Zoom lens magnification following method

Abstract

The application provides a zoom lens zooming following method, which can ensure that the current scene is clear during zooming and the zooming is smooth without the phenomenon of jamming. The application comprises the following steps: subtracting a basic focusing curve and an object distance focusing curve to obtain a plurality of difference curves of the basic focusing curve and the object distance focusing curve, and fitting the difference curves into high-order polynomials; obtaining the positions of a Focus motor corresponding to the most clear positions of each different object distance under the zoom motor magnification; obtaining the positions of the Focus motor corresponding to the object distance focusing curves of the two object distances most adjacent to the point value according to the point value of the most clear point of the current scene image; and obtaining the zooming tracking curve under the current focusing object distance, wherein each zoom motor position corresponds to a Focus motor position during the zooming process, so that the zoom lens can always keep the clear image within the depth of field range during the zooming process.

Description

TECHNICAL FIELD

[0001] The present application relates to a zoom lens magnification following method, used in the field of video cameras. BACKGROUND

[0002] In recent years, online video conferencing and video education are increasingly popular with the public, and the industry is developing rapidly, greatly promoting remote video communication and education dissemination.

[0003] Online video conferencing is mainly in an indoor environment, with a small conference room length of about 3-4 meters, a medium conference room length of about 6-8 meters, and a large conference room length of about 10-15 meters. In such an environment, it is difficult to cover the details with a fixed-focus lens, so a high-magnification zoom camera is usually used to zoom in on the details in real time, including close-up shots of a conference participant and tracking of relevant conference participants.

[0004] During a video conference, a high-magnification zoom camera is usually placed in front. When no one is speaking, the high-magnification zoom camera is stretched to the maximum field of view, i.e. the zoom motor is at the minimum magnification position, which covers the entire conference scene. If someone speaks, the camera head will adjust the lens position to center the speaker in the field of view, and then pull the zoom motor to the high magnification position to zoom in on the speaker and make them the focus. When the speaker finishes speaking, the zoom motor is pulled again to the low magnification position to restore the large field of view.

[0005] In addition, in the context of video education, such as in the current campus education environment, high-definition high-magnification zoom cameras are installed on the front and back walls of most classrooms to monitor, track, and record the teaching environment in real time. When the teacher walks around the classroom, the camera will zoom in to ensure the teacher's position in the close-up shot. When a student stands up to answer a question during class, the close-up shot will zoom in to ensure that the student is the focus of the scene.

[0006] Whether it's online video conferencing or video education, the use of high-magnification zoom cameras will inevitably involve zoom lens magnification stretching. During the magnification stretching process, it is necessary to ensure that the current scene is clear during the stretching process and that the zoom is smooth during the stretching process without any lag, giving a soft and smooth change in the sense of sight. SUMMARY

[0007] The purpose of this invention is to overcome the above-mentioned shortcomings in the prior art and to provide a zoom lens zoom-following method with a reasonable structural design, which can ensure that the current scene is clear during the zooming process and that the zooming is smooth without any stuttering.

[0008] The technical solution adopted by the present invention to solve the above problems is: a zoom lens zoom-following method, characterized by including the following steps:

[0009] Step 1: Using a calibration light box, align the zoom lens with the light port of the calibration light box, drive the zoom motor, and move it from the near focal end to the far focal end, point by point, and search for the corresponding focus motor position under each zoom motor position, record it, and store it as the basic focus curve.

[0010] Step 2: Use a teleconverter and align the zoom lens with the teleconverter; set multiple teleconverters to simulate object distances; drive the zoom motors, moving from the near focal end to the far focal end, point by point, and search for the corresponding focus motor position under each zoom motor position, record it, and store it as multiple object distance focusing curves.

[0011] Step 3: Using the basic focusing curve from Step 1, calculate the difference between it and the multiple focusing curves with different object distances obtained in Step 2 to obtain multiple difference curves between the basic focusing curve and the object distance focusing curve.

[0012] Step 4: Using curve fitting, fit the difference curve obtained in Step 3 into a higher-order polynomial.

[0013] Step 5: Based on the hill-climbing algorithm, search for the Focus motor position value corresponding to the clearest point in the current scene image at the current magnification;

[0014] Step 6: Using the current zoom motor position, and the higher-order polynomials of different object distances from Step 4, obtain the Focus motor position at the clearest position for each different object distance corresponding to the zoom motor magnification.

[0015] Based on the Focus motor position value corresponding to the clearest point in the current scene image obtained in step five, the Focus motor positions corresponding to the two object distance focusing curves closest to this position value are obtained. That is, the upper and lower limits FP of the Focus motor position corresponding to the two closest object distance focusing curves at the current magnification are obtained. UP FP DOWN ;

[0016] Step 7: Convert the upper and lower position limits FP obtained in Step 6. UP FP DOWNInterval subdivision is carried out, and the current Focus motor position is obtained after subdivision, the nearest point position FP' of the Focus motor in the interval is obtained, and two object distance focusing curves in which FP UP and FP DOWN are located are taken as references, the proportion of FP' in FP up -FP down is calculated, and the Focus position {FP1, FP2, FP3, …, FP n} is calculated from the near focus end to the far focus end point by point, n is the total point number of the zoom motor interval, and the zoom tracking curve L = {FP1, FP2, FP3, …, FP n} under the current focusing object distance is obtained, and each zoom motor position corresponds to a focus motor position in the zooming process, so that the zoom lens can always maintain a clear image in the depth of field range during zooming.

[0017] In step two of the present application, the simulated object distances of the extender lens are set to 1m, 2m, 3m, 4m, 5m, 6m, 7m, 8m, 9m, 10m, 11m, 12m, 13m, 14m, 15m, 16m, 17m, 18m, 19m and 20m, respectively.

[0018] In step four of the present application, the high-order polynomial is an 8-order polynomial, and the corresponding polynomial analytical expression is recorded.

[0019] delta (x) = ax 8 + bx 7 + cx 6 + dx 5 + ex 4 + fx 3 + gx 2 + hx 1 + i (a not equal to 0) ;

[0020] Where [a, b, c, d, e, f, g, h, i] are polynomial coefficients, x is the point number of the zoom motor, and delta is the focus motor position corresponding to x.

[0021] In step five of the present application, the specific implementation manner is: before starting the algorithm, the lens basic curve data, i.e. the infinite focus curve data, is acquired, the current Focus motor forward speed is set according to the lens curve and the image sharpness, wherein the Focus motor speed is positively correlated with the image sharpness change trend, when the Focus motor is along the current direction, the multi-frame image sharpness value is rising trend, then the Focus motor running speed needs to be appropriately increased, if the descending trend appears afterwards, and the obvious descending trend of the image sharpness value of continuous several frames appears, then it can be judged that the highest point of the current scene along the direction has been searched, the Focus motor direction is reversed to continue searching, if the image sharpness value appears continuous rising, and the obvious descending trend of the image sharpness value of continuous several frames appears, then it can be determined that the clearest point of the scene is searched, the Focus motor goes to the highest position of the scene sharpness, the current mountain climbing search algorithm ends, i.e. the position of the current scene image clearest point corresponds to the Focus motor point value.

[0022] In step seven of the present application, the upper and lower limit values FP UP , FP DOWN are adopted to subdivide.

[0023] Compared with the prior art, the present application has the following advantages and effects: 1. In the actual use of lens stretching scene, the current zoom following curve under the current object distance can be quickly determined, the picture lag caused by the non-smooth change of the curve in the lens zoom pulling process is avoided, and the feeling of frustration is given to people; 2. It has strong universality on the same specification product, and only one basic focusing curve needs to be corrected when the same new lens is introduced, and other difference curves can be derived, thereby optimizing the engineering quantity. 3. The focusing curve of the current object distance can be known in advance, the motor is smoothly moved according to the focusing curve, the zoom motor and the focus motor reach the determined clear position at the same time, and it is ensured that each point is clear. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 It is a mountain climbing algorithm schematic diagram of the embodiment of the present application.

[0025] Figure 2 It is a focusing curve diagram of the 12x zoom lens actually measured under the object distance of 1m-20m in the embodiment of the present application.

[0026] Figure 3 It is a focus motor position schematic diagram corresponding to the focus motor position of the nearest object distance focusing curve when the image is clearest under the current magnification in the embodiment of the present application. DETAILED DESCRIPTION

[0027] The present application will be further described in detail below in combination with the drawings and through embodiments, and the following embodiments are the explanation of the present application, and the present application is not limited to the following embodiments.

[0028] The embodiment of the present application comprises the following steps:

[0029] Step one, using a correction light box to simulate infinite object distance, taking a 12x optical zoom camera as an example, aiming the zoom lens at the light port of the correction light box, according to the zoom lens auto-focusing method, driving the zoom motor to travel from the wide end to the tele end, point by point, and searching for the corresponding focus point motor position under each zoom motor point, and recording it down, storing it as the basic focus curve L0.

[0030] Step two, using a magnifying mirror, an optical device with the function of simulating any object distance, which is different from the correction light box in step one, the correction light box can only simulate infinite object distance, the magnifying mirror can simulate any object distance, by setting the distance between the baffle and the magnifying mirror corresponding to the vertical distance, different object distances can be simulated. Aim the zoom lens at the magnifying mirror and observe the baffle through the magnifying mirror.

[0031] Set the magnifying mirror to simulate object distances of 1m, 2m, 3m, 4m, 5m, 6m, 7m, 8m, 9m, 10m, 11m, 12m, 13m, 14m, 15m, 16m, 17m, 18m, 19m, 20m respectively. Adjust the distance between the baffle and the magnifying mirror corresponding to the vertical distance to simulate the corresponding object distance size. Repeat the method in step one, drive the zoom motor from the wide end to the tele end, point by point, and search for the corresponding focus point motor position under each zoom motor point, and record it down, store it as 20 different object distance focus curves (such as Figure 2 ) : {L i ∶ i = 1, 2, 3, ……, 20}.

[0032] Step three, export the basic focus curve L0 and the 20 different object distance focus curves (L0, L1, L2, ……, L 20 ) stored in step one and step two to an Excel table. The present application adopts the bias method, that is, by subtracting the 20 different object distance focus curves obtained in step two from the basic focus curve L0 in step one, 20 difference curves of the basic focus curve and the different object distance focus curve {L'1, L'2, L'3, ……, L' 20} are obtained.

[0033] Step four, using curve fitting method, the difference curves {L'1, L'2, L'3, ……, L' 20} obtained in step three are fitted into high-order polynomials, which is conducive to the dynamic storage of data in the device. The present application sets the high-order polynomial to be an 8th order polynomial, and records the corresponding polynomial analytical expression:

[0034] δ(x)=ax 8 +bx 7 +cx 6 +dx 5 +ex 4 +fx 3 +gx 2 +hx 1 +i(a≠0);

[0035] Where [a, b, c, d, e, f, g, h, i] are polynomial coefficients, x is the number of zoom motor points, and δ is the position of the focus motor corresponding to x.

[0036] Polynomial coefficients can be obtained through fitting. Specifically, taking L'1 as an example, the difference curve {L'1, L'2, L'3, ..., L'...} obtained in step five can be fitted... 20}, taking L'1, and substituting the corresponding zoom motor position and focus motor position into the formula, we can obtain the polynomial:

[0037]

[0038] With the help of the algorithm platform, the coefficients of higher-order polynomials are solved by encoding, thereby obtaining the analytical coefficients of the difference curve L'1. The coefficients of other curves are solved in the same way.

[0039] Step 5: Using a hill-climbing algorithm, search for the Focus motor position value (FP value) corresponding to the sharpest point in the current scene image at the current magnification. The specific implementation is as follows: Before starting the algorithm, acquire the lens's basic curve data, i.e., the infinity focus curve data. Set the current Focus motor speed based on the lens curve and image sharpness. The Focus motor speed is positively correlated with the image sharpness trend. If the image sharpness value increases over multiple frames along the current direction, the Focus motor speed needs to be appropriately increased. If a decreasing trend subsequently appears, and the image sharpness value shows a significant decreasing trend for three consecutive frames, it can be determined that the highest point of the current scene has been found along this direction. The Focus motor direction is then reversed to continue the search. If the image sharpness value increases continuously, followed by two consecutive frames showing a significant decreasing trend, the sharpest point in the scene has been found. The Focus motor moves to the position with the highest scene sharpness, and the current hill-climbing search algorithm ends. The position of the sharpest point in the current scene image corresponds to the Focus motor position value (FP value).

[0040] Step six, through the current zoom motor point, using the high-order polynomial of each different object distance in step four, the zoom motor ratio corresponding to each different object distance is obtained respectively, and the focus motor position is the most clear position, respectively {FP1, FP2, …, FP i , i = 1, 2, 3 … 20}. According to the FP value of the current scene image most clear point corresponding to the focus motor point value (FP value) under the current magnification searched by the hill climbing algorithm automatic focusing in step five, the squeeze method can be used to obtain the focus motor positions [FP n , FP n+1 ] corresponding to the two object distance focusing curves most adjacent to the FP value. n ∈ [0, 19], FP UP , FP DOWN correspond to FP n and FP n+1 respectively. As shown in Figure 3 .

[0041] Step seven, the upper and lower limit values FP UP , FP DOWN obtained in step six are divided into intervals according to a certain ratio, and the present application adopts 1:2:3:4 for subdivision. After subdivision, the current FP position is obtained according to the squeeze method, the nearest point position FP' of FP in the interval is obtained by using the nearest neighbor method, and the two focusing curves where FP UP and FP DOWN are located are taken as the reference. From the wide end to the tele end, the focus motor position {FP1, FP2, FP3, …, FP n} is calculated by taking the ratio Ratio of FP' in FP up -FP down , n is the total number of zoom motor intervals, and the current focusing object distance variable tracking curve L = {FP1, FP2, FP3, …, FP n} is obtained. During the variable magnification process, each zoom motor position corresponds to a focus motor position, which can ensure that the variable magnification lens always remains within the depth of field during the variable magnification process.

[0042] In addition, it should be noted that the specific embodiments described in the specification, the shape of the components, the name taken, etc. can be different, the above described in the specification is only an example of the structure of the present application. Any equivalent changes or simple changes made in accordance with the structure, features and principles described in the patent concept of the present application are included in the protection scope of the present application. Those skilled in the art can make various modifications or supplements to the described specific embodiments or use similar ways to replace, as long as they do not deviate from the structure of the present application or exceed the scope defined by the present claims, which shall belong to the protection scope of the present application.

Claims

1. A zoom lens magnification following method characterized by: It comprises the following steps: Step one, using the correction light box, zoom lens is aimed at the correction light box light port, drive zoom motor, from the near end to the far end, point by point, and search for each zoom motor point under the corresponding focus point focus motor position, and record down, store as the basis of focusing curve; Step two, using the magnifying mirror, zoom lens is aimed at the magnifying mirror; Set up multiple magnifying mirror simulation object distance respectively; Drive zoom motor, from the near end to the far end, point by point, and search for each zoom motor point under the corresponding focus point focus motor position, and record down, store as multiple object distance focusing curve; Step three, through the basis of focusing curve in step one, respectively, with the multiple object distance focusing curve obtained in step two, get the difference curve of the basis of focusing curve and object distance focusing curve; Step four, using curve fitting method, the difference curve obtained in step three is fitted into high order polynomial; Step five, based on hill climbing algorithm, search for the current scene image clearest point corresponding focus motor point value under the current magnification; Step six, through the current zoom motor point, using the high order polynomial of each different object distance in step four, respectively, get the focus motor position when the zoom motor magnification corresponds to the clearest position under each different object distance; According to the Focus motor point position value corresponding to the sharpest point of the current scene image obtained in step five, the Focus motor positions corresponding to the two object distance focusing curves closest to the point position value are obtained, i.e. the upper and lower limit values FP of the Focus motor position under the current magnification corresponding to the two object distance focusing curves closest to the point position value UP , FP DOWN ; Step 7: Convert the upper and lower position limits FP obtained in Step 6. UP FP DOWN The interval is subdivided, and the current Focus motor position is obtained after subdivision. The position FP' of the Focus motor in the interval is obtained, and then the position FP' of the nearest point of the Focus motor in step six is ​​obtained. UP and FP DOWN Based on the two object distance focusing curves, with FP' at FP up -FP down The proportion of the focal length is calculated point by point from the near focal end to the far focal end, and the focus motor position {FP1, FP2, FP3, ..., FP} is calculated. n }, where n is the total number of points in the zoom motor range. At this point, the zoom tracking curve L = {FP1, FP2, FP3, ..., FP} can be obtained at the current focusing distance. n During zooming, each zoom motor position corresponds to a focus motor position, ensuring that the zoom lens maintains a clear image within the depth of field throughout the zooming process.

2. The zoom lens magnification following method according to claim 1, characterized by: In step two, set the magnifying mirror simulation object distance to 1m, 2m, 3m, 4m, 5m, 6m, 7m, 8m, 9m, 10m, 11m, 12m, 13m, 14m, 15m, 16m, 17m, 18m, 19m, 20m.

3. The zoom lens magnification following method of claim 1, wherein: In step four, the high order polynomial is 8 order polynomial, and the corresponding polynomial expression is recorded: δ(x) = ax 8 + bx 7 + cx 6 + dx 5 + ex 4 + fx 3 + gx 2 + hx 1 + i (a≠0); Where [a, b, c, d, e, f, g, h, i] is the polynomial coefficient, x is the zoom motor point, and δ is the focus motor position corresponding to x.

4. The zoom lens magnification following method of claim 1, wherein: In step five, the specific implementation is: before the algorithm starts, get the lens basis curve data, that is, the infinite focusing curve data, set the current focus motor forward speed according to the lens curve and image sharpness, wherein the focus motor speed is positively correlated with the image sharpness change trend, when the focus motor is along the current direction, the image sharpness value is rising trend, the focus motor running speed should be appropriately improved, if the downward trend appears after that, and the image sharpness value of several continuous frames appears obvious downward trend, it can be judged that the highest point of the current scene has been searched along the direction, so the focus motor direction is reversed to continue searching, if the image sharpness value appears continuous rise, and the image sharpness value of several continuous frames appears obvious downward trend, it can be determined that the clearest point of the scene is searched, the focus motor goes to the highest position of the scene sharpness, the current hill climbing search algorithm ends, that is, the current scene image clearest point position corresponds to the focus motor point value.

5. The zoom lens magnification change following method according to claim 1, characterized by: In step seven, the upper and lower values FP UP , FP DOWN are subdivided using 1:2:3:4.

Images