A multi-optical axis optical imaging infrared night vision device and control method thereof

Abstract

The present invention relates to the technical field of monitoring equipment, in particular to a multi-optical axis optical imaging infrared night vision device and a control method thereof. The real-time image data of each optical axis are fused to generate a fused real-time imaging image; if the image entropy of the real-time imaging image is greater than a preset threshold, the real-time imaging parameters of the infrared night vision device are not regulated; if the image entropy of the real-time imaging image is not greater than the preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree index spectrum matching, and the real-time imaging parameters of the infrared night vision device are regulated according to the preset correction measure. The imaging quality, stability and adaptability of the infrared night vision device in monitoring work are effectively improved by the present invention.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring equipment, in particular to a multi-optical axis optical imaging infrared night vision equipment and a control method thereof. Background Art

[0002] Infrared night vision technology is an important technical means in modern military reconnaissance, security monitoring, industrial detection and other fields. With the continuous development of science and technology, the requirements for infrared night vision equipment are getting higher and higher, especially in terms of imaging quality and stability in complex environments. Multi-axis optical imaging infrared night vision equipment can achieve wider field of view coverage, longer detection distance and higher image resolution by integrating multiple optical channels. However, the introduction of multi-axis systems also brings new challenges. First, image registration and fusion between multiple optical axes are the key to achieving high-quality imaging, which requires precise control and calibration. Secondly, due to environmental factors (such as temperature changes, mechanical vibrations, etc.), imaging parameters may shift, resulting in reduced image quality or even failure to image normally. Therefore, how to achieve effective control of multi-axis infrared night vision equipment to ensure that it can work stably and output high-quality images in various complex environments has become a technical problem that needs to be solved urgently.

[0003] In addition, the existing control methods of infrared night vision equipment mainly focus on the adjustment of single imaging parameters, such as focal length, exposure time, etc., and lack the optimization and control of the overall performance of the multi-axis system. For the correction of imaging parameter offset, manual adjustment or simple feedback control is usually used, which is inefficient and difficult to ensure real-time performance and accuracy. In order to solve these problems, a more intelligent and adaptive control method is needed to comprehensively control and optimize multi-axis infrared night vision equipment to ensure that it can work stably and output high-quality images in various complex environments. Summary of the invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a multi-optical axis optical imaging infrared night vision device and a control method thereof.

[0005] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is:

[0006] The invention discloses a control method for a multi-optical axis optical imaging infrared night vision device, comprising the following steps:

[0007] Obtaining a prefabricated control scheme for an infrared night vision device, and evaluating and optimizing the prefabricated control scheme to obtain a standard control scheme for the infrared night vision device and a standard working parameter set for each module when performing a preset working task;

[0008] Controlling the infrared night vision device to perform monitoring work according to the standard control scheme, and acquiring the real-time image data of each optical axis through the image sensor at a preset time point, fusing the real-time image data of each optical axis, and generating a fused real-time imaging image;

[0009] Calculating the image entropy of the real-time imaging image; if the image entropy of the real-time imaging image is greater than a preset threshold, not adjusting the real-time imaging parameters of the infrared night vision device;

[0010] Preset deviation correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the prefabricated infrared night vision device, and constructing a B-tree-index map according to the preset deviation correction measures corresponding to various imaging parameter offset values;

[0011] If the image entropy of the real-time imaging image is not greater than a preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure.

[0012] Furthermore, a prefabricated control scheme of the infrared night vision device is obtained, and the prefabricated control scheme is evaluated and optimized to obtain a standard control scheme of the infrared night vision device and a standard working parameter set of each module when performing a preset working task, specifically:

[0013] Obtaining a structural model diagram of the infrared night vision device based on the structural parameter specification of the infrared night vision device, obtaining a scene feature model diagram within the working area of ​​the infrared night vision device based on remote sensing technology, and obtaining environmental feature parameters of the infrared night vision device within a preset working time period based on meteorological software;

[0014] According to the structural model diagram, the scene feature model diagram and the environmental feature parameters, a simulation imaging system of an infrared night vision device is constructed using simulation software;

[0015] Obtaining a prefabricated control scheme for an infrared night vision device, importing the prefabricated control scheme into the infrared night vision device simulation imaging system for simulation imaging analysis, and obtaining a simulated imaging image obtained after the infrared night vision device is simulated controlled based on the prefabricated control scheme;

[0016] Calculating the image entropy of the simulated imaging image, and comparing the image entropy of the simulated imaging image with a preset threshold;

[0017] If the image entropy of the simulated imaging image is greater than a preset threshold, the prefabricated control scheme is used as a standard control scheme for the infrared night vision device.

[0018] Furthermore, a prefabricated control scheme of the infrared night vision device is obtained, and the prefabricated control scheme is evaluated and optimized to obtain a standard control scheme of the infrared night vision device and a standard working parameter set of each module when performing a preset working task, and the following steps are also included:

[0019] If the image entropy of the simulated imaging image is not greater than a preset threshold, the prefabricated control scheme is optimized and iteratively processed using a particle swarm optimization algorithm to obtain an iteratively optimized prefabricated control scheme;

[0020] The prefabricated control scheme after iterative optimization is introduced into the infrared night vision device simulation imaging system for simulation imaging analysis, and after each iterative optimization cycle is completed, a simulation imaging image is obtained, and the image entropy of the simulation imaging image is calculated; the image entropy of the simulation imaging image is compared with a preset threshold;

[0021] If the image entropy of the simulated imaging image is still not greater than the preset threshold, the next iterative optimization cycle is executed;

[0022] When the image entropy of the simulated imaging image of the iteratively optimized prefabricated control scheme is greater than a preset threshold, the iteration is stopped, and the prefabricated control scheme after the last iterative optimization is used as the standard control scheme of the infrared night vision device;

[0023] After the standard control scheme is obtained, the standard working parameters corresponding to each time node of each module in the infrared night vision device when the infrared night vision device is simulated controlled according to the standard control scheme are obtained in the infrared night vision device simulation imaging system, and the standard working parameter set of each module in the infrared night vision device when performing the preset working task is obtained;

[0024] The standard working parameter set includes the standard working parameters corresponding to each working time node when the corresponding module executes the preset working task.

[0025] Furthermore, the real-time image data of each optical axis is fused to generate a fused real-time imaging image, specifically:

[0026] Acquire real-time image data of each optical axis, and combine the real-time image data of each optical axis into a combination matrix of image data;

[0027] Centralizing the combined matrix, and calculating the covariance matrix of the combined matrix after the centralization, wherein the covariance matrix can reflect the correlation between different image data;

[0028] Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix. The eigenvalues ​​represent the variance of the data in the direction of the corresponding eigenvectors, and the eigenvectors determine the direction of the data in the new space.

[0029] Sort the eigenvectors according to the size of the eigenvalues, and select the eigenvectors of the first k eigenvalues ​​as the selected eigenvectors;

[0030] Construct a transformation matrix according to the selected eigenvectors, perform linear transformation on the combination matrix of the original image data through the transformation matrix to obtain a number of principal components; construct a principal component matrix according to the obtained principal components;

[0031] The principal component matrix is ​​inversely transformed to obtain the original spatial form of the image data, thereby obtaining a fused real-time imaging image.

[0032] Furthermore, preset correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the prefabricated infrared night vision device are constructed according to the preset correction measures corresponding to the various imaging parameter offset values. Specifically, the B-tree-index map is:

[0033] Preset deviation correction measures corresponding to various imaging parameter deviation values ​​of the imaging module in the prefabricated infrared night vision device;

[0034] Construct a B-tree and initialize the root node and preset splitting rules of the B-tree; use the offset value of each imaging parameter as a data item; perform feature extraction processing on each data item to obtain a unique identification value of each data item;

[0035] Starting from the root node of the B-tree, compare the data item identification value with the key value of the current node; if the identification value is less than the minimum key value of the current node, search downward along the leftmost subtree; if it is greater than the maximum key value of the current node, search downward along the rightmost subtree; otherwise, find a suitable subtree within the key value range of the current node and search downward;

[0036] When a leaf node is reached, the preset position is located in the leaf node and the data item is inserted; if the leaf node is full, the leaf node is split into two nodes according to the splitting rule of the B-tree, and the key values ​​and pointers of the parent node and higher-level nodes are adjusted;

[0037] Perform split iteration according to the above steps to store each data item in order in each leaf node of the B-tree until there are no other data items in each leaf node, then stop split iteration;

[0038] The preset correction measures corresponding to various imaging parameter offset values ​​are respectively stored in the leaf nodes to which the corresponding data items belong, and a B-tree-index map is obtained.

[0039] Furthermore, if the image entropy of the real-time imaging image is not greater than a preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure, specifically:

[0040] If the image entropy of the real-time imaging image is not greater than a preset threshold, obtaining the real-time imaging parameters of the imaging module in the infrared night vision device at the current working time node;

[0041] and obtaining a standard working parameter set of an imaging module in an infrared night vision device when performing a preset working task, and obtaining standard working parameters of the imaging module at a current working time node in the standard working parameter set; wherein the standard working parameters are standard imaging parameters;

[0042] Calculating the difference between the standard working parameter of the imaging module at the current working time node and the real-time imaging parameter to obtain an imaging parameter deviation value, and acquiring an identification value of the imaging parameter deviation value;

[0043] Importing the identification value of the imaging parameter deviation value into the B-tree-index map for matching, matching to obtain leaf nodes with the same identification items, and extracting preset deviation correction measures from the matched leaf nodes;

[0044] The real-time imaging parameters of the infrared night vision device are adjusted and processed according to the extracted preset correction measures.

[0045] The present invention also discloses a multi-axis optical imaging infrared night vision device, which is applied to a control method of a multi-axis optical imaging infrared night vision device as described in any one of the above items. The infrared night vision device comprises:

[0046] A support frame and a night vision device body, wherein the night vision device body and the support frame are connected via an angle adjuster;

[0047] An optical module is arranged in the main body of the night vision device, and the optical module includes a plurality of optical axes, a lens group and a reflector, and is used to collect and focus infrared light;

[0048] The main body of the night vision device is provided with a detection module for converting infrared light into electrical signals;

[0049] An imaging module is provided in the main body of the night vision device to collect and analyze image signals to generate images;

[0050] A display module is provided in the main body of the night vision device, for displaying the generated image;

[0051] A control module is provided in the main body of the night vision device, which is used to control the working parameters of each module;

[0052] A signal transmission module is arranged in the main body of the night vision device, which is used to send the collected image signal to a preset terminal.

[0053] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: by evaluating and optimizing the prefabricated control scheme, a standard control scheme and a standard working parameter set are determined, thereby providing a good start for the operation of the equipment; multi-optical axis image fusion improves the comprehensiveness and accuracy of imaging, and the imaging quality is judged based on image entropy to decide whether to adjust the imaging parameters; a B-tree-index map is constructed to store the corresponding relationship between the imaging parameter offset value and the preset correction measures, which can quickly match the correction measures according to the imaging parameter deviation value and adjust the real-time imaging parameters, thereby overall improving the imaging quality, stability and adaptability of the infrared night vision equipment in monitoring work. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0055] Figure 1 This is a flow chart of the first method of the infrared night vision device control method;

[0056] Figure 2 This is a flow chart of the second method of the infrared night vision device control method;

[0057] Figure 3 This is a simplified schematic diagram of the structure of the infrared night vision device. DETAILED DESCRIPTION

[0058] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0060] like Figure 1The present invention discloses a control method for a multi-optical axis optical imaging infrared night vision device, comprising the following steps:

[0061] S102, obtaining a prefabricated control scheme for an infrared night vision device, and evaluating and optimizing the prefabricated control scheme to obtain a standard control scheme for the infrared night vision device and a standard working parameter set for each module when performing a preset working task;

[0062] S104, controlling the infrared night vision device to perform monitoring work according to the standard control scheme, and obtaining real-time image data of each optical axis through an image sensor at a preset time point, fusing the real-time image data of each optical axis, and generating a fused real-time imaging image;

[0063] S106, calculating the image entropy of the real-time imaging image; if the image entropy of the real-time imaging image is greater than a preset threshold, no adjustment or processing is performed on the real-time imaging parameters of the infrared night vision device;

[0064] S108, prefabricating preset correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the infrared night vision device, and constructing a B-tree-index map according to the preset correction measures corresponding to the various imaging parameter offset values;

[0065] S110. If the image entropy of the real-time imaging image is not greater than a preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure.

[0066] It should be noted that the control method of the multi-axis optical imaging infrared night vision device determines the standard control scheme and the standard working parameter set by evaluating and optimizing the prefabricated control scheme, providing a good start for the operation of the equipment; multi-axis image fusion improves the comprehensiveness and accuracy of imaging, and judges the imaging quality based on image entropy to decide whether to adjust the imaging parameters; constructs a B-tree-index map to store the correspondence between the imaging parameter offset value and the preset correction measures, which can quickly match the correction measures according to the imaging parameter deviation value and adjust the real-time imaging parameters, thereby overall improving the imaging quality, stability and adaptability of the infrared night vision equipment in monitoring work.

[0067] Furthermore, a prefabricated control scheme of the infrared night vision device is obtained, and the prefabricated control scheme is evaluated and optimized to obtain a standard control scheme of the infrared night vision device and a standard working parameter set of each module when performing a preset working task, such as Figure 2 As shown, specifically:

[0068] S202, obtaining a structural model diagram of the infrared night vision device based on the structural parameter specification of the infrared night vision device, obtaining a scene feature model diagram within the working area of ​​the infrared night vision device based on remote sensing technology, and obtaining environmental feature parameters of the infrared night vision device within a preset working time period based on meteorological software;

[0069] S204, constructing an infrared night vision device simulation imaging system according to the structural model diagram, the scene feature model diagram and the environmental feature parameters using simulation software;

[0070] S206, obtaining a prefabricated control scheme for an infrared night vision device, importing the prefabricated control scheme into the infrared night vision device simulation imaging system for simulation imaging analysis, and obtaining a simulated imaging image obtained after the infrared night vision device is simulated controlled based on the prefabricated control scheme;

[0071] S208, calculating the image entropy of the simulated imaging image, and comparing the image entropy of the simulated imaging image with a preset threshold;

[0072] S210: If the image entropy of the simulated imaging image is greater than a preset threshold, the prefabricated control scheme is used as a standard control scheme for infrared night vision equipment.

[0073] Among them, the prefabricated control scheme refers to a set of control strategies and parameter combinations that are pre-set before the actual operation of the infrared night vision device. It covers the control instructions and parameter settings for each module of the infrared night vision device (such as optical module, detection module, imaging module, etc.), aiming to enable the device to operate in the expected manner and realize the imaging function. For example, there are pre-set values ​​for the direction of the optical axis in the optical module, the focusing parameters of the lens group, the sensitivity setting of the detection module, and the image acquisition frequency of the imaging module. This scheme is formulated by relevant technical personnel based on comprehensive consideration of multiple factors such as the functional requirements of the equipment, possible working scenarios, and relevant technical experience. Therefore, in order to eliminate human experience errors, it is necessary to undergo a subsequent evaluation and optimization process based on the simulation imaging system to determine whether it can meet the imaging quality requirements of the equipment in actual work, and finally become the standard control scheme of the equipment.

[0074] It should be noted that the structural model diagram is first obtained from the structural parameter specification of the infrared night vision device, which provides a basis for understanding the structural characteristics of the device itself. For example, the layout of each module inside the device, the relative position relationship between the optical axes and other structural information are all included in the structural model diagram. The scene feature model diagram in the working area is obtained through remote sensing technology, which can reflect the terrain, landform, distribution of objects and other characteristics of the scene where the device will work. Different scenes (such as mountainous areas, plains, cities, etc.) will have different reflections and scattering of infrared light. The scene feature model diagram helps to consider the impact of these factors on imaging in advance. The environmental characteristic parameters within the preset working time period are obtained using meteorological software. Environmental factors such as temperature, humidity, and atmospheric visibility have a great impact on the imaging effect of infrared night vision equipment. For example, high humidity will affect the propagation of infrared light, and temperature changes may cause changes in the performance of components inside the device. According to the structural model diagram, scene feature model diagram and environmental characteristic parameters obtained above, simulation software (such as Zemax, OptiFDTD and other software) is used to build an infrared night vision device simulation imaging system. This simulation system is a virtual test platform that can simulate the working state of infrared night vision equipment in a real environment. Import the prefabricated control scheme into the simulation imaging system for simulation imaging analysis to obtain a simulated imaging image. Then calculate the image entropy of the simulated imaging image. Image entropy is an indicator to measure the amount of image information. The larger the image entropy, the richer the information contained in the image, and the higher the image quality may be. Compare the image entropy of the simulated imaging image with the preset threshold. If it is greater than the preset threshold, it means that the quality of the simulated imaging image obtained under the prefabricated control scheme is good, and this prefabricated control scheme can be used as the standard control scheme for infrared night vision equipment. This standard control scheme and the standard working parameter set when each module performs the preset work task under this scheme can provide an optimized control and working parameter reference for the equipment in actual work, ensuring that the equipment can achieve better imaging effects in actual operation. Through the above series of operations, the structure, working scene and environmental factors of the infrared night vision equipment itself can be comprehensively considered, and the prefabricated control scheme can be evaluated by using the simulation imaging system. The standard control scheme and standard working parameter set that can enable the equipment to achieve better imaging effects in actual work can be screened out using image entropy as a measurement indicator, thereby improving the imaging quality and work efficiency of the equipment in actual operation.

[0075] Furthermore, a prefabricated control scheme of the infrared night vision device is obtained, and the prefabricated control scheme is evaluated and optimized to obtain a standard control scheme of the infrared night vision device and a standard working parameter set of each module when performing a preset working task, and the following steps are also included:

[0076] If the image entropy of the simulated imaging image is not greater than a preset threshold, the prefabricated control scheme is optimized and iteratively processed using a particle swarm optimization algorithm to obtain an iteratively optimized prefabricated control scheme;

[0077] The method steps for optimizing and iterating the prefabricated control scheme using the particle swarm optimization algorithm to obtain the iteratively optimized prefabricated control scheme are as follows: Determine the relevant parameters of the particle swarm algorithm, such as the number of particles, the number of iterations, etc. Encode the parameters of the prefabricated control scheme into a particle position vector and initialize the particle position and speed. Substitute the scheme represented by the particle into the simulation imaging system, and calculate the image entropy of the simulated imaging image as the fitness value. During the iteration, the particle updates the speed and position according to its own historical optimality and the group's historical optimality. The speed update considers the current speed, inertia weight, cognition and social parts, and the position changes according to the updated speed. Repeat the operations of calculating fitness, updating extreme values, updating speed and position until the iteration stop condition is reached to obtain the iteratively optimized prefabricated control scheme;

[0078] The prefabricated control scheme after iterative optimization is introduced into the infrared night vision device simulation imaging system for simulation imaging analysis, and after each iterative optimization cycle is completed, a simulation imaging image is obtained, and the image entropy of the simulation imaging image is calculated; the image entropy of the simulation imaging image is compared with a preset threshold;

[0079] If the image entropy of the simulated imaging image is still not greater than the preset threshold, the next iterative optimization cycle is executed;

[0080] When the image entropy of the simulated imaging image of the iteratively optimized prefabricated control scheme is greater than a preset threshold, the iteration is stopped, and the prefabricated control scheme after the last iterative optimization is used as the standard control scheme of the infrared night vision device;

[0081] After the standard control scheme is obtained, the standard working parameters corresponding to each time node of each module in the infrared night vision device when the infrared night vision device is simulated controlled according to the standard control scheme are obtained in the infrared night vision device simulation imaging system, and the standard working parameter set of each module in the infrared night vision device when performing the preset working task is obtained;

[0082] The standard working parameter set includes the standard working parameters corresponding to each working time node when the corresponding module executes the preset working task.

[0083] It should be noted that when the image entropy of the simulated imaging image obtained based on the initial prefabricated control scheme is not greater than the preset threshold, it indicates that the imaging effect under the prefabricated control scheme does not meet the expected standard. At this time, the prefabricated control scheme needs to be optimized. The particle swarm optimization algorithm is used to optimize the prefabricated control scheme iteratively. The particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence. It simulates the behavior of bird flocks foraging. Multiple particles (representing possible solutions) continuously move and adjust their positions in the search space to find the optimal solution. Here, each particle represents a possible parameter combination of the prefabricated control scheme. In each iteration, the algorithm updates the position of the particle according to preset rules (such as the historical optimal position of the particle itself and the historical optimal position of the group, etc.), that is, updates the parameters of the prefabricated control scheme. After each iterative optimization cycle is completed, the iteratively optimized prefabricated control scheme is imported into the infrared night vision device simulation imaging system for simulation imaging analysis. Then the image entropy of the simulated imaging image is calculated and compared with the preset threshold. If the image entropy is still not greater than the preset threshold, it means that the current prefabricated control scheme has not achieved a satisfactory imaging effect, and it is necessary to continue to execute the next iterative optimization cycle. This process is repeated until the image entropy of the simulated imaging image of the iteratively optimized prefabricated control scheme is greater than the preset threshold. When the condition that the image entropy is greater than the preset threshold is met, the iteration is stopped, and the prefabricated control scheme after the last iterative optimization is determined as the standard control scheme for the infrared night vision device. Then, in the infrared night vision device simulation imaging system, the device is simulated and controlled according to this standard control scheme to obtain the standard working parameters corresponding to each module in the device at each time node when performing the preset work task. These parameters constitute the standard working parameter set, which contains the standard working parameters corresponding to each working time node when the corresponding module performs the preset work task. This means that for each module in the device, there are clear standard working parameters at different time stages of the entire preset work task, which provides a basis for the precise control of the device in actual operation.

[0084] When the image entropy of the simulated imaging image of the prefabricated control scheme does not meet the requirements, the particle swarm optimization algorithm is used for iterative optimization, and the imaging effect is evaluated in each iterative cycle. Finally, a standard control scheme that can make the image entropy of the simulated imaging image greater than the preset threshold is determined, and then the standard working parameter set corresponding to each module of the equipment at each time node when executing the preset work task is obtained. This helps to improve the imaging quality of infrared night vision equipment and ensure that each module of the equipment operates according to the optimal parameter combination in actual work, so as to adapt to complex working environments and task requirements.

[0085] Furthermore, the real-time image data of each optical axis is fused to generate a fused real-time imaging image, specifically:

[0086] Acquire real-time image data of each optical axis, and combine the real-time image data of each optical axis into a combination matrix of image data;

[0087] Centralizing the combined matrix, and calculating the covariance matrix of the combined matrix after the centralization, wherein the covariance matrix can reflect the correlation between different image data;

[0088] Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix. The eigenvalues ​​represent the variance of the data in the direction of the corresponding eigenvectors, and the eigenvectors determine the direction of the data in the new space.

[0089] Sort the eigenvectors according to the size of the eigenvalues, and select the eigenvectors of the first k eigenvalues ​​as the selected eigenvectors;

[0090] Construct a transformation matrix according to the selected eigenvectors, perform linear transformation on the combination matrix of the original image data through the transformation matrix to obtain a number of principal components; construct a principal component matrix according to the obtained principal components;

[0091] The principal component matrix is ​​inversely transformed to obtain the original spatial form of the image data, thereby obtaining a fused real-time imaging image.

[0092] It should be noted that first, the real-time image data of each optical axis is obtained and combined into a combination matrix of image data. This step is to integrate the image data from different optical axes into a data structure to provide a unified data source for subsequent processing. The combination matrix is ​​centralized to make the mean of the data 0, which can simplify subsequent calculations. The covariance matrix of the centralized combination matrix is ​​calculated, and the covariance matrix can reflect the correlation between different image data. For example, if the covariance between two image data is positive, it means that they are positively correlated to some extent; if it is negative, they are negatively correlated; if it is close to 0, the correlation is weak. Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix. The eigenvalue represents the variance of the data in the direction of the corresponding eigenvector. The larger the variance, the greater the amount of information contained in the data in this direction. The eigenvector determines the direction of the data in the new space. It can project the original data onto the new coordinate axis, thereby realizing the dimensionality reduction and feature extraction of the data. Sort the eigenvectors according to the size of the eigenvalue, and select the eigenvectors of the first k eigenvalues ​​as the selected eigenvectors. This is a feature selection method. Selecting eigenvectors with larger eigenvalues ​​means selecting directions with larger variance and more information in the data. Then, a transformation matrix is ​​constructed based on the selected eigenvectors. This transformation matrix will be used to perform linear transformation on the original image data. The combination matrix of the original image data is linearly transformed through the transformation matrix to obtain several principal components. The principal component is the representation of the original image data in the new feature space, which contains the main information of the original data. A principal component matrix is ​​constructed based on the several principal components obtained, and then the principal component matrix is ​​inversely transformed to convert the data from the principal component space back to the original space form, thereby obtaining a fused real-time imaging image. This inverse transformation can restore the original structure of the image data while integrating the main information of the image data of each optical axis.

[0093] Based on the correlation analysis of image data, feature selection is performed by calculating eigenvalues ​​and eigenvectors, extracting the main information and constructing the principal components, and finally inverse transforming to obtain the fused image. This process can reduce the redundant information in the image data, enhance the overall feature performance of the image, improve the quality and recognizability of the image, and thus provide clearer and more accurate real-time imaging images for infrared night vision equipment.

[0094] Furthermore, preset correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the prefabricated infrared night vision device are constructed according to the preset correction measures corresponding to the various imaging parameter offset values. Specifically, the B-tree-index map is:

[0095] Preset deviation correction measures corresponding to various imaging parameter deviation values ​​of the imaging module in the prefabricated infrared night vision device;

[0096] Construct a B-tree and initialize the root node and preset splitting rules of the B-tree; use the offset value of each imaging parameter as a data item; perform feature extraction processing on each data item to obtain a unique identification value of each data item;

[0097] Starting from the root node of the B-tree, compare the data item identification value with the key value of the current node; if the identification value is less than the minimum key value of the current node, search downward along the leftmost subtree; if it is greater than the maximum key value of the current node, search downward along the rightmost subtree; otherwise, find a suitable subtree within the key value range of the current node and search downward;

[0098] When a leaf node is reached, the preset position is located in the leaf node and the data item is inserted; if the leaf node is full, the leaf node is split into two nodes according to the splitting rule of the B-tree, and the key values ​​and pointers of the parent node and higher-level nodes are adjusted;

[0099] Perform split iteration according to the above steps to store each data item in order in each leaf node of the B-tree until there are no other data items in each leaf node, then stop split iteration;

[0100] The preset correction measures corresponding to various imaging parameter offset values ​​are respectively stored in the leaf nodes to which the corresponding data items belong, and a B-tree-index map is obtained.

[0101] It should be noted that, firstly, the preset correction measures corresponding to the offset values ​​of various imaging parameters of the imaging module in the infrared night vision device are prefabricated by relevant technicians. This means that corresponding solutions are formulated in advance for the offset of different imaging parameters that may occur in the imaging module. For example, when the focal length, exposure time, contrast and other parameters of the imaging module are offset, it is determined how to adjust these parameters to correct the offset.

[0102] Construct a B-tree and initialize its root node and preset splitting rules. B-tree is a self-balancing multi-way search tree that is efficient in storing and searching data. The root node is initialized as the starting point of the B-tree. The preset splitting rules are used to split the node when the amount of node data reaches a certain limit to maintain the balance of the B-tree. Each imaging parameter offset value is used as a data item, and feature extraction is performed on each data item to obtain a unique identification value for each data item. This identification value will be used to locate and compare data items in the B-tree to ensure that the data items are correctly stored and queried in the B-tree.

[0103] When inserting a data item into a B-tree, start from the root node and compare the data item identification value with the key value of the current node. If the identification value is less than the minimum key value of the current node, search down along the leftmost subtree; if it is greater than the maximum key value of the current node, search down along the rightmost subtree; otherwise, find a suitable subtree within the key value range of the current node and search down. When a leaf node is reached, locate the preset position in the leaf node and insert the data item. This process ensures that the data items are stored in a certain order in the B-tree.

[0104] If the leaf node is full, according to the splitting rules of the B-tree, the leaf node is split into two nodes, and the key values ​​and pointers of the parent node and higher-level nodes are adjusted. This split operation is to maintain the balance of the B-tree and prevent the tree from being too high, thereby ensuring search efficiency.

[0105] Through the above split iteration, each data item is stored in order in each leaf node of the B-tree until there are no other data items in each leaf node. Finally, the preset correction measures corresponding to various imaging parameter offset values ​​are stored in the leaf nodes to which the corresponding data items belong, and a B-tree-index map is obtained. This map establishes an efficient index relationship between the imaging parameter offset value and the corresponding preset correction measures, which is convenient for subsequent query and use.

[0106] In summary, through the above process of constructing the B-tree-index map, an efficient index relationship is established between the imaging parameter offset value of the imaging module of the infrared night vision device and the corresponding preset correction measures. When it is found that the imaging module has an imaging parameter offset during the operation of the device, the corresponding preset correction measures can be quickly located in the B-tree-index map according to the offset value, so as to adjust the imaging parameters in a timely and accurate manner, thereby improving the stability and accuracy of the device imaging and reducing problems such as image quality degradation caused by imaging parameter offset.

[0107] Furthermore, if the image entropy of the real-time imaging image is not greater than a preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure, specifically:

[0108] If the image entropy of the real-time imaging image is not greater than a preset threshold, obtaining the real-time imaging parameters of the imaging module in the infrared night vision device at the current working time node;

[0109] and obtaining a standard working parameter set of an imaging module in an infrared night vision device when performing a preset working task, and obtaining standard working parameters of the imaging module at a current working time node in the standard working parameter set; wherein the standard working parameters are standard imaging parameters;

[0110] Calculating the difference between the standard working parameter of the imaging module at the current working time node and the real-time imaging parameter to obtain an imaging parameter deviation value, and acquiring an identification value of the imaging parameter deviation value;

[0111] Importing the identification value of the imaging parameter deviation value into the B-tree-index map for matching, matching to obtain leaf nodes with the same identification items, and extracting preset deviation correction measures from the matched leaf nodes;

[0112] The real-time imaging parameters of the infrared night vision device are adjusted and processed according to the extracted preset correction measures.

[0113] It should be noted that when the image entropy of the real-time imaging image is not greater than the preset threshold, this indicates that the current imaging quality does not meet the expected standard and the imaging parameters need to be adjusted. At this time, the parameters related to the imaging module are obtained. First, the real-time imaging parameters of the imaging module in the infrared night vision device at the current working time node are obtained, which reflects the current actual working state of the imaging module. At the same time, the standard working parameters (i.e., standard imaging parameters) of the imaging module at the current working time node are obtained from the previously obtained standard working parameter set, and the standard working parameter set is a parameter set under the ideal state obtained through evaluation and optimization. The imaging parameter deviation value is obtained by calculating the difference between the standard working parameters of the imaging module at the current working time node and the real-time imaging parameters. This deviation value can quantify the degree of deviation between the current imaging parameters and the ideal parameters. Then, the identification value of the imaging parameter deviation value is obtained, which will be used for matching in the B-tree-index map. The identification value of the imaging parameter deviation value is imported into the B-tree-index map for matching. Since the B-tree-index map stores the corresponding relationship between the imaging parameter offset value and the preset correction measure, the leaf node with the same identification item can be matched, and then the preset correction measure is extracted from the matched leaf node. The real-time imaging parameters of the infrared night vision device are regulated and processed according to the extracted preset correction measures. This step can adjust the parameters of the imaging module in the direction close to the standard imaging parameters, which is expected to improve the imaging quality and make the image entropy of the real-time imaging image reach or exceed the preset threshold. When the real-time imaging image quality is poor (the image entropy is not greater than the preset threshold), the deviation value is obtained by comparing the real-time imaging parameters of the imaging module with the standard imaging parameters, and then the corresponding preset correction measures are quickly matched using the B-tree-index map, and then the real-time imaging parameters are regulated. This process realizes the rapid and precise regulation of the imaging parameters of the infrared night vision device, which helps to improve the imaging quality, so that the device can be adjusted in time when the imaging parameters deviate from the standard, and ensure that the device stably and efficiently outputs high-quality imaging results.

[0114] In addition, the method may further comprise the following steps:

[0115] Collect image datasets for training. The datasets should contain the target objects to be identified and be labeled with the category and location information of the targets. Preprocess the datasets, including normalizing and resizing the images to meet the model input requirements.

[0116] Build the network structure of the YOLO model, determine the number of network layers and convolution kernel size parameters; define the loss function, which is used to measure the difference between the model prediction results and the actual annotations, such as target positioning loss, target confidence loss, and category prediction loss;

[0117] The labeled image data is input into the model for forward propagation to obtain the prediction result. The gradient of the loss function to the model parameters is calculated through the back-propagation algorithm. The model parameters are updated according to the gradient using the optimization algorithm (such as stochastic gradient descent). The process of forward propagation, back-propagation and parameter update is repeated continuously. After multiple rounds of iterative training, the performance indicators of the model on the validation set (such as accuracy, recall rate, etc.) reach a satisfactory level. At this time, the model training is completed and the pre-trained YOLO model is obtained.

[0118] Adjust the fused real-time imaging image to a size suitable for the YOLO algorithm input, and then input the image into the pre-trained YOLO model. The network structure of the YOLO model will quickly process the image.

[0119] In the convolution layer of the model, various features of the image are automatically extracted through the convolution kernel. These features include the shape and texture information of the target. Then, in the detection layer, the extracted features are matched according to the pre-set anchor boxes. Each anchor box predicts the bounding box coordinates, confidence level, and category probability of the target.

[0120] The confidence level is used to filter out areas where targets may exist. For prediction results with confidence levels higher than a set threshold, the corresponding target type is determined, and the position of the target in the image is obtained based on the predicted bounding box coordinates.

[0121] Calculate the deviation between the optical axis and the target based on the position of the target. If the deviation exceeds the allowable range, adjust the direction of the optical axis, focal length and other parameters according to the deviation to make the target located at the center of the imaging screen or a specific tracking area.

[0122] As the target moves, new fused real-time imaging images are continuously input into the YOLO model and the above steps are repeated to achieve accurate tracking of the target.

[0123] In addition, the method may further comprise the following steps:

[0124] Obtaining operation log information of the infrared night vision device, and obtaining characteristic working parameters corresponding to the infrared night vision device in various operation states according to the operation log information; wherein the operation state includes a normal state, a parameter offset state, and a fault state;

[0125] A fuzzy evaluation algorithm is introduced, and fuzzy evaluation is performed on characteristic working parameters corresponding to various operating states according to the fuzzy evaluation algorithm to obtain a fuzzy matrix between various operating states;

[0126] The Monte Carlo analysis algorithm is introduced to analyze the fuzzy matrix between various operating states based on the Monte Carlo analysis algorithm, and the conditional transition probability of the fuzzy matrix of each operating state transferring to the fuzzy matrix of another operating state is obtained;

[0127] Constructing a conditional transfer probability matrix according to the conditional transfer probability of the fuzzy matrix of each operating state transferring to the fuzzy matrix of another operating state;

[0128] Building a prediction model based on a deep learning network, importing the conditional transfer probability matrix into the prediction model for coding learning, and outputting the conditional transfer probability prediction model after the model parameters meet the requirements;

[0129] Acquire the real-time working parameters of the infrared night vision device at the preset time node, import the real-time working parameters of the infrared night vision device at the preset time node into the conditional transfer probability prediction model for prediction, and acquire the fuzzy matrix of the infrared night vision device at the current preset time node and the conditional transfer probability of the fuzzy matrix at the current preset time node being transferred to other fuzzy matrices;

[0130] When the conditional transfer probability of the fuzzy matrix of the current preset time node being transferred to other fuzzy matrices is greater than the preset probability threshold, the fuzzy matrix of the current preset time node is output and transferred to other fuzzy matrices as the fuzzy matrix of the current preset time node;

[0131] When the conditional transfer probability of the fuzzy matrix of the current preset time node being transferred to other fuzzy matrices is not greater than the preset probability threshold, the fuzzy matrix of the current preset time node is directly output;

[0132] It is determined whether the fuzzy matrix of the current preset time node is the fuzzy matrix corresponding to the fault state. If so, fault warning information is generated.

[0133] It should be noted that the operation log information of the infrared night vision device is first obtained, which is a record of various situations during the operation of the device. The characteristic working parameters corresponding to the device in various operating states such as normal state, parameter offset state and fault state are obtained from the operation log. These characteristic working parameters can reflect the working characteristics of the device in different operating states. For example, in the normal state, the working parameters of each module are in the normal range, while in the fault state, some parameters may exceed the normal range or fluctuate abnormally. The fuzzy evaluation algorithm is introduced to perform fuzzy evaluation on the characteristic working parameters corresponding to different operating states. Since the relationship between the operating state and the working parameters of the device may not be absolutely clear, there is a certain fuzziness, and the fuzzy evaluation algorithm can handle this uncertainty. The fuzzy matrix between various operating states is obtained through the algorithm. This fuzzy matrix describes the fuzzy relationship between different operating states, such as the degree of fuzzy relationship from the normal state to the parameter offset state. Then the fuzzy matrix is ​​analyzed using the Monte Carlo analysis algorithm. The Monte Carlo algorithm handles complex probability problems through random simulation and statistical analysis. Here, the conditional transition probability of the fuzzy matrix of each operating state to the fuzzy matrix of another operating state is obtained. This conditional transition probability reflects the possibility of the device switching from one operating state to another operating state. According to the obtained conditional transition probability, a conditional transition probability matrix is ​​constructed. This matrix comprehensively describes the probability relationship of the device switching between different operating states. Then, a prediction model is constructed based on a deep learning network, and the conditional transition probability matrix is ​​imported into the prediction model for coding learning. The deep learning network has powerful learning and fitting capabilities and can learn the complex relationships in the conditional transition probability matrix. When the model parameters meet the requirements, the conditional transition probability prediction model is output. The real-time working parameters of the device at the preset time node are obtained and imported into the conditional transition probability prediction model for prediction. The fuzzy matrix of the device at the current preset time node and the conditional transition probability of the fuzzy matrix at the current preset time node being transferred to other fuzzy matrices are obtained. If this conditional transition probability is greater than the preset probability threshold, the fuzzy matrix at the current preset time node is output and transferred to other fuzzy matrices as the fuzzy matrix at the current preset time node; if it is not greater than the preset probability threshold, the fuzzy matrix at the current preset time node is directly output. Finally, it is determined whether the fuzzy matrix at the current preset time node is the fuzzy matrix corresponding to the fault state. If so, fault warning information is generated, thereby realizing early warning of equipment failure. This method can effectively deal with the ambiguity of the relationship between the equipment operating status and working parameters, accurately predict the conversion probability of the equipment operating status, and ultimately achieve early warning of equipment failures, thereby improving the reliability and safety of the operation of infrared night vision equipment.

[0134] In addition, the method may further comprise the following steps:

[0135] Acquire a fused real-time imaging image, and calculate a structural similarity index between the real-time imaging image and a standard imaging image;

[0136] If the structural similarity index is not greater than a preset index threshold, matrix conversion processing is performed on the real-time imaging image and the standard imaging image to obtain a gray level co-occurrence matrix corresponding to the real-time imaging image and a gray level co-occurrence matrix corresponding to the standard imaging image;

[0137] Compare and analyze the elements of the same matrix position in the gray level co-occurrence matrix corresponding to the real-time imaging image and the gray level co-occurrence matrix corresponding to the standard imaging image, obtain the matrix position corresponding to the unequal elements in the two matrices, and obtain the image deviation area between the real-time imaging image and the standard imaging image;

[0138] Marking the imaging module responsible for the image deviation area as a suspicious imaging module, obtaining the real-time working parameters of the suspicious imaging module, and calculating the difference between the real-time imaging parameters and the standard imaging parameters;

[0139] If the difference between the real-time imaging parameter and the standard imaging parameter is greater than a preset difference threshold, the real-time imaging parameter of the suspicious imaging module is corrected based on the difference.

[0140] It should be noted that by calculating the structural similarity index between the real-time imaging image and the standard imaging image, the grayscale co-occurrence matrix is ​​further analyzed to determine the image deviation area when the index does not meet the standard, thereby locating the suspicious imaging module, and then comparing the difference between its working parameters and the standard parameters, and making corrections when the difference is too large. This method can effectively locate the cause of poor imaging quality, accurately find imaging modules that may have problems, and improve imaging quality by correcting their working parameters, thereby enhancing the stability and accuracy of imaging of multi-axis optical imaging infrared night vision equipment.

[0141] The present invention also discloses a multi-axis optical imaging infrared night vision device, which is applied to a control method of a multi-axis optical imaging infrared night vision device as described in any one of the above. Figure 3 As shown, the infrared night vision device comprises:

[0142] A support frame 1011 and a night vision device body 1012, wherein the night vision device body and the support frame are connected via an angle adjuster 1013;

[0143] An optical module is arranged in the main body of the night vision device, and the optical module includes a plurality of optical axes, a lens group and a reflector, and is used to collect and focus infrared light;

[0144] The main body of the night vision device is provided with a detection module for converting infrared light into electrical signals;

[0145] An imaging module is provided in the main body of the night vision device to collect and analyze image signals to generate images;

[0146] A display module is provided in the main body of the night vision device, for displaying the generated image;

[0147] A control module is provided in the main body of the night vision device, which is used to control the working parameters of each module;

[0148] A signal transmission module is arranged in the main body of the night vision device, which is used to send the collected image signal to a preset terminal.

[0149] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A control method for a multi-axis optical imaging infrared night vision device, characterized in that: The following steps are involved: Obtaining a prefabricated control scheme for an infrared night vision device, and evaluating and optimizing the prefabricated control scheme to obtain a standard control scheme for the infrared night vision device and a standard working parameter set for each module when performing a preset working task; Controlling the infrared night vision device to perform monitoring work according to the standard control scheme, and acquiring the real-time image data of each optical axis through the image sensor at a preset time point, fusing the real-time image data of each optical axis, and generating a fused real-time imaging image; Calculating the image entropy of the real-time imaging image; If the image entropy of the real-time imaging image is greater than a preset threshold, no adjustment or processing is performed on the real-time imaging parameters of the infrared night vision device; Preset deviation correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the prefabricated infrared night vision device, and constructing a B-tree-index map according to the preset deviation correction measures corresponding to various imaging parameter offset values; If the image entropy of the real-time imaging image is not greater than a preset threshold, an imaging parameter deviation value of the imaging module is obtained, a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure; Among them, the preset correction measures corresponding to various imaging parameter offset values ​​of the imaging module in the prefabricated infrared night vision device are used to construct a B-tree-index map according to the preset correction measures corresponding to various imaging parameter offset values, specifically: Preset deviation correction measures corresponding to various imaging parameter deviation values ​​of the imaging module in the prefabricated infrared night vision device; Construct a B-tree and initialize the root node and preset splitting rules of the B-tree; use the offset value of each imaging parameter as a data item; perform feature extraction processing on each data item to obtain a unique identification value of each data item; Starting from the root node of the B-tree, compare the data item identification value with the key value of the current node; if the identification value is less than the minimum key value of the current node, search downward along the leftmost subtree; if it is greater than the maximum key value of the current node, search downward along the rightmost subtree; otherwise, find a suitable subtree within the key value range of the current node and search downward; When a leaf node is reached, the preset position is located in the leaf node and the data item is inserted; if the leaf node is full, the leaf node is split into two nodes according to the splitting rule of the B-tree, and the key values ​​and pointers of the parent node and higher-level nodes are adjusted; Perform split iteration according to the above steps to store each data item in order in each leaf node of the B-tree until there are no other data items in each leaf node, then stop split iteration; The preset correction measures corresponding to various imaging parameter offset values ​​are respectively stored in the leaf nodes to which the corresponding data items belong, and a B-tree-index map is obtained.

2. The control method of a multi-axis optical imaging infrared night vision device according to claim 1, characterized in that: Obtain a prefabricated control scheme for infrared night vision equipment, evaluate and optimize the prefabricated control scheme, and obtain a standard control scheme for infrared night vision equipment and a standard working parameter set for each module when performing a preset working task, specifically: Obtaining a structural model diagram of the infrared night vision device based on the structural parameter specification of the infrared night vision device, obtaining a scene feature model diagram within the working area of ​​the infrared night vision device based on remote sensing technology, and obtaining environmental feature parameters of the infrared night vision device within a preset working time period based on meteorological software; According to the structural model diagram, the scene feature model diagram and the environmental feature parameters, a simulation imaging system of an infrared night vision device is constructed using simulation software; Obtaining a prefabricated control scheme for an infrared night vision device, importing the prefabricated control scheme into the infrared night vision device simulation imaging system for simulation imaging analysis, and obtaining a simulated imaging image obtained after the infrared night vision device is simulated controlled based on the prefabricated control scheme; Calculating the image entropy of the simulated imaging image, and comparing the image entropy of the simulated imaging image with a preset threshold; If the image entropy of the simulated imaging image is greater than a preset threshold, the prefabricated control scheme is used as a standard control scheme for the infrared night vision device.

3. The control method of a multi-axis optical imaging infrared night vision device according to claim 2, characterized in that: Obtaining a prefabricated control scheme of an infrared night vision device, and evaluating and optimizing the prefabricated control scheme to obtain a standard control scheme of the infrared night vision device and a standard working parameter set of each module when performing a preset working task, further comprising the following steps: If the image entropy of the simulated imaging image is not greater than a preset threshold, the prefabricated control scheme is optimized and iteratively processed using a particle swarm optimization algorithm to obtain an iteratively optimized prefabricated control scheme; The prefabricated control scheme after iterative optimization is introduced into the infrared night vision device simulation imaging system for simulation imaging analysis, and after each iterative optimization cycle is completed, a simulation imaging image is obtained, and the image entropy of the simulation imaging image is calculated; the image entropy of the simulation imaging image is compared with a preset threshold; If the image entropy of the simulated imaging image is still not greater than the preset threshold, the next iterative optimization cycle is executed; When the image entropy of the simulated imaging image of the iteratively optimized prefabricated control scheme is greater than a preset threshold, the iteration is stopped, and the prefabricated control scheme after the last iterative optimization is used as the standard control scheme of the infrared night vision device; After the standard control scheme is obtained, the standard working parameters corresponding to each time node of each module in the infrared night vision device when the infrared night vision device is simulated controlled according to the standard control scheme are obtained in the infrared night vision device simulation imaging system, and the standard working parameter set of each module in the infrared night vision device when performing the preset working task is obtained; The standard working parameter set includes the standard working parameters corresponding to each working time node when the corresponding module executes the preset working task.

4. The control method of a multi-axis optical imaging infrared night vision device according to claim 1, characterized in that: The real-time image data of each optical axis are fused to generate a fused real-time imaging image, specifically: Acquire real-time image data of each optical axis, and combine the real-time image data of each optical axis into a combination matrix of image data; Performing a centering process on the combined matrix, and calculating a covariance matrix of the combined matrix after the centering process, wherein the covariance matrix can reflect the correlation between different image data; Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix. The eigenvalues ​​represent the variance of the data in the direction of the corresponding eigenvectors, and the eigenvectors determine the direction of the data in the new space. Sort the eigenvectors according to the size of the eigenvalues, and select the eigenvectors of the first k eigenvalues ​​as the selected eigenvectors; Construct a transformation matrix according to the selected eigenvectors, perform linear transformation on the combination matrix of the original image data through the transformation matrix to obtain a number of principal components; construct a principal component matrix according to the obtained principal components; The principal component matrix is ​​inversely transformed to obtain the original spatial form of the image data, thereby obtaining a fused real-time imaging image.

5. The control method of a multi-optical axis optical imaging infrared night vision device according to claim 1, characterized in that: If the image entropy of the real-time imaging image is not greater than a preset threshold, the imaging parameter deviation value of the imaging module is obtained, and a preset correction measure is obtained according to the imaging parameter deviation value and combined with B-tree-index map matching, and the real-time imaging parameters of the infrared night vision device are regulated and processed according to the preset correction measure, specifically: If the image entropy of the real-time imaging image is not greater than a preset threshold, obtaining the real-time imaging parameters of the imaging module in the infrared night vision device at the current working time node; and obtaining a standard working parameter set of an imaging module in an infrared night vision device when performing a preset working task, and obtaining standard working parameters of the imaging module at a current working time node in the standard working parameter set; wherein the standard working parameters are standard imaging parameters; Calculating the difference between the standard working parameter of the imaging module at the current working time node and the real-time imaging parameter to obtain an imaging parameter deviation value, and acquiring an identification value of the imaging parameter deviation value; Importing the identification value of the imaging parameter deviation value into the B-tree-index map for matching, matching to obtain leaf nodes with the same identification items, and extracting preset deviation correction measures from the matched leaf nodes; The real-time imaging parameters of the infrared night vision equipment are regulated and processed according to the extracted preset correction measures.

6. A multi-axis optical imaging infrared night vision device, applied to a control method of a multi-axis optical imaging infrared night vision device according to any one of claims 1 to 5, characterized in that: The infrared night vision device comprises: A support frame and a night vision device body, wherein the night vision device body and the support frame are connected via an angle adjuster; An optical module is arranged in the main body of the night vision device, and the optical module includes a plurality of optical axes, a lens group and a reflector, and is used to collect and focus infrared light; The main body of the night vision device is provided with a detection module for converting infrared light into electrical signals; An imaging module is provided in the main body of the night vision device to collect and analyze image signals to generate images; A display module is provided in the main body of the night vision device, for displaying the generated image; A control module is provided in the main body of the night vision device, which is used to control the working parameters of each module; A signal transmission module is arranged in the main body of the night vision device, which is used to send the collected image signal to a preset terminal.

Images