Joint zooming method and system

Abstract

The invention relates to the technical field of zoom lenses, in particular to a joint zoom method and system, which comprises the following steps: acquiring a zoom instruction, calling image features, judging an illumination state, determining a dominant control core channel type, identifying optical axis offset, generating a reconstruction direction pointing angle, detecting boundary matching, and adjusting the aperture width. And triggering light path extension to verify that image surfaces are consistent, analyzing a track trend to confirm a picture, and forming a combined zooming scheme. According to the method, the main control channel is dynamically identified through the imaging features and the brightness coverage state, accurate extraction of the optical axis offset direction and reconstruction of the optical axis direction are achieved, edge coincidence detection and aperture synchronous finishing actions in the image boundary registration process are guided, and the image boundary matching compactness is improved; furthermore, the imaging consistency between the channels is kept through light path extension and image plane direction linkage control, the visual angle fusion precision is improved in combination with multi-frame image edge trajectory comparison, and the image synchronism and the picture stability in the continuous zooming process are effectively enhanced.

Description

Technical Field

[0001] This invention relates to the field of zoom lens technology, and in particular to a combined zoom method and system. Background Technology

[0002] The field of zoom lens technology involves adjusting the position of one or more lens groups in an optical system to change the field of view to meet the needs of different observation or imaging distances. This technology is widely used in scenarios such as surveillance, aerial photography, imaging measurement, autonomous driving, and infrared thermal imaging. Its core aspects include optical zoom structure design, zoom drive control, field of view matching strategy, image synchronous processing, and multi-sensor coordinated control. The overall technical field covers the integration of mechanical structure and electronic control, as well as the joint acquisition and processing of visible light and infrared images. Among them, the traditional joint zoom method refers to the method of achieving consistent viewing angles of dual-path images by simultaneously controlling visible light cameras and infrared cameras in multimodal imaging systems such as UAV pods. The technical issue addressed by this patent is the problem of field of view misalignment and image mismatch caused by differences in optical parameters during the zooming process of dual-path cameras. Traditional joint zoom usually adopts an optical zoom mechanism that independently controls two channels and is supplemented by digital image cropping and scaling to achieve field of view unification. In the execution process, the corresponding strain factor is often found by pre-setting a focal length table and the image cropping area and scaling factor are determined by software calculation to match the target field of view adjustment.

[0003] Traditional combined zoom relies on a preset focal length table to drive optical zoom linkage, making it difficult to judge illumination adaptability based on actual imaging conditions. It lacks a unified axis offset judgment mechanism and synchronous calibration strategy across different channels, resulting in a lag in the optical axis alignment process. In the image boundary cropping and registration stage, no dynamic detection basis and real-time correction methods are introduced, making edge matching susceptible to short-term deviations caused by the lens zoom drive inertia. This leads to image plane instability in continuous images. In particular, the lack of an edge trajectory continuity judgment mechanism during multi-frame image analysis makes image fusion prone to mismatch during zoom operation, and image frame consistency is difficult to maintain. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a combined zoom method, comprising the following steps: S1: After obtaining the zoom command, retrieve the image clarity and texture area of ​​the infrared lens, and determine the illumination status by comparing it with the outline and brightness of the visible light image, and select the dominant control core channel type. S2: Based on the dominant control core channel type, read the lens focal length number and identify the optical axis offset direction. By calculating the change in the angle between the center lines of the two channels, generate the reconstructed direction pointing angle. S3: Based on the reconstructed direction pointing angle, detect the matching of the viewpoint edge with the main channel, determine whether the focal length boundary is overlapped through the zoom component, adjust the aperture to trim the boundary width, and generate the viewpoint edge position that is synchronously covered. S4: Based on the edge position of the synchronously covered viewpoint, trigger the optical path extension, adjust the viewing axis to the main channel direction, verify that the two channel image planes are projected in the same way, determine whether the image is fused and reconstructed, and generate a master-slave image superimposed frame. S5: Based on the superimposed frame size of the master and slave images, analyze the displacement trend of the boundary points, compare the overlap of the edge trajectories, and if no offset or boundary-crossing trajectory occurs, form a joint zoom scheme.

[0005] As a further aspect of the present invention, the dominant control core channel type includes image sharpness criteria, texture area index, contour distribution density, and brightness distribution ratio; the reconstruction direction pointing angle includes focal length corresponding number, optical axis offset azimuth, and channel center angle change; the synchronously covered viewing angle edge position includes main channel edge coordinates, secondary channel boundary matching degree, and edge cropping width value; the master-slave image superposition frame includes image plane projection direction consistency, image fusion integrity, and channel exit angle synergy; and the joint zoom scheme includes boundary point displacement trend line, image edge trajectory overlap, and continuous zoom segment stability judgment result.

[0006] As a further aspect of the present invention, the boundary point displacement trend refers to the change pattern of the positional movement direction and magnitude of corresponding points at the edges of the master and slave images during image zooming, as time and focal length change.

[0007] As a further aspect of the present invention, the offset out-of-bounds trajectory refers to the motion path in which the image edge trajectory deviates from the main channel's viewpoint range and exceeds the overlap error range during the boundary point displacement process.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: After obtaining the zoom command from the control end, call the infrared lens image frame, extract the gray-scale gradient and texture distribution of the edge region, calculate the gradient standard deviation of adjacent regions and count the area pixels of continuous texture regions, and generate the edge sharpness and continuous texture area value of the infrared image. S102: Based on the edge sharpness and continuous texture area value of the infrared image, obtain the visible light image frame, extract the contour region to calculate the edge line density and brightness pixel ratio of the unit pixel, and combine the contour line density judgment threshold and the brightness coverage ratio judgment threshold to make a joint judgment and generate the current image contour density and brightness coverage comparison factor. S103: Call the infrared image edge sharpness and continuous texture area value and the current image contour density and brightness coverage comparison factor to perform a difference comparison analysis on the image feature stability of infrared and visible light images. If the gradient standard deviation of the infrared image is lower than that of the visible light image and the corresponding brightness coverage ratio is lower than the edge threshold benchmark, generate the dominant control core channel type.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the dominant control core channel type, read the lens configuration parameter set associated with the current channel, retrieve the list of focal length indexes that can be matched collaboratively, filter the focal length range called by the current channel in combination with the channel identifier parameter, and extract the number parameter with the focal length linkage control identifier from it to generate a set of matching lens focal length numbers. S202: Call the set of matching lens focal length numbers, access the displacement information of the optical axis correction lens group built into the structure, extract the direction mark vector of the lens assembly, retrieve the correction mark corresponding to the main channel focal length, and obtain the offset direction vector of the optical axis of the secondary channel by comparing the direction relationship between the mark and the initial attitude of the optical axis of the secondary channel. S203: Based on the optical axis offset direction vector of the secondary channel, combined with the initial angle value between the center lines of the primary channel and the secondary channel, continuously calculate the change in the center line angle under the differentiated focal length number, extract the primary direction component in the change trend, and after performing consistency correction on the sign of the primary direction component, establish the reconstructed direction pointing angle.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the reconstructed direction pointing angle, extract the imaging view range of the main channel image frame, and calibrate the boundary coordinate group of the image edge under the current view. Obtain the horizontal and vertical borders within the image range from the channel image frame, compare them with the main channel boundary coordinate group, calculate the root mean square of the coordinate difference between corresponding edge points between the borders, and obtain the image border matching deviation value. S302: Based on the image border matching deviation value, retrieve the status of the front zoom component under the main channel configuration, set the boundary coincidence error threshold as the boundary coordinate difference limit range, and verify whether the zoom component is called to adjust the focal length position by judging whether the current deviation value is lower than the boundary coincidence error threshold, and obtain the boundary coincidence adjustment status identifier. S303: Invoke the boundary coincidence adjustment status flag, control the channel aperture position according to the boundary coincidence status, and adjust the width of the boundary area formed after the field of view is clipped by synchronously adjusting the aperture edge opening angle. Combine the view coordinate group after boundary adjustment to establish the view edge position of synchronous coverage.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Based on the edge position of the synchronously covered viewpoint, trigger the extension command of the lens middle section structure, monitor the state of the optical path axis inside the channel, extract the angle difference between the current direction vector of the optical axis and the corresponding direction vector of the main channel, control the optical axis to move in the direction of movement through the advancement action of the structural components, so that the angle between the two vectors gradually approaches zero, and obtain the optical axis alignment offset. S402: Call the optical axis alignment offset, extract the lens displacement data of the lens group during the axial advancement process, and combine the difference of the normal vectors of the two channel image planes before and after advancement to calculate the image projection direction consistency error, determine whether the image plane normal angle is within the set image plane parallelism error tolerance range, and obtain the image plane alignment state parameters. S403: Based on the image alignment state parameters, analyze the continuous change trend curve of the emission angle direction. If the emission direction remains stable and the rate of change is lower than the projection consistency judgment threshold, verify that the current image has been fused and reconstructed, and perform frame-level synchronous fusion of the main channel and the slave channel image matrices to generate a master-slave image superimposed frame.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the superimposed frame of the master and slave images, obtain the image frame sequence within the current continuous imaging cycle, extract the coordinates of the boundary points at the upper left and lower right corners of each frame, calculate the coordinate difference of the same boundary point in adjacent frames, and perform moving average processing on the difference sequence to obtain the trend of boundary point displacement. S502: Based on the trend of boundary point displacement change, extract the coordinate set of edge trajectory lines in the corresponding frame sequence, perform alignment processing on the edge trajectory lines in the main channel and the secondary channel image frames respectively, calculate the overlap rate index between the corresponding trajectory lines, and calculate the average area of ​​the overlapping area of ​​trajectory lines in all frames to obtain the edge trajectory line overlap degree value. S503: Call the edge trajectory line overlap value, combine it with the boundary point coordinate sequence of the master and slave channels in the current continuous zoom operation segment, determine whether there is an offset trajectory of the channel boundary that exceeds the image edge. If all boundary trajectories remain stable in the overlap area, they are marked as fusionable intervals, and a joint zoom scheme is established.

[0013] A combined zoom system, comprising: The channel selection module is used to implement S1: after obtaining the zoom command, it retrieves the image sharpness and texture area of ​​the infrared lens, and determines the illumination state by comparing it with the outline and brightness of the visible light image, and selects the dominant control core channel type. The optical axis correction module is used to achieve S2: based on the type of the dominant control core channel, read the lens focal length number and identify the optical axis offset direction, and generate the reconstruction direction pointing angle by calculating the change in the angle between the center lines of the two channels; The viewpoint alignment module is used to implement S3: based on the angle of the reconstruction direction, detect the match between the viewpoint edge and the main channel, determine whether the focal length boundary is overlapped by the zoom component, adjust the aperture trimming boundary width, and generate the viewpoint edge position that is synchronously covered. The image fusion module is used to implement S4: based on the position of the viewing angle edge of the synchronous coverage, trigger the optical path extension, adjust the viewing axis to the main channel direction, verify that the image planes of the two channels are projected in the same way, determine whether the image is fused and reconstructed, and generate a master-slave image superimposed frame. The zoom judgment module is used to implement S5: based on the frame size of the master and slave image superimposed image, analyze the displacement trend of the boundary points, compare the overlap of the edge trajectories, and if no offset or boundary-crossing trajectory occurs, form a joint zoom scheme.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the main control channel is dynamically identified by imaging features and brightness coverage status, which enables accurate extraction of the visual axis offset direction and reconstruction of the optical axis direction. This guides edge coincidence detection and aperture synchronous adjustment during image boundary registration, improving the tightness of image boundary matching. Furthermore, the consistency of imaging between channels is maintained by optical path extension and image plane direction linkage control. Combined with multi-frame image edge trajectory comparison, the accuracy of viewpoint fusion is improved, effectively enhancing image synchronization and frame stability during continuous zoom. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] Please see Figure 1 This invention provides a combined zoom method, comprising the following steps: S1: After obtaining the zoom command from the control terminal, retrieve the edge sharpness and continuous texture area range of the infrared lens image. By comparing the stability trends of the two types of indicators horizontally, and combining the density of contour lines and the brightness coverage ratio in the current visible light image, determine the current state of the two lenses in the lighting environment, select infrared and visible light as the current main control path, and generate the dominant control core channel type. S2: Based on the type of the dominant control core channel, read the lens focal length number that can be matched collaboratively, call the displacement direction mark of the optical axis correction lens group built into the structure, identify the offset direction of the optical axis of the secondary channel under the focal length of the main channel, extract the actual direction displacement of the optical axis reconstruction by calculating the changing trend of the angle between the center lines of the two channels, and generate the reconstruction direction pointing angle. S3: Based on the direction of reconstruction, the edge of the main channel imaging view is used as the reference boundary. By detecting the matching degree between the horizontal and vertical borders in the channel image range and the main channel, it is determined whether the boundary within the focal length is overlapped through the front zoom component. At the same time, the aperture position is controlled to register and adjust the width of the cropped boundary, generating the synchronously covered view edge position. S4: Based on the edge position of the synchronously covered viewpoint, the optical path extension action of the middle section structure of the lens is triggered, the visual axis of the channel is adjusted to the direction corresponding to the main channel, and the projection direction of the two channel image planes is kept consistent through the axial advancement of the lens group. The image fusion and reconstruction are judged by combining the trend of the change of the exit angle direction, and the master and slave image superimposed image frame is generated. S5: Based on the image size of the master-slave image overlay, analyze the displacement trend of the boundary points at the upper left and lower right corners of the image in the continuous imaging cycle, compare the degree of overlap of the edge trajectory lines in the previous and subsequent frames, and if no offset or out-of-bounds trajectory occurs at the boundary of the master and slave channels in the current continuous zoom operation segment, mark and form a joint zoom scheme.

[0023] The dominant control core channel types include image sharpness criteria, texture area index, contour distribution density, and brightness distribution ratio. The reconstruction direction pointing angle includes focal length corresponding number, optical axis offset orientation, and channel center angle change. The synchronously covered viewpoint edge position includes master channel edge coordinates, slave channel boundary matching degree, and edge cropping width value. The master-slave image superposition frame includes image plane projection direction consistency, image fusion integrity, and channel exit angle synergy. The joint zoom scheme includes boundary point displacement trend line, image edge trajectory overlap degree, and continuous zoom segment stability judgment result.

[0024] Please see Figure 2 The specific steps of S1 are as follows: S101: After obtaining the zoom command from the control end, call the infrared lens image frame, extract the gray-scale gradient and texture distribution of the edge region, calculate the gradient standard deviation of adjacent regions and count the area pixels of continuous texture regions, and generate the edge sharpness and continuous texture area value of the infrared image. First, the preset magnification in the instruction is parsed, and this magnification is compared item by item with the current magnification parameter of the infrared lens in the current imaging. If there is a discrepancy, image acquisition is immediately invoked to obtain an infrared lens image frame. During the acquisition process, a timestamp is used to lock the current image acquisition time point to ensure that the processed image is the latest image frame. Then, the image frame is loaded into image processing for edge region extraction. Specifically, spatial filtering is performed on the grayscale values ​​of the image frame to highlight areas with obvious edge changes. A 5×5 sliding window is used to scan the entire image area to extract the grayscale value variation range of pixels in all edge regions. The grayscale value variation results are then recorded according to the pixel row and column directions. Next, the gradient value set is extracted for the sub-blocks formed by adjacent pixels in each region. Based on this, the difference of the pixel grayscale gradient within each sub-block is calculated and... The sum of squares is accumulated, and the average deviation is calculated based on the total number of pixels in the sub-block to obtain the standard deviation. A regional standard deviation threshold is set to determine whether the sub-block is a boundary region. This threshold is dynamically adjusted according to the background noise environment of the image. It is usually set to 25 in static monitoring scenarios and 35 in outdoor environments. Regions with a standard deviation higher than this threshold are selected as valid edge regions. Adjacent edge blocks are identified one by one to determine whether there are structurally continuous and uniformly distributed regions. The minimum number of continuous blocks is set to 6 blocks, and the standard deviation difference between them is required to be no more than 10 gray levels. This is used to identify continuous texture regions. Finally, the number of pixels in all such texture regions is counted and recorded as the continuous texture area. At the same time, the standard deviations of all valid edge sub-blocks are statistically averaged as a reference value for the edge sharpness of the infrared image. Finally, the edge sharpness and continuous texture area values ​​of the infrared image are generated.

[0025] S102: Based on the edge sharpness and continuous texture area value of infrared image, obtain visible light image frame, extract contour region to calculate unit pixel edge line density and brightness pixel ratio, combine contour line density judgment threshold and brightness coverage ratio judgment threshold for joint judgment, and generate current image contour density and brightness coverage comparison factor. A visible light camera is simultaneously activated to acquire the visible light image frame at the current moment. Upon receiving the image frame, contour extraction processing is immediately performed. After grayscale conversion and contrast enhancement, edge extraction is performed. For each 20×20 pixel region of the image, the number of edge pixels generated by grayscale abrupt changes within that region is counted, and the ratio of this number to the total number of pixels in the region is calculated to determine the edge line density per unit pixel. The preset edge density judgment threshold is 0.12, which is derived by analyzing the structural edge distribution characteristics in standard images. For typical scenes such as urban roads or indoor architectural environments, this threshold has high stability. After contour extraction, the brightness coverage area is further analyzed. By performing a global scan of the image grayscale values, pixels with brightness values ​​greater than 200 are marked as bright pixels, and the number of such pixels is counted. The proportion of pixels in the total number of pixels in the image is the brightness coverage ratio. Based on the standard lighting conditions of the image, the brightness coverage ratio judgment threshold is set to 0.08. Above this value, it indicates that the overall brightness of the image is relatively strong. The above two indicators are jointly judged for each region. Only when the edge line density of the region is not less than 0.12 and the brightness coverage ratio is not less than 0.08, the region is considered to be a structurally significant region. The joint judgment statistics are performed on all regions of the entire image, and finally the edge line density and brightness coverage ratio of all regions are averaged to synthesize the current image contour density and brightness coverage comparison factor. The edge density judgment threshold is 0.12 (dimensionless, representing the proportion of edge pixels), and the brightness coverage ratio threshold is set to 0.08 (that is, the proportion of bright pixels in this region is 8%).

[0026] S103: Call the infrared image edge sharpness and continuous texture area value and the current image contour density and brightness coverage comparison factor to perform a difference comparison analysis on the image feature stability of infrared and visible light images. If the gradient standard deviation of the infrared image is lower than that of the visible light image and the corresponding brightness coverage ratio is lower than the edge threshold benchmark, generate the dominant control core channel type. To perform image channel feature stability analysis, the corresponding feature value sequences of the previous 10 frames of infrared and visible light images are retrieved. Cross-frame stability analysis is then conducted on the edge sharpness values ​​in the infrared images. A variation sequence is constructed using the absolute value of the difference between each frame's feature value and the value of the previous frame. The average variation amplitude of this variation sequence is calculated to quantify the edge feature variability of the infrared image. Simultaneously, the same method is applied to the contour density and brightness coverage comparison factor sequences in the visible light image. If the average variation value of the infrared image is lower than that of the visible light image, it indicates that the infrared image is more stable in the current scene. Next, the brightness of the current frame of the visible light image is retrieved. The average coverage value is compared with the set edge threshold benchmark. This benchmark value is set based on the brightness coverage ratio of statistical images concentrated in strong daylight environments, and is usually set to 0.10. If the brightness coverage ratio of the current frame is lower than 0.10 (which actually means that "the proportion of brighter pixels in this frame is less than 10%), it is determined that the current visible light image is greatly affected by insufficient illumination. The final judgment condition is that the fluctuation value of the edge feature of the infrared image is less than the corresponding fluctuation value of the visible light image, and the brightness coverage ratio of the current visible light image is lower than the set benchmark value of 0.10. If these two judgment conditions are met, the dominant control core channel type of the current image is marked.

[0027] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the type of the dominant control core channel, read the lens configuration parameter set associated with the current channel, retrieve the list of focal length indexes that can be matched collaboratively, filter the focal length range called by the current channel in combination with the channel identifier parameter, and extract the number parameter with the focal length linkage control identifier from it to generate a set of matching lens focal length numbers. First, the lens configuration parameter set bound to this channel is read from the control unit. The read parameters include the lens's minimum and maximum focal length, current focal length, supported zoom levels, focal length numbers and control codes, and whether the lens has control markers for linkage with other channels. By calling the upper and lower focal length limit information from these parameters, the standard focal length index list registered in the lens control parameter table is traversed. The current focal length value is compared to see if it falls within a certain standard focal length range. This standard focal length range is set in 10mm increments, such as 35mm, 45mm, and 55mm. Based on this, it is determined whether the current lens focal length matches the standard focal length index. If the focal length is 53mm, it is divided into a 55mm segment and marked as a match, using the current channel identifier... The parameters determine whether the channel is the dominant channel. If it is the dominant channel, it allows further filtering of focal length numbers with linkage control identifiers. In the settings, the number parameters with focal length linkage capability usually have linkage control flag bits. A flag bit of "1" indicates that linkage is available, and "0" indicates that it is not available. By reading the control flag bits of the focal length index one by one and removing the focal length numbers marked with "0", only the focal length numbers with linkage control capability are retained. From this, the final set of focal length numbers that can be used for linkage control operations in the current channel is generated. For example, if the current lens configuration focal length index list is [35mm, 50mm, 70mm, 105mm], where 50mm and 105mm have control flag bits of "1", the final set of matching lens focal length numbers is [50mm, 105mm].

[0028] S202: Call the set of matching lens focal length numbers, access the displacement information of the optical axis correction lens group built into the structure, extract the direction mark vector of the lens assembly, retrieve the correction mark corresponding to the main channel focal length, and obtain the offset direction vector of the optical axis of the secondary channel by comparing the direction relationship between the mark and the initial attitude of the optical axis of the secondary channel. The structural configuration is loaded, and the displacement information of the optical axis correction lens group is saved. This information includes the current position of each lens component, the component's movement history, displacement deviation data in each direction, and the reference optical axis correction mark of the lens group at a specific focal length. By reading the lens component orientation mark vector at each focal length, the orientation mark vector corresponding to the current focal length of the main channel is obtained. The correction mark is extracted by combining the spatial positioning information of the lens body. This mark records the offset direction between the optical axis and the theoretical central axis, which is used to characterize the pointing relationship of the lens components. Furthermore, the orientation vector corresponding to the initial optical axis attitude is obtained from the lens configuration of the secondary channel, and the orientation marks of the main channel and the secondary channel are established. By comparing the values ​​of the direction vectors along the X, Y, and Z axes, the magnitude and polarity of the offset difference in each direction are analyzed. A positive offset value indicates an outward offset, while a negative offset value indicates an inward offset. A significant offset threshold of 0.02 units of length is set. This threshold is determined by the accuracy range of the calibrated measuring instrument and is typically set to 2% of the lens field of view at the millimeter level. For example, if the field of view is 100mm, the threshold is 2mm. Finally, the offset direction vector of the optical axis of the slave channel is constructed based on the numerical difference between the direction markers of the master and slave channels. This vector will be used in the subsequent direction control process.

[0029] S203: Based on the optical axis offset direction vector of the channel, combined with the initial angle value between the center lines of the main channel and the channel, continuously calculate the change in the center line angle under the differentiated focal length number, extract the main direction component in the change trend, and after performing consistency correction on the sign of the main direction component, establish the reconstructed direction pointing angle. The initial angle between the centerlines of the main and secondary channels, measured at the factory during the equipment's manufacturing process, is read from the lens initialization calibration table. This value is set according to the equipment's alignment accuracy, typically between 1 and 5 degrees, and is determined based on the accuracy level. For example, in industrial-grade lenses, the initial angle is set to 3.5 degrees. Subsequently, each lens focal length number is sequentially read from the set of compatible lens focal length numbers, and the centerline angle is recalculated for each focal length number. The direction vectors of the main and secondary channels at the corresponding focal length are read, and the angle is calculated by recording the changes in these direction vectors. By continuously calculating the difference between the angle value corresponding to each focal length and the initial value, the amount of angle change caused by the focal length change is obtained. This is then processed according to the focal length. The difference in the included angle is recorded sequentially and a trend vector is constructed. In this trend vector, the change components in the three axes (X, Y, Z) are extracted one by one, and their change magnitude is judged. The threshold of the main direction component is set to 0.5 degrees. When the included angle change value in a certain direction is continuously greater than or less than the threshold and occupies more than 60% of the consecutive change times, it is determined to be the main direction component. The sign of the component is judged and a normalization operation is performed. That is, if the direction change is positive, it is marked as "+1" and if it is negative, it is marked as "-1". The magnitude of the value is ignored and only the directional information is retained. The finally generated normalized direction label will be used to construct the reconstructed direction pointing angle of the current frame.

[0030] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the direction of reconstruction, extract the imaging view range of the main channel image frame, and calibrate the boundary coordinate group of the image edge under the current view. Obtain the horizontal and vertical borders within the image range from the channel image frame, compare them with the boundary coordinate group of the main channel, calculate the root mean square of the coordinate difference between corresponding edge points between the borders, and obtain the image border matching deviation value. First, the current image frame captured by the main channel is retrieved. Based on the current focal length of the lens and the physical size of the image sensor, the actual imaging angle range of the image in the current state is calculated. This angle range is represented by a rectangular bounding box. Next, the lens calibration parameters and lens extrinsic parameter matrix are retrieved to calculate the boundary coordinate set from the four corners of the image. This includes the positions of the four edge points (upper left, upper right, lower left, and lower right) in the pixel coordinate system and is recorded as the main channel boundary reference coordinate set. For example, at a resolution of 1920×1080, the boundary coordinates covered by the current imaging angle of the main channel can be recorded as [(48, 42), (1872, 42), (48, 1038), (1872, 1038)]. Then, the current image frame from the secondary channel is retrieved, and the frame image is processed by border analysis. The horizontal and vertical border edge positions of the complete imaging area are extracted, and the four corner coordinate values ​​are marked. By comparing the values ​​from the secondary channel, the boundary coordinates of the image frame are determined. The corner coordinates of the channel image borders are compared with the corresponding corner coordinates of the main channel boundary. The differences in the horizontal and vertical directions are calculated point by point. For example, if the top left corner of the channel is (55, 50) and the bottom right corner is (1865, 1025), these two points are compared with the top left and bottom right corners of the main channel, respectively. The offsets in the X and Y axes are recorded. This process is repeated for all four sets of corresponding corner points. The squared differences are averaged and then the square root is taken. Finally, a numerical index representing the degree of matching between the two image borders in the overall space is obtained, namely the image border matching deviation value. The smaller the value, the higher the degree of overlap between the two channel boundaries. For example, if the differences of the four corner points are 10, 7, 6, and 8 pixels in the X direction and 12, 10, 11, and 9 pixels in the Y direction, the final matching deviation value is the square root of the sum of the squares, which is approximately 10.5 pixels. This value is used for subsequent boundary adjustment judgment.

[0031] S302: Based on the image border matching deviation value, retrieve the status of the front zoom component under the main channel configuration, set the boundary coincidence error threshold as the boundary coordinate difference limit range, and verify whether the zoom component is called to adjust the focal length position by judging whether the current deviation value is lower than the boundary coincidence error threshold, and obtain the boundary coincidence adjustment status identifier. Entering the zoom component status verification, the current working status of the front zoom component is extracted from the main channel configuration parameters, including focal length encoding, zoom motor position feedback value, and zoom adjustable range. Based on this, a boundary overlap error threshold is set to determine whether a zoom action needs to be performed. This threshold is calculated based on the image sensor width and the main channel lens resolution. For example, if the sensor width is 1920 pixels, and the allowable boundary error is set to 0.6% of the image width, or approximately 11.5 pixels, then the boundary overlap error threshold is set to 11 pixels, rounded down to ensure error control. This value is used as the judgment benchmark. By comparing the current image border matching deviation value with the error threshold, if the matching deviation value is less than the threshold, it means that the boundary has coincided within an acceptable range, and the status flag is set to no adjustment is required. If the matching deviation value is greater than the threshold, the front zoom component needs to be triggered to perform a fine adjustment of the focal length position. The control motor is called to adjust the current focal length position. Each adjustment unit is set to the pixel angle of view adjustment amount corresponding to a focal length change of 0.5mm. The feedback sensor confirms whether the adjustment is completed. After one adjustment is completed, the adjustment status is recorded as valid. Finally, the boundary coincidence adjustment status flag is output. A status of "1" indicates that the adjustment has been performed, and a status of "0" indicates that no adjustment has been performed.

[0032] S303: Call the boundary coincidence adjustment status flag, control the aperture position from the channel according to the boundary coincidence status, and register and adjust the width of the boundary area formed after the field of view is clipped by synchronously adjusting the aperture edge opening angle. Combine the view coordinate group after the boundary adjustment to establish the view edge position of synchronous coverage. First, determine if the current state is "1", indicating whether the main channel has already performed a zoom adjustment operation at the boundary position. If it is "1", it means that the secondary channel needs to perform field of view registration and adjustment. Read the aperture opening angle parameter of the current secondary channel and extract the actual visible area boundary formed by the aperture limitation in the image frame. Set the opening angle step size to 0.25 degrees and the adjustment range to ±5 degrees in the aperture control unit. Analyze the alignment difference between the boundary position and the main channel boundary to calculate the angle value that should be contracted or expanded. For example, if the current horizontal field of view boundary of the secondary channel exceeds the main channel by 8 pixels, it needs to be contracted by 0.6 degrees to match the edge position of the main channel according to the imaging relationship. Control the aperture blades to perform a closing operation until the viewing angle matching difference is controlled within the set allowable range, such as an allowable error of ±3 pixels. After the operation is performed, obtain the boundary coordinates of the cropped image again, extract the four corner points and perform coordinate synchronization calibration with the main channel boundary. After confirming the overlap, record the cropped boundary formed by the aperture boundary at this time as the final effective image edge. Establish the synchronously covered viewing angle edge position as the final result of the current secondary channel and main channel viewing angle overlap control.

[0033] Please see Figure 5 The specific steps of S4 are as follows: S401: Based on the edge position of the synchronously covered viewpoint, trigger the extension command of the lens middle section structure, monitor the state of the optical path axis inside the channel, extract the angle difference between the current direction vector of the optical axis and the corresponding direction vector of the main channel, control the optical axis to move in the direction of movement through the advancement action of the structural components, so that the angle between the two vectors gradually approaches zero, and obtain the optical axis alignment offset. First, the viewpoint edge coordinates of the main channel and the secondary channel are read, and combined with the focal length and mounting pose parameters of the two channels, the spatial direction vector of the optical axis of the main channel is calculated. This vector is formed by the direction from the center point of the lens to the center point of the image plane. Then, the optical axis direction vector of the secondary channel is extracted. Both are measured by the angle sensor in the lens's internal attitude detection. According to the control logic of the lens's middle section structure, an extension command is issued to make the middle section propulsion component move along the axis in a fixed step manner. After each step, the optical axis direction vector of the secondary channel is re-acquired. The angle difference between the direction vectors of the main channel and the secondary channel is gradually compared, and this angle difference is recorded as the optical axis offset angle. If the offset angle exceeds the set threshold, the propulsion or retraction command is continued to be issued until the error converges to the allowable range. The optical axis angle error threshold is set based on the factory calibration results of the equipment. It usually depends on the lens focal length, image resolution, and optical center offset. For lenses with a focal length in the range of 50mm to 100mm, the optical axis angle error is allowed to be controlled between 0.2 and 0.3 degrees. When the current angle difference is detected to be 0.7 degrees, the control middle section structure is continuously advanced along the optical axis direction several times. Each time, the equivalent angle is adjusted in a step size of 0.05 degrees. The comparison operation is repeated until the angle difference is lower than 0.3 degrees. When the difference remains stable and the fluctuation is less than 0.05 degrees in three consecutive measurements, the advancement action is stopped. The optical axis alignment offset of the channel relative to the main channel at this moment is recorded and used as the optical axis alignment result.

[0034] S402: Call the optical axis alignment offset, extract the lens displacement data during the axial advancement of the lens group, and combine it with the difference in the normal vectors of the two channel image planes before and after advancement to calculate the image projection direction consistency error, determine whether the image plane normal angle is within the set image plane parallelism error tolerance range, and obtain the image plane alignment state parameters. Entering the lens assembly axial displacement monitoring stage, the displacement data of the lens assembly at various time points during the advancement process are read in real time. The starting position, ending position, and displacement data corresponding to each fine-tuning step of the lens assembly are recorded. By comparing the displacement information before and after advancement, and combining the normal directions of the imaging planes of the main channel and the slave channel, image plane normal vector sets for both are established. The angle difference between the two sets of normal vectors is extracted and angle analysis is performed to calculate the consistency error of the image projection direction. A tolerance range for image plane parallelism error is set during judgment. This range is determined based on the installation flatness of the imaging sensor and the assembly error of the lens assembly, and is usually controlled within 1 degree. For equipment with high structural precision, this tolerance can be reduced to 0.5 degrees. When the detected angle between the normals of the two channel image planes is less than this tolerance, the image plane parallelism is judged to be qualified; otherwise, it is considered misaligned and the lens advancement process needs to be returned for compensation and adjustment. For example, when the difference in the angle between the image planes is 0.6 degrees, which is within the set tolerance range of 1 degree, it is determined to be an acceptable state, and the currently measured error value is stored together with the lens advance record to generate the corresponding image plane alignment state parameters.

[0035] S403: Based on the image alignment state parameters, analyze the continuous change trend curve of the emission angle direction. If the emission direction remains stable and the rate of change is lower than the projection consistency judgment threshold, then verify that the current image has been fused and reconstructed, and perform frame-level synchronous fusion of the main channel and the slave channel image matrices to generate a master-slave image superimposed frame. First, within a continuous ten-frame imaging cycle, the angular changes of the emission directions of the two channels are recorded frame by frame. The angular difference between the emission direction vector of each frame and the previous frame is extracted to construct an emission direction change trend curve. The angular change values ​​between adjacent frames are compared, and the maximum change amplitude and average change rate within ten frames are statistically analyzed. A projection consistency judgment threshold is set to measure the stability of the emission direction. This threshold is set during calibration based on the optical imaging stability accuracy, typically 0.3 degrees. If the average change rate is less than 0.1 degrees and the maximum change amplitude does not exceed 0.3 degrees, the emission direction is considered stable. Simultaneously, the temporal stability parameter is monitored when judging the emission direction change trend. The angular change difference fluctuation of five consecutive frames is required to not exceed 0.05 degrees. When this condition is met, the projection direction is considered stable. Then, master-slave image synchronization is triggered, aligning the spatial position and optical axis direction of the two channel image frames. Frame-by-frame matrix data temporal synchronization is performed, and color mapping and pixel brightness equalization are conducted respectively, enabling the two channel images to spatially overlap in the same frame sequence, ultimately generating a master-slave image overlay.

[0036] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the overlay frame of the master and slave images, obtain the image frame sequence within the current continuous imaging cycle, extract the coordinates of the boundary points at the upper left and lower right corners of each frame, calculate the coordinate difference of the same boundary point in adjacent frames, and perform moving average processing on the difference sequence to obtain the trend of boundary point displacement. First, during continuous zoom operation, a sequence of image frames is acquired over a period of time. The imaging cycle is set to 10 consecutive frames. The coordinates of the top-left and bottom-right boundary points in each frame are extracted. These coordinates are obtained using image frame metadata and pixel localization to ensure accurate pixel numbering. Then, pixel-level coordinate difference calculations are performed on the same boundary points in adjacent frames. Specifically, the X-axis and Y-axis differences of the top-left corner point between frame n and frame (n+1) are denoted as ΔXn and ΔYn, respectively. The same operation is performed on the bottom-right corner point, forming a sequence of boundary coordinate differences in two directions. For example, the top-left X-coordinate difference sequence... The sequence is [3, -2, 1, 0, -1, 2, -3, 1, 0]. Based on this difference sequence, a moving average is performed to reduce the impact of instantaneous fluctuations. The window width is set to 3 frames, that is, the data of the previous frame and the current frame are averaged at each time to generate a smoothed displacement change trend sequence. The calculation is performed independently in the X and Y directions to obtain the smooth displacement trajectory change trend of the two points at the top left and bottom right corners in the continuous imaging cycle. This trend is used to reflect the boundary stability and position drift of the image during zooming. Finally, the trend calculation is completed and the boundary point displacement change trend result is output.

[0037] S502: Based on the trend of boundary point displacement change, extract the coordinate set of edge trajectory lines in the corresponding frame sequence, align the edge trajectory lines in the main channel and the secondary channel image frames respectively, calculate the overlap rate index between the corresponding trajectory lines, and calculate the average area of ​​the overlapping area of ​​trajectory lines in all frames to obtain the edge trajectory line overlap degree value. The process then proceeds to edge trajectory processing. The coordinate set of the outer edge trajectory lines for each frame in the image frame sequence is extracted sequentially. These edge trajectory lines indicate the spatial boundaries of the image's visible range. Using the trajectory lines extracted from the main channel image frames as a reference, the edge trajectory lines of the secondary channel image frames are matched frame by frame. The spatial relationship between corresponding trajectories is identified by comparing pixel positions. A direct coordinate comparison method is used to calculate the overlap rate between the main and secondary channel trajectory lines in each frame. The overlap rate is defined as the proportion of pixels in the secondary channel trajectory line that completely overlap with the main channel trajectory line out of the total number of trajectory line pixels. For example, if a secondary channel trajectory line contains 100 pixels in a frame, and 87 of these pixels overlap with the main channel trajectory line at the same position, then the overlap rate is calculated. The overlap rate is 87%. This process is repeated for consecutive frame sequences to record the trajectory overlap rate of all frames and calculate its average value. If the average value is higher than the set threshold, it is considered that the alignment is stable. The trajectory overlap rate threshold is set based on actual fusion requirements and is usually set to 80%. Based on the trajectory line overlap rate calculation, the overlapping area between the master and slave channel trajectory lines in each frame image is further calculated. The overlapping area values ​​in all frames are added together and divided by the number of frames to obtain the average overlapping area value. For example, the overlapping areas of ten frames are 950, 980, 1020, 1005, etc. The average value is used to help judge the spatial consistency of the trajectory lines. Finally, the average overlap rate and the average area value are combined together to output the edge trajectory line overlap degree value.

[0038] S503: Call the edge trajectory line overlap value, combine it with the boundary point coordinate sequence of the master and slave channels in the current continuous zoom operation segment, determine whether there is an offset trajectory of the channel boundary that exceeds the image edge. If all boundary trajectories remain stable in the overlap area, they are marked as fusionable intervals, and a joint zoom scheme is established. Further stability assessment is performed by combining the boundary point coordinate sequences of the main and secondary channels within the current continuous zoom segment. First, the coordinate change data of the upper left and lower right corners in consecutive image frames are extracted to establish the movement trajectory of each boundary point in the time dimension. Each point is then compared to check for deviations from the edge of the main channel image. Boundary trajectory offset is defined as the boundary point coordinates in a frame exceeding the defined range of the image size. For example, if the image width is 1920 pixels and the height is 1080 pixels, if the X-coordinate of the lower right corner point in a frame is greater than 1920 or the Y-coordinate is greater than 1080, then that point is considered to have offset. This process is repeated for each frame. If all boundary points meet the condition, and all boundary points remain within the image edge frame in all frames, and the overlap value reaches the preset standard, then it is further determined whether the boundary points remain stable within the overlap area. The trajectory stability judgment threshold is set to ±5 pixels, that is, the change in boundary point coordinates in consecutive frames shall not exceed 5 pixels. If this condition is met and all boundary trajectory points are within the overlap area, the zoom segment is marked as a fusionable interval, and a joint zoom scheme is generated in the control logic. This scheme will be synchronously called in subsequent zoom adjustment tasks to unify the linkage control and field of view coordination of the main channel and the slave channel lenses.

[0039] Please see Figure 7 A combined zoom system, comprising: The channel selection module is used to implement S1: after obtaining the zoom command, it retrieves the image sharpness and texture area of ​​the infrared lens, and determines the illumination state by comparing it with the outline and brightness of the visible light image, and selects the dominant control core channel type. The optical axis correction module is used to achieve S2: based on the type of the dominant control core channel, read the lens focal length number and identify the optical axis offset direction, and generate the reconstruction direction pointing angle by calculating the change in the angle between the center lines of the two channels; The viewpoint alignment module is used to implement S3: based on the angle of the reconstruction direction, it detects the match between the viewpoint edge and the main channel, determines whether the focal length boundary is overlapped through the zoom component, adjusts the aperture trimming boundary width, and generates the viewpoint edge position that is synchronously covered. The image fusion module is used to implement S4: based on the edge position of the synchronously covered viewpoint, it triggers the optical path extension, adjusts the viewing axis to the main channel direction, verifies that the two channel image planes are projected in the same way, determines whether the image is fused and reconstructed, and generates a master-slave image superimposed frame. The zoom judgment module is used to implement S5: based on the superimposed frame of the master and slave images, analyze the displacement trend of boundary points, compare the overlap of edge trajectories, and if no offset or boundary-crossing trajectory occurs, form a joint zoom scheme.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A combined zoom method, characterized in that, Includes the following steps: S1: After obtaining the zoom command, retrieve the image clarity and texture area of ​​the infrared lens, and determine the illumination status by comparing it with the outline and brightness of the visible light image, and select the dominant control core channel type. S2: Based on the dominant control core channel type, read the lens focal length number and identify the optical axis offset direction. By calculating the change in the angle between the center lines of the two channels, generate the reconstructed direction pointing angle. S3: Based on the reconstructed direction pointing angle, detect the matching of the viewpoint edge with the main channel, determine whether the focal length boundary is overlapped through the zoom component, adjust the aperture to trim the boundary width, and generate the viewpoint edge position that is synchronously covered. S4: Based on the edge position of the synchronously covered viewpoint, trigger the optical path extension, adjust the viewing axis to the main channel direction, verify that the two channel image planes are projected in the same way, determine whether the image is fused and reconstructed, and generate a master-slave image superimposed frame. S5: Based on the superimposed frame size of the master and slave images, analyze the displacement trend of the boundary points, compare the overlap of the edge trajectories, and if no offset or boundary-crossing trajectory occurs, form a joint zoom scheme.

2. The combined zoom method according to claim 1, characterized in that, The dominant control core channel type includes image sharpness criteria, texture area index, contour distribution density, and brightness distribution ratio. The reconstruction direction pointing angle includes focal length corresponding number, optical axis offset azimuth, and channel center angle change. The synchronously covered viewpoint edge position includes master channel edge coordinates, slave channel boundary matching degree, and edge cropping width value. The master-slave image superposition frame includes image plane projection direction consistency, image fusion integrity, and channel exit angle synergy. The joint zoom scheme includes boundary point displacement trend line, image edge trajectory overlap, and continuous zoom segment stability judgment result.

3. The combined zoom method according to claim 1, characterized in that, The boundary point displacement trend refers to the change pattern of the positional movement direction and magnitude of corresponding points at the edges of the master and slave images during image zooming, as time and focal length change.

4. The combined zoom method according to claim 1, characterized in that, The offset out-of-bounds trajectory refers to the motion path in which the image edge trajectory deviates from the main channel's viewpoint range and exceeds the overlap error range during the boundary point displacement process.

5. The combined zoom method according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: After obtaining the zoom command from the control end, call the infrared lens image frame, extract the gray-scale gradient and texture distribution of the edge region, calculate the gradient standard deviation of adjacent regions and count the area pixels of continuous texture regions, and generate the edge sharpness and continuous texture area value of the infrared image. S102: Based on the edge sharpness and continuous texture area value of the infrared image, obtain the visible light image frame, extract the contour region to calculate the edge line density and brightness pixel ratio of the unit pixel, and combine the contour line density judgment threshold and the brightness coverage ratio judgment threshold to make a joint judgment and generate the current image contour density and brightness coverage comparison factor. S103: Call the infrared image edge sharpness and continuous texture area value and the current image contour density and brightness coverage comparison factor to perform a difference comparison analysis on the image feature stability of infrared and visible light images. If the gradient standard deviation of the infrared image is lower than that of the visible light image and the corresponding brightness coverage ratio is lower than the edge threshold benchmark, generate the dominant control core channel type.

6. The combined zoom method according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the dominant control core channel type, read the lens configuration parameter set associated with the current channel, retrieve the list of focal length indexes that can be matched collaboratively, filter the focal length range called by the current channel in combination with the channel identifier parameter, and extract the number parameter with the focal length linkage control identifier from it to generate a set of matching lens focal length numbers. S202: Call the set of matching lens focal length numbers, access the displacement information of the optical axis correction lens group built into the structure, extract the direction mark vector of the lens assembly, retrieve the correction mark corresponding to the main channel focal length, and obtain the offset direction vector of the optical axis of the secondary channel by comparing the direction relationship between the mark and the initial attitude of the optical axis of the secondary channel. S203: Based on the optical axis offset direction vector of the secondary channel, combined with the initial angle value between the center lines of the primary channel and the secondary channel, continuously calculate the change in the center line angle under the differentiated focal length number, extract the primary direction component in the change trend, and after performing consistency correction on the sign of the primary direction component, establish the reconstructed direction pointing angle.

7. The combined zoom method according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Based on the reconstructed direction pointing angle, extract the imaging view range of the main channel image frame, and calibrate the boundary coordinate group of the image edge under the current view. Obtain the horizontal and vertical borders within the image range from the channel image frame, compare them with the main channel boundary coordinate group, calculate the root mean square of the coordinate difference between corresponding edge points between the borders, and obtain the image border matching deviation value. S302: Based on the image border matching deviation value, retrieve the status of the front zoom component under the main channel configuration, set the boundary coincidence error threshold as the boundary coordinate difference limit range, and verify whether the zoom component is called to adjust the focal length position by judging whether the current deviation value is lower than the boundary coincidence error threshold, and obtain the boundary coincidence adjustment status identifier. S303: Invoke the boundary coincidence adjustment status flag, control the channel aperture position according to the boundary coincidence status, and adjust the width of the boundary area formed after the field of view is clipped by synchronously adjusting the aperture edge opening angle. Combine the view coordinate group after boundary adjustment to establish the view edge position of synchronous coverage.

8. The combined zoom method according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the edge position of the synchronously covered viewpoint, trigger the extension command of the lens middle section structure, monitor the state of the optical path axis inside the channel, extract the angle difference between the current direction vector of the optical axis and the corresponding direction vector of the main channel, control the optical axis to move in the direction of movement through the advancement action of the structural components, so that the angle between the two vectors gradually approaches zero, and obtain the optical axis alignment offset. S402: Call the optical axis alignment offset, extract the lens displacement data of the lens group during the axial advancement process, and combine the difference of the normal vectors of the two channel image planes before and after advancement to calculate the image projection direction consistency error, determine whether the image plane normal angle is within the set image plane parallelism error tolerance range, and obtain the image plane alignment state parameters. S403: Based on the image alignment state parameters, analyze the continuous change trend curve of the emission angle direction. If the emission direction remains stable and the rate of change is lower than the projection consistency judgment threshold, verify that the current image has been fused and reconstructed, and perform frame-level synchronous fusion of the main channel and the slave channel image matrices to generate a master-slave image superimposed frame.

9. The combined zoom method according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the superimposed frame of the master and slave images, obtain the image frame sequence within the current continuous imaging cycle, extract the coordinates of the boundary points at the upper left and lower right corners of each frame, calculate the coordinate difference of the same boundary point in adjacent frames, and perform moving average processing on the difference sequence to obtain the trend of boundary point displacement. S502: Based on the trend of boundary point displacement change, extract the coordinate set of edge trajectory lines in the corresponding frame sequence, perform alignment processing on the edge trajectory lines in the main channel and the secondary channel image frames respectively, calculate the overlap rate index between the corresponding trajectory lines, and calculate the average area of ​​the overlapping area of ​​trajectory lines in all frames to obtain the edge trajectory line overlap degree value. S503: Call the edge trajectory line overlap value, combine it with the boundary point coordinate sequence of the master and slave channels in the current continuous zoom operation segment, determine whether there is an offset trajectory of the channel boundary that exceeds the image edge. If all boundary trajectories remain stable in the overlap area, they are marked as fusionable intervals, and a joint zoom scheme is established.

10. A combined zoom system, characterized in that, The system is used to implement the combined zoom method according to any one of claims 1-9, the system comprising: The channel selection module is used to implement S1: after obtaining the zoom command, it retrieves the image sharpness and texture area of ​​the infrared lens, and determines the illumination state by comparing it with the outline and brightness of the visible light image, and selects the dominant control core channel type. The optical axis correction module is used to achieve S2: based on the type of the dominant control core channel, read the lens focal length number and identify the optical axis offset direction, and generate the reconstruction direction pointing angle by calculating the change in the angle between the center lines of the two channels; The viewpoint alignment module is used to implement S3: based on the pointing angle of the reconstruction direction, detect the match between the viewpoint edge and the main channel, determine whether the focal length boundary is overlapped through the zoom component, adjust the aperture trimming boundary width, and generate the viewpoint edge position that is synchronously covered. The image fusion module is used to implement S4: based on the position of the viewing angle edge of the synchronous coverage, trigger the optical path extension, adjust the viewing axis to the main channel direction, verify that the image planes of the two channels are projected in the same way, determine whether the image is fused and reconstructed, and generate a master-slave image superimposed frame. The zoom judgment module is used to implement S5: based on the frame size of the master and slave image superimposed image, analyze the displacement trend of the boundary points, compare the overlap of the edge trajectories, and if no offset or boundary-crossing trajectory occurs, form a joint zoom scheme.

Images