of the light source control flash setting in step 1.1 of the embodiment of the application;

[0052] Figure 3 is a schematic view of the acquisition timing cycle and initial exposure time in step 1.2 of the embodiment of the application;

[0053] Figure 4 is a view of the alignment of the three timing and the set acquisition timing cycle in step 1.2 of the embodiment of the application;

[0054] Figure 5 is a schematic view of the selected light spot area in step 3.1.2 of the embodiment of the application;

[0055] Figure 6 is a schematic view of the four quadrants divided in step 3.2 of the embodiment of the application.

[0056] Reference signs:

[0057] 01-control module, 02-detector, 021-optical lens, 022-high-speed image chip, 023-data processing module, 03-light source. DETAILED DESCRIPTION

[0058] The technical solutions of the present application will be described clearly and completely in combination with the drawings and embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0059] The dynamic four-quadrant underwater high-speed guidance method provided by the embodiment of the present application comprises the following steps:

[0060] Step 1, setting light source serial numbers and flashing periods for N light sources 03 on a recovery platform, and determining working time and exposure time of a detector 02 arranged on a mobile platform through flashing of the N light sources 03; specifically comprising:

[0061] Step 1.1, setting light source serial numbers for the N light sources 03 on the recovery platform, the light source serial numbers being 1, 2…i…N in turn, N≥3, and setting a flashing period of each light source 03 as (2N+2)T, and controlling the N light sources 03 to perform periodic flashing respectively, wherein, in the first (2N)T, the N light sources 03 flash according to their serial numbers in turn, only one light source 03 is on each time, and the duration of the on and off is T, in the last 2T, the N light sources 03 are all off; the range of T is matched with the frame frequency of the detector 02 arranged on the mobile platform for receiving light signals, that is, the frame frequency of the detector 02 is 1 / (2T), when the range of T is set, it is ensured that the frame frequency 1 / (2T) is within the frame frequency range that can be set by the detector 02.

[0062] In this embodiment, as shown in Figure 1 , three light sources 03 are arranged on the recovery platform A, the light source serial numbers are 1, 2 and 3 in turn, the flashing period of each light source 03 is set as 8×0.1 ms, and the three light sources 03 are controlled to perform periodic flashing by the control module 01, as shown in Figure 2As shown in the figure, the three light sources 03 in the first 6x0.1 ms are sequentially turned on and off, only one light source 03 is turned on each time, and the duration of turning on and off is 0.1 ms, and in the last 2x0.1 ms, the three light sources 03 are all turned off; the frame frequency of the detector 02 in this embodiment is 1 / (2x0.0001) fps, and the detector 02 specifically comprises an optical lens 021, a high-speed image chip 022 and a data processing module 023, and the field angle of the optical lens 021 is α=60°, which can be larger or smaller in other embodiments.

[0063] Step 1.2, as shown in Figure 3 and Figure 4 The detector 02 receives the light signals emitted by the light source 03 and collects images in three time sequences, and the collection time sequence period of each time sequence is (2N+2)T, the delay time of the first time sequence is 0, the delay time of the second time sequence is T / 2, and the delay time of the third time sequence is -T / 2, N+1 frames of images are collected in each time sequence, and the collection period of each frame of image is 2T, and the exposure time is E t , wherein the initial exposure time is T.

[0064] In this embodiment, the detector 02 receives the light signals emitted by the three light sources 03 and collects images in three time sequences, and the collection time sequence period of each time sequence is 8x0.1 ms, the delay time of the first time sequence is 0, the delay time of the second time sequence is 0.1 / 2 ms, and the delay time of the third time sequence is -0.1 / 2 ms, 4 frames of images are collected in each time sequence, and the collection period of each frame of image is 2x0.1 ms, and the exposure time is E t , wherein the initial exposure time is 0.1 ms, that is, initially, E t =0.1 ms.

[0065] Step 1.3, compare all the images collected in the three time sequences, and select the time sequence corresponding to the image with the maximum brightness as the working time sequence.

[0066] In this embodiment, the effective exposure time is 0.1 ms when the delay time is 0, the effective exposure time is 0.1 / 2 ms when the delay time is 0.1 / 2 ms, and the effective exposure time is 0.1 / 2 ms when the delay time is -0.1 / 2 ms, the effective exposure time is the largest when the delay time is 0, and the brightness is the highest, so the first time sequence is determined as the working time sequence.

[0067] Step 2, the detector 02 collects images based on the working time sequence and the exposure time, with a collection time sequence period of (2N+2)T, a frame frequency of 1 / (2T), and an exposure time of E tAgain, continuous acquisition of light signals, then a working timing cycle can be collected N+1 frame image, wherein N frame is N light source 03 respectively when the bright image, a frame is all light source 03 when the dark image, mark N+1 frame image number in order according to the acquisition sequence is P1, P2…P N+1 , through the circular ring sorting method, according to the dark image position to determine the corresponding relationship between image number and light source number.

[0068] In this embodiment, the probe 02 based on the determination of the working timing, with the acquisition timing cycle of 8x0.1 ms, frame frequency of 1 / (2x0.0001)fps, exposure time E t Again, continuous acquisition of light signals, then a working timing cycle can be collected 4 frame image, then there must be 3 frame respectively as 3 light source 03 when the bright image, 1 frame is all light source 03 when the dark image, mark 4 frame image number in order according to the acquisition sequence is P1, P2…P4, through the circular ring sorting method, according to the dark image position to determine the corresponding relationship between image number and light source number.

[0069] Through the circular ring sorting method, according to the dark image position to determine the corresponding light source number of each frame image number, as follows:

[0070] The collected 4 frame image is sorted in the same circular ring along the clockwise direction in order, to determine the dark image, with the dark image along the clockwise direction of the next image, corresponding to the light source number 1, 2…3, the dark image has no corresponding light source number.

[0071] Threshold filtering for 3 frame bright image: from 3 frame bright image to get the maximum pixel gray value L, the pixel gray value of 3 frame bright image below LxG1 is assigned to 0, wherein G1 is the first preset proportion; In this embodiment, G1=10%.

[0072] Step 3, for 3 frame bright image, the following steps are executed:

[0073] 3.1, for each frame image, the sum of all pixel gray values of each row and the sum of all pixel gray values of each column are calculated respectively, and are screened to obtain The sum of the row and The sum of the column, which requires And ; Specifically, it includes:

[0074] 3.1.1, mark each frame image has m rows x n columns of pixel gray value, for each frame image, the sum of all pixel gray values of each row and the sum of all pixel gray values of each column are calculated respectively, to obtain m row sum and n column sum.

[0075] In this embodiment, m=1000, n=800.

[0076] 3.1.2. Select a sums of rows greater than a row threshold from the sums of m rows, and select b sums of columns greater than a column threshold from the sums of n columns; the row threshold is H×G2, and the column threshold is Q×G3, where H is the maximum value among the sums of m rows, Q is the maximum value among the sums of n columns, G2 is the second preset ratio, and G3 is the third preset ratio.

[0077] In this embodiment, as Figure 5 As shown, the bottom left corner of the image is defined as the origin, the rightward direction is the positive x-axis, and the upward direction is the positive y-axis. The maximum value H among the sums of m rows is multiplied by G2 = 10% as the row threshold. Rows m1-m that are consecutive and higher than the row threshold are selected. a The sum of a rows is used as the column threshold. The maximum value Q among the sums of n columns is multiplied by G3 = 10%. Columns n1-n that are consecutive and exceed the column threshold are selected. b The sum of a rows and b columns is the light spot region, and the size of the region is a rows and b columns.

[0078] 3.1.3, if and Based on The sum of each row and The sum of each column is directly executed in step 3.2.

[0079] like and / or If so, then repeat step 3.1.2, and correspondingly reduce the row threshold and / or column threshold by half (i.e., if only If only the row threshold is reduced by half, then... If so, then only the column threshold will be reduced by half. and If the threshold is halved, then both the row threshold and column threshold will be halved; and the filtering process will be repeated until... and , get new The sum of individual lines and the new The sum of each column, then based on the new... The sum of individual lines and the new The sum of each column is executed in step 3.2.

[0080] 3.2. Extract the data from each frame of the image. The region is divided into four quadrants. The sum of the row sum of the pixel gray values ​​in the first and second quadrants, the row sum of the pixel gray values ​​in the third and fourth quadrants, the column sum of the pixel gray values ​​in the first and fourth quadrants, and the column sum of the pixel gray values ​​in the second and third quadrants are calculated.

[0081] like Figure 6 As shown, each frame of the image contains The region is divided into four quadrants, and the sum of the gray values ​​of the pixels in the first and second quadrants is calculated. The sum of the gray values ​​of the pixels in the third and fourth quadrants. The sum of the column values ​​of the pixel grayscale values ​​in the first and fourth quadrants. The sum of the column values ​​of the pixel grayscale values ​​in the second and third quadrants. .

[0082] 3.3. Based on the sum of rows and columns accumulated in step 3.2, calculate the centroid coordinates of each frame of the image.

[0083] In this embodiment, the centroid coordinates (X, Y, X) of each image frame are calculated based on the four-quadrant measurement principle. i Y i ):

[0084]

[0085] in, , Right now and Take the larger value.

[0086] 3.4 Obtain the deflection angle and elevation angle corresponding to the centroid coordinates of each frame image. Based on the correspondence between the image number and the light source number, obtain the angle of each light source 03 relative to the detector 02.

[0087] In this embodiment, the centroid coordinates (X) of each frame of image are obtained from a database of correspondences between the deflection angle and elevation angle of the independent light source relative to the detector 02 and the centroid coordinates of the light spot detected by the detector 02. i Y i The corresponding deflection angle ψ i and pitch angle θ i Based on the correspondence between image number and light source number, the angle (ψ) of each light source 03 relative to detector 02 is obtained. i θ i ).

[0088] 3.5. Based on the angles of each light source 03 relative to the detector 02 and the relative positions of the N light sources 03, calculate the angle between the line OO′ and the detection surface and the coordinates of point O′ in the detection surface coordinate system. Then, obtain the angle between the line OO′ and the detection surface and the coordinates of point O′ in the detection surface coordinate system within the working time period. Among them, point O in the line OO′ is the center point of the detection surface, point O′ is the geometric center point of the N light sources 03, and the detection surface coordinate system is the detection surface coordinate system of the detector 02 itself.

[0089] In this embodiment, based on the angle (ψ) of each light source 03i , θ i ) and the relative position of the three light sources 03, the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface are calculated through geometric relationship, and then the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface in the working time sequence cycle are obtained.

[0090] Step 4, summing the a rows determined in step 3.1, and averaging the sum of the a rows in the N frames of bright images in the corresponding working time sequence cycle to obtain L m .

[0091] In this embodiment, the a rows determined in step 3.1.3 are summed, and the average value L m is obtained by averaging the sum of the a rows corresponding to the three frames of bright images in the corresponding working time sequence cycle.

[0092] Step 5, when L min ≤L m ≤L max , the exposure time is kept unchanged, and the step 2 is returned to collect the next working time sequence cycle, so as to continuously obtain the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface.

[0093] When L m >L max or L m <L min , the exposure time is updated to , and the step 2 is returned to collect the next working time sequence cycle, so as to continuously obtain the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface, wherein L max is a preset upper limit value, and L min is a preset lower limit value.

[0094] In this embodiment, the preset upper limit value L max =1500, and the preset lower limit value L min =150; when the updated exposure time is greater than T, the step 2 is returned to continue to execute with T as the exposure time; and when the updated exposure time is less than the minimum value supported by the detector, the step 2 is returned to continue to execute with the minimum value supported by the detector as the exposure time.

[0095] Until the dynamic four-quadrant underwater high-rate guidance is completed, the field of view angle of 60° and the guidance pose information output of 1 / (8*0.0001)=1250Hz are realized.

[0096] The above merely illustrates the specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any change or replacement within the technical scope disclosed by the present application should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A dynamic four-quadrant underwater high-rate guidance method, characterized in that, The method comprises the following steps: Step 1, setting light source serial numbers and flashing periods for N light sources on a recovery platform, and determining working time sequence and exposure time of a detector arranged on a mobile platform by controlling flashing of the N light sources; Step 2, the detector performs image acquisition based on the working time sequence and the exposure time, so that N+1 frames of images can be acquired within a corresponding working time sequence period, wherein N frames are bright images when the N light sources are bright respectively, and one frame is a dark image when all the light sources are off, each frame of image is numbered, and the corresponding relationship between the image number and the light source serial number is determined according to the position of the dark image; Step 3, performing the following steps on the N frames of bright images: Step 3.1, for each frame of image, calculate the sum of all pixel gray values of each row and the sum of all pixel gray values of each column respectively, and screen to obtain the sum of row and the sum of column, so that and ; Step 3.2, divide the region in each frame image into four quadrants, accumulate the row sum of the pixel gray scale values of the 1st and 2nd quadrants and the 3rd and 4th quadrants, and accumulate the column sum of the pixel gray scale values of the 1st and 4th quadrants and the 2nd and 3rd quadrants; Step 3.2, divide the region in each frame image into four quadrants, accumulate the row sum of the pixel gray scale values of the 1st and 2nd quadrants and the 3rd and 4th quadrants, and accumulate the column sum of the pixel gray scale values of the 1st and 4th quadrants and the 2nd and 3rd quadrants; Step 3.3, calculating the centroid coordinates of each frame of image according to the accumulated row sum and column sum of step 3.2; Step 3.4, obtaining the deflection angle and the pitch angle corresponding to the centroid coordinates of each frame of image, obtaining the angle of each light source relative to the detector based on the corresponding relationship between the image number and the light source serial number; Step 3.5, calculating the included angle between the straight line OO' and the detection surface and the coordinates of point O' in the detection surface coordinate system within the working time sequence period according to the angle of each light source relative to the detector and the relative position of the N light sources; wherein point O in the straight line OO' is the center point of the detection surface, and point O' is the geometric center point of the N light sources; Step 4, sum the sums of the a rows determined in Step 3.1 and average the sums of the a rows in the N frame bright image for the corresponding work timing period L m ; Step 5, average value L-based m The exposure time of the next working time cycle is determined, and the step 2 is returned to continuously acquire the included angle between the straight line OO' and the detection surface and the coordinates of the point O' in the detection surface coordinate system until the dynamic four-quadrant underwater high-rate guidance is completed.

2. The dynamic four-quadrant underwater high-rate guidance method according to claim 1, characterized in that, Step 1 comprises: Step 1.1, setting light source serial numbers for N light sources on a recovery platform, the light source serial numbers are 1, 2…i…N in turn, N≥3, and setting the flashing period of each light source as (2N+2)T, controlling the N light sources to perform periodic flashing respectively, wherein for the first (2N)T, the N light sources flash according to their serial numbers in turn, only one light source is bright each time, and the duration of bright and off is T, for the last 2T, the N light sources are all off; the range of T is matched with the frame frequency of a detector arranged on a mobile platform for receiving light signals; Step 1.2, the detector receives the light signal emitted by the light source and carries out image acquisition in three time sequences, the acquisition time sequence period of each time sequence is (2N+2)T, the delay time of the first time sequence is 0, the delay time of the second time sequence is T / 2, the delay time of the third time sequence is -T / 2, N+1 frames of images are acquired in each time sequence, and the acquisition period of each frame of image is 2T, and the exposure time is E t wherein the initial exposure time is T; Step 1.3, comparing all the images acquired in the three time sequences, and selecting the time sequence corresponding to the image with the maximum brightness as the working time sequence.

3. The dynamic four-quadrant underwater high-rate guidance method according to claim 2, characterized in that, Step 2 specifically comprises: The detector acquires a timing cycle of (2N+2)T based on the working timing, a frame frequency of 1 / (2T), and an exposure time of E t Continuously acquiring the light signal again, N+1 frames of images can be acquired in one working timing cycle, wherein N frames are bright images when N light sources are bright, and one frame is a dark image when all light sources are off, and the serial numbers of the N+1 frames of images are P1, P2…P N+1 The light source serial number corresponding to the serial number of each frame of image is determined according to the dark image position through a circular ring sorting method.

4. The dynamic four-quadrant underwater high-rate guidance method according to claim 1 or 2 or 3, characterized in that, Before step 3, the method further comprises the step of performing threshold filtering on the N frames of bright images: Obtaining the maximum pixel gray value L from the N frames of bright images, and assigning the pixel gray value less than L×G1 in the N frames of bright images as 0, wherein G1 is a first preset proportion.

5. The dynamic four-quadrant underwater high-rate guidance method according to claim 4, characterized in that, Step 3.1 comprises: Step 3.1.1, recording m rows×n columns of pixel gray values of each frame of image, calculating the sum of all pixel gray values of each row and the sum of all pixel gray values of each column for each frame of image, and obtaining m row sums and n column sums; Step 3.1.2, selecting a row sum greater than a row threshold value from the m row sums, and selecting b column sums greater than a column threshold value from the n column sums; the row threshold value is H×G2, and the column threshold value is Q×G3, wherein H is the maximum value in the m row sums, Q is the maximum value in the n column sums, G2 is a second preset proportion, and G3 is a third preset proportion; Step 3.1.3, if and then directly proceed to Step 3.2; If and / or then re-perform step 3.1.2 and correspondingly reduce the row threshold and / or the column threshold by half, re-screen, until and resulting in new sums of rows and new sums of columns, after which step 3.2 is performed again.

6. The dynamic four-quadrant underwater high-rate guidance method according to claim 5, characterized in that, Step 3.2 specifically comprises: In each frame of the image The region is divided into four quadrants, and the sum of the gray values ​​of the pixels in the first and second quadrants is calculated. The sum of the gray values ​​of the pixels in the third and fourth quadrants. The sum of the column values ​​of the pixel grayscale values ​​in the first and fourth quadrants. The sum of the gray values ​​of the pixels in the second and third quadrants. .

7. The dynamic four-quadrant underwater high-rate guidance method according to claim 6, characterized in that, Step 3.3 specifically comprises: According to the four-quadrant measurement principle, the centroid coordinates (X i , Y i ) of each frame of image are calculated: ； wherein .

8. The dynamic four-quadrant underwater high-rate guidance method according to claim 7, characterized in that, Step 3.4 specifically comprises: From the calibrated database of the corresponding relationship between the yaw angle, the pitch angle of the independent light source relative to the detector and the spot centroid coordinates detected by the detector, the centroid coordinates (X i , Y i ) of each frame of image are obtained, and the corresponding yaw angle ψ i and the pitch angle θ i are obtained, so as to obtain the angle (ψ i , θ i ) of each light source relative to the detector.

9. The dynamic four-quadrant underwater high-rate guidance method according to claim 8, characterized in that, Step 3.5 specifically comprises: According to the angle (ψ i , θ i ) of each light source, the relative position of N light sources and the light source serial number corresponding to the corresponding image number, the included angle between the straight line OO' and the detection surface and the coordinates of the point O' in the detection surface coordinate system are calculated through geometric relationship, and then the included angle between the straight line OO' and the detection surface and the coordinates of the point O' in the detection surface coordinate system in the working timing cycle are obtained.

10. The dynamic four-quadrant underwater high-rate guidance method according to claim 9, characterized in that, Step 5 specifically comprises: When L min ≤ L m ≤ L max , the exposure time is kept unchanged, and the process returns to step 2 to acquire the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface in the next working timing cycle, so as to continuously acquire the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the coordinate system of the detection surface. When L m > L max or L m <L min , the exposure time is updated to , and the process returns to step 2 to acquire the next working cycle, thereby continuously acquiring the angle between the straight line OO' and the detection surface and the coordinates of the point O' in the detection surface coordinate system, wherein L max is a preset upper limit value, and L min is a preset lower limit value. Until the dynamic four-quadrant underwater high-rate guidance is completed.

Images