An optimization method for the cam table of a continuous zoom system

Abstract

The present invention discloses an optimization method for a cam table of a continuous zoom system, which relates to the field of electronic cam curve generation. The method adopts a segmented idea. First, using the theoretical cam data, the moving range of the lens is equally divided to obtain sampling points. Then, taking the reciprocal of the difference at the sampling points as the weight, the sampling points are reset to segment the cam curve. After obtaining the cam data at the sampling points through the focusing technology, the cubic Hermite interpolation algorithm is used to calculate other cam data. The beneficial effects of the present invention: simulation experiments are carried out by simulating with the theoretical cam curve. Different numbers of sampling points n are defined respectively, and the sampling points corresponding to the theoretical cam table are obtained according to the form length optimization algorithm. Then, according to the main axis data of the theoretical cam table, the corresponding slave axis positions are obtained through cubic Hermite interpolation in sequence, and the error after interpolation is calculated. Since when the system is at short focus, that is, when the position of the variable magnification group lens is less than 50,000 steps, the theoretical cam data can meet the system accuracy requirements. Therefore, when analyzing the error, the cam data with the position of the variable magnification group greater than 50,000 steps is mainly analyzed.

Description

Technical Field

[0001] The present invention relates to the field of electronic cam curve generation, and more specifically to the fields of non-uniform curve sampling reconstruction and research and application of optimizing the cam data length of an infrared continuous zoom system. Prior Art

[0002] The continuous zoom function of military optoelectronic systems is realized through cam technology. Cams are divided into mechanical cams and electronic cams. Mechanical cams require high machining precision and are prone to wear. Compared with mechanical cams, electronic cams have better flexibility, lower cost, and lower mechanical noise. The zoom system usually consists of four groups of lenses: a front fixed group, a zoom group, a compensation group, and a rear fixed group. The cam table is the position data of the zoom group and the compensation group at the corresponding points when the system forms an image within the moving range, with the position unit being the step size, corresponding to the position of the lens motor encoder disk. The two groups of lenses, the zoom group and the compensation group, move together according to the preset cam data, so that the image remains clear during the zoom process of the system, thereby realizing the continuous zoom function of the system.

[0003] The minimum resolution of the system corresponds to a fixed step size Step fixed , and the change in image clarity caused by the movement of the lens within this step size is not obvious to the human eye. Therefore, in order to ensure the accuracy of focusing during continuous zooming, the data of adjacent two groups of the electronic cam data of the zoom group are spaced by Step fixed step sizes, and the ortho-focus position of any field of view can be obtained by interpolating the cam data. For a certain type of infrared optoelectronic continuous zoom system, the step size corresponding to the minimum resolution of the system is 500, the moving range of its zoom group is 0 to 500,000 step sizes, and the moving range of the corresponding compensation group is 0 to 160,000 step sizes. Therefore, there are 1,000 sets of theoretical data corresponding to the electronic cam table, which serves as the theoretical support for the optical design and mechanical design of the optoelectronic system. The cam curve is as Figure 1 shown. If a margin is left in the optical design, it can cover the errors caused by mechanical processing and installation. However, currently, the focal length of optoelectronic systems is getting longer, and the smaller depth of focus brought by the longer focal length greatly increases the requirements for design and installation accuracy. Therefore, the cam table is generated through focusing technology calibration after the equipment is produced to meet the accuracy requirements of the system.

[0004] Currently, the way to generate the cam table is to obtain the corresponding compensation group position through focusing technology for each position of the zoom group, thereby obtaining the entire electronic cam table. Summary of the Invention

[0005] The object of the present invention is: It takes about 5 seconds on average to obtain a set of cam data using the focusing technology, and it takes up to about 90 minutes to calibrate 1,000 sets of cam data. A large amount of data greatly increases the task volume of generating the cam table. Therefore, it is necessary to find a suitable optimization strategy to generate the cam table faster without losing the focusing accuracy of the cam table.

[0006] To solve the problems of heavy workload and long time consumption in generating the cam table, the present invention adopts a segmented idea. First, using the theoretical cam data, the moving range of the compensation lens is equally divided to obtain sampling points. Then, taking the reciprocal of the difference at the sampling points as the weight, the sampling points are reset to segment the cam curve. After obtaining the cam data at the sampling points through the focusing technology, the cubic Hermite interpolation algorithm is used to calculate other cam data. The specific segmented optimization strategy is as follows:

[0007] Step 1: Define the number of sampling points

[0008] Define the number of points n to be sampled. When generating the cam table using the focusing technology later, only n - 1 groups of data except the first group of cam data (0,0) need to be sampled and generated.

[0009] Step 2: Obtain the equally spaced difference data of the theoretical cam curve

[0010] On the cam curve, the moving range L of the compensation lens group is equally divided into n - 1 segments, and the difference value at each end point is d. Take the reciprocal of the difference and denote it as α = 1 / d.

[0011] Step 3: Calculate the position weights of the sampling points

[0012] Add the reciprocals α of the difference data at all n - 1 points to get sum(α), and calculate the proportion of α at the n - 1 end points in the total sum to obtain the weight w of the sampling position i = α i / sum(α).

[0013] Step 4: Allocate the positions of the sampling points according to the weights

[0014] Allocate the movable range L of the zoom lens group according to the weight w i to determine the position p of the cam data to be sampled i = L·w i .

[0015] Step 5: Generate the corresponding cam table using autofocus

[0016] Using the focusing technology, the zoom lens group focuses at the n - 1 end points to obtain the corresponding positions of the compensation lens group and record them, obtaining cam data with a length of n.

[0017] Step 6: Obtain all the cam data through cubic Hermite interpolation

[0018] Corresponding to the position of the zoom lens group in the theoretical cam table, perform cubic Hermite interpolation through the generated n groups of cam tables to obtain the corresponding positions of the compensation lens group and obtain the complete cam data.

[0019] Advantages of the present invention: Simulation experiments are carried out using the theoretical cam curve simulation. Different sampling point numbers n are respectively defined, and the sampling points corresponding to the theoretical cam table are obtained according to the form length optimization algorithm. Then, according to the main axis data of the theoretical cam table, the corresponding slave axis positions are obtained through cubic Hermite interpolation in sequence, and the error after interpolation is calculated. Since the system is in the short focal length, that is, when the position of the varifocal lens group is less than 50,000 step lengths, the theoretical cam data can meet the system accuracy requirements. Therefore, when analyzing the error, the cam data with the varifocal lens group position greater than 50,000 step lengths is mainly analyzed.

[0020] As Figure 2 shown, when the sampling point number is 10, the cam curve is divided into 9 segments. As the slope of the cam curve decreases, the interval of the sampling points gradually increases. After segmentation, one thousand groups of cam data are obtained by using cubic Hermite interpolation, and the error curve is obtained by taking the difference between the interpolated data and the original data, as Figure 3 shown.

[0021] The experimental results show that: when the sampling point number is 10, the error curve of the main axis with a step length greater than 500,000 basically remains horizontal, and the difference value remains within [-47, 100], which is less than the 500 step lengths corresponding to the system accuracy and meets the system accuracy requirements. Therefore, theoretically, 10 sampling points can generate the cam table.

[0022] When the sampling point number is increased to 20, the sampling position and the error curve after interpolation are as Figure 4 shown in Figure 5 and

[0023] shown. The experimental results show that: when the sampling point number is 20, the difference value of the error curve of the main axis with a step length greater than 50,000 fluctuates within ±10, which is much less than the step lengths corresponding to the system accuracy and meets the system accuracy requirements.

[0024] In summary, this optimization strategy can greatly reduce the data required for generating the cam table, from the original 1006 groups of data to about 10 groups, and a cam table that meets the system accuracy can be generated. Without affecting the focusing accuracy, the time for generating the cam table can be greatly reduced.

[0025] The cam table optimization strategy proposed by the present invention reduces the time for generating the cam table, which is reduced to about 1 / 20 of the original generation time. From the above analysis, taking 10 sampling points can meet the system index requirements, and the accuracy of the cam table generated in about 3 minutes is the same as that of the real cam table generated in 80 minutes. The optimization strategy of obtaining sampling points using the speed of change of the cam curve as the weight and then obtaining cam data through segmented cubic Hermite interpolation greatly reduces the workload of generating the cam table, improves the production efficiency of optoelectronic devices, and reduces the loss of the mechanical structure during the generation of the cam table, thereby improving the service life of optoelectronic devices. Description of the Drawings

[0026] Figure 1 is the theoretical cam curve of a certain type of infrared optoelectronic continuous zoom system in the prior art;

[0027] Figure 2 is the position distribution diagram with 10 sampling points in the embodiment;

[0028] Figure 3 is the interpolation error curve with 10 sampling points in the embodiment;

[0029] Figure 4 is the position distribution diagram with 20 sampling points in the embodiment;

[0030] Figure 5 is the interpolation error curve with 20 sampling points in the embodiment;

[0031] Figure 6 is the error of interpolating 1006 groups of data from 10 groups of cam data in the embodiment;

[0032] Figure 7 is the error of interpolating 1006 groups of data from 20 groups of cam data in the embodiment. Detailed Embodiment

[0033] The effect of the present invention is verified by a certain domestic infrared optoelectronic continuous zoom system. The moving range of the variable magnification lens (main axis) of a certain infrared optoelectronic continuous zoom system is 0 to 500,000 steps, and the corresponding moving range of the compensation lens (secondary axis) is 0 to 160,000 steps. Its theoretical cam table has 1006 groups of cam data. The 500-step length of the system is the step length corresponding to the minimum resolution of the system. When the step length is within 500, the change in image clarity is not obvious to the human eye.

[0034] When the system is in the short focal length, that is, when the position of the variable magnification lens is less than 50,000 steps, the theoretical cam data can meet the system accuracy requirements. Therefore, when verifying the optimization performance of the cam table, the cam data with the variable magnification lens position greater than 50,000 steps is mainly analyzed. In the optoelectronic system, cam tables with 10 and 20 sampling points are respectively generated using the cam table optimization strategy, denoted as Cam10 and Cam20, as shown in Table 1.

[0035] Table 1 Cam data of 10 groups and 20 groups generated by the optimization strategy

[0036]

[0037] The generation steps of Cam10 and Cam20 are as follows:

[0038] Step 1: Define the number of sampling points n = 10 and n = 20 respectively.

[0039] Step 2: Divide the moving range of the compensation group lens into 9 segments and 19 segments on the cam curve, and obtain the difference values of the corresponding cam data.

[0040] When n = 10, divide the moving range of the compensation group lens equally, and the cam data groups corresponding to the endpoints are 1, 113, 224, 336, 448, 559, 671, 783, 894, 1006 respectively.

[0041] When n = 20, divide the moving range of the compensation group lens equally, and the cam data groups corresponding to the endpoints are 1, 54, 107, 160, 213, 265, 318, 371, 424, 477, 530, 583, 636, 689, 742, 794, 847, 900, 953, 1006 respectively. Then, obtain the reciprocals of the differences of each group of cam data.

[0042] Step 3: Obtain the position weights of the sampling points as shown in Table 2

[0043] Table 2 Cam data of 10 groups and 20 groups of data generated by the optimization strategy

[0044]

[0045]

[0046] Step 4: Allocate the positions of the sampling points according to the weights as shown in Table 3

[0047] Table 3 Positions of the sampling points

[0048]

[0049] Step 5: Use autofocus to generate the corresponding cam table

[0050] Using the focusing technology, obtain the cam data with lengths of 10 and 20.

[0051] Step 6: Obtain all the cam data through cubic Hermite interpolation

[0052] Corresponding to the position of the variable magnification group in the theoretical cam table, perform cubic Hermite interpolation through the generated 10 - group and 20 - group cam tables to obtain the corresponding positions of the compensation group one by one, and obtain the cam data with a length of 1006.

[0053] The generation times of the two groups of cam tables are 2.3 min and 3.5 min respectively. Then, cubic Hermite interpolation is used. According to the main shaft positions of the real cam table, the corresponding slave shaft positions are interpolated. The 1006 groups of cam data generated by 10 groups of interpolation are denoted as Cam10to1006, and the 1006 groups of cam data generated by 20 groups of interpolation are denoted as Cam20to1006. Error analysis is performed with the real cam table respectively, and the difference results are as Figure 6 and Figure 7 shown.

[0054] The experimental results in this embodiment show that:

[0055] (1) When the number of sampling points is 10, the difference curve between the optimized cam table and the real cam table fluctuates within [-391, 477], meeting the requirement that the system index is less than ±500.

[0056] (2) When the number of sampling points is 20, the difference curve between the optimized cam table and the real cam table fluctuates within [-441, 387], meeting the requirement that the system index is less than ±500.

[0057] (3) When the number of sampling points is 10, the cam table generation time is 2.3 min. When the number of sampling points is 20, the cam table generation time is 3.5 min. Both meet the requirement that the system index is less than 10 min.

[0058] Therefore, the cam table optimization strategy proposed in this paper greatly reduces the workload of generating the cam table, improves the production efficiency of optoelectronic devices, and reduces it to about 1 / 20 of the original generation time. It reduces the loss of the mechanical structure during the generation of the cam table and improves the service life of optoelectronic devices.

Claims

1. An optimization method for a cam table of a continuous zoom system, characterized in that, It includes the following steps: Step 1: Define the number of sampling points: Define the number of points n to be sampled. When generating the cam table using the focusing technique later, only n - 1 groups of data except the first group of cam data (0, 0) need to be sampled and generated; Step 2: Obtain the equally-spaced differential data of the theoretical cam curve: Divide the moving range L of the compensation group lens on the cam curve into n - 1 equal segments. The differential value at each end point is d, and take the reciprocal of the differential, denoted as α = 1 / d; Step 3: Calculate the position weights of the sampling points: Add the reciprocals α of the differential data at all n - 1 points to obtain sum(α), calculate the proportion of α at the n - 1 end points in the total, and obtain the weight of the sampling position as w i = α i / sum(α); Step 4: Allocate the positions of the sampling points according to the weights: The movable range L of the zooming lens group is weighted by w i to allocate and determine the position p where the cam data needs to be sampled i = L·w i ; Step 5: Generate the corresponding cam table using autofocus: Using the focusing technique, the variable magnification group focuses at n - 1 end points to obtain the corresponding positions of the compensation group and record them, obtaining cam data with a length of n; Step 6: Obtain all the cam data through cubic Hermite interpolation: For the positions of the variable magnification group corresponding to the theoretical cam table, perform cubic Hermite interpolation through the generated n groups of cam tables to obtain the corresponding positions of the compensation group and obtain the complete cam data.

Images