Hunting camera imaging quality optimization method and system based on multi-mode data fusion

Abstract

The present invention discloses a hunting camera imaging quality optimization method and system based on multi-mode data fusion, which relates to the field of image processing. The method includes: obtaining data source features of multi-mode data, establishing a multi-mode data feature description library; analyzing imaging component factors, performing hierarchical decomposition on the imaging component factors, and obtaining imaging levels; performing imaging quality dimension correlation analysis on the imaging levels, and obtaining positive quality factors and negative quality factors of the imaging levels; establishing a compensation effect relationship between each data feature in the multi-mode data feature description library and the positive quality factor and the negative quality factor; and performing imaging compensation optimization on each data feature in the multi-mode data feature description library based on the compensation effect relationship. The method solves the technical problems of limited optimization effect, insufficient robustness and accuracy of imaging quality in existing hunting camera imaging quality, and achieves the technical effects of improving the imaging quality optimization effect, as well as improving the robustness and accuracy of imaging quality.

Description

Technical Field

[0001] The present application relates to the field of image processing, and in particular to a method and system for optimizing the imaging quality of a hunting camera based on multi-mode data fusion. Background Art

[0002] In wildlife monitoring and hunting activities, the imaging quality of hunting cameras is crucial. High-quality imaging can clearly capture the moments of animal activity and provide valuable data for scientific research and entertainment. However, due to the complexity and variability of the wild environment, the imaging quality of hunting cameras is often affected by many factors, such as lighting conditions, weather conditions, animal behavior, etc. Among the existing methods for optimizing the imaging quality of hunting cameras, they mainly rely on processing and analyzing data from a single imaging mode. These methods usually include steps such as image enhancement, denoising, and color correction, aiming to improve the clarity and color reproduction of the image. However, these methods ignore the information provided by other imaging modes, resulting in a significant decrease in imaging quality at night or in severe weather conditions.

[0003] In the current related technologies, there are technical problems in the imaging quality of hunting cameras, such as limited optimization effect, insufficient robustness and accuracy of imaging quality. Summary of the invention

[0004] The present application provides a hunting camera imaging quality optimization method and system based on multi-mode data fusion, collects data from different imaging modes, performs imaging feature analysis on the data of each imaging mode, establishes a mapping relationship between imaging features and imaging modes, constructs a multi-mode data feature description library containing imaging feature identifiers, analyzes the imaging quality of the hunting camera, identifies the component factors that affect the imaging quality, obtains different levels of the imaging system by hierarchical decomposition of these factors, identifies the key dimensions that affect the imaging quality for each imaging level, fits the imaging influence relationship between the key dimensions and the imaging level, identifies the positive quality factor and the reverse quality factor, matches the features in the multi-mode data feature description library with the quality factors in the imaging level, analyzes how each feature affects the positive quality factor and the reverse quality factor, establishes a compensation effect relationship between them, determines compensation strategies for different imaging levels and imaging features based on the compensation effect relationship, and quantitatively adjusts each data feature in the multi-mode data feature description library by applying the compensation strategy, thereby achieving the technical effect of improving the imaging quality optimization effect and improving the robustness and accuracy of the imaging quality.

[0005] This application provides a hunting camera imaging quality optimization method based on multi-mode data fusion, including:

[0006] Acquire data source features of multimodal data and establish a multimodal data feature description library; parse imaging component factors, perform hierarchical decomposition on the imaging component factors, and obtain imaging levels; perform imaging quality dimension correlation analysis on the imaging levels, and obtain positive quality factors and negative quality factors of the imaging levels; establish a compensation effect relationship between each data feature in the multimodal data feature description library and the positive quality factor and negative quality factor; and perform imaging compensation optimization on each data feature in the multimodal data feature description library based on the compensation effect relationship.

[0007] In a possible implementation, the multi-modal data feature description library is established, and the following processing is performed:

[0008] The multimodal data at least includes visible light imaging data, infrared imaging data, and thermal imaging data, and imaging advantage features and imaging disadvantage features of the multimodal data are analyzed respectively to obtain imaging features of each multimodal data; a mapping relationship between the imaging features and imaging levels, and a mapping relationship between the imaging features and multimodal data are established; based on the mapping relationships, imaging features, imaging levels, and multimodal data are mapped and integrated to construct the multimodal data feature description library, wherein the multimodal data feature description library has feature identifiers of imaging advantage features and imaging disadvantage features.

[0009] In a possible implementation, the imaging component factors are analyzed, the imaging component factors are hierarchically decomposed to obtain imaging levels, and the following processing is performed:

[0010] The imaging quality factors of the hunting camera are analyzed according to the sensor performance, optical system characteristics and environmental conditions to obtain the imaging component factors; the imaging component factors are decomposed according to the multiple levels of physics, signal processing, space, time, spectrum and perception to establish the correlation between the imaging component factors and the multiple levels; based on the correlation, the imaging component factors are projected to the corresponding multiple levels to obtain the imaging level.

[0011] In a possible implementation, the imaging level is subjected to an imaging quality dimension correlation analysis to obtain a positive quality factor and a negative quality factor of the imaging level, and the following processing is performed:

[0012] For each imaging level, the key dimensions that affect the imaging quality are identified; the influence relationship between the key dimensions and the single-layer imaging of the level is fitted, and the single-layer positive quality factor and the single-layer negative quality factor are identified; the inter-layer imaging influence relationship between the imaging levels is fitted, and the inter-layer positive quality factor and the inter-layer negative quality factor are identified; based on the imaging level, the quality influence relationship of the single-layer positive quality factor, the single-layer negative quality factor, the inter-layer positive quality factor, and the inter-layer negative quality factor are fused to obtain the positive quality factor and the negative quality factor of each imaging level.

[0013] In a possible implementation, a compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor is established, and the following processing is performed:

[0014] Obtain the quality influence relationship between the positive quality factor and the reverse quality factor on the imaging level, perform basic gradient conversion on the quality influence relationship, and obtain the imaging influence gradient of each quality factor; extract the imaging features of the multimodal data corresponding to the imaging level according to the multimodal data feature description library; based on the corresponding relationship between the imaging features and the quality factors, perform imaging influence gradient projection on the imaging features of the multimodal data of each imaging level to obtain the imaging influence gradient of the multimodal data; perform imaging quality compensation analysis based on the imaging influence gradient of the multimodal data to obtain the compensation effect relationship, and the compensation effect relationship has a gradient identifier.

[0015] In a possible implementation, based on the compensation effect relationship, imaging compensation optimization is performed on each data feature in the multi-mode data feature description library, and the following processing is performed:

[0016] Based on the gradient identifier, feature gradient segmentation is performed on each data feature in the multi-modal data feature description library to construct multi-scale features; according to the compensation effect relationship corresponding to the gradient identifier, scale feature mapping interaction is performed on the multi-scale features, and imaging compensation optimization is performed according to the multi-scale interaction relationship to obtain hierarchical imaging parameters.

[0017] In a possible implementation, scale feature mapping interaction is performed on the multi-scale features, imaging compensation optimization is performed according to the multi-scale interaction relationship, hierarchical imaging parameters are obtained, and the following processing is performed:

[0018] The multi-scale features are sorted according to the size relationship of the parameters to determine the scale adjacency relationship; according to the scale adjacency relationship, the adjacent scale features are interacted in sequence, wherein the adjacent scale feature interaction includes: inputting the features of two adjacent scales into the feature enhancement layer for feature enhancement through the attention mechanism to obtain a first enhanced feature and a second enhanced feature; performing an inner product calculation and a softmax calculation on the first enhanced feature and the second enhanced feature to obtain a corresponding relationship between the two adjacent scale features, and fusing the two scale features based on the corresponding relationship to obtain a fused feature; completing the feature interaction of all adjacent scales according to the scale adjacency relationship, performing imaging compensation optimization evaluation on the fused feature, and obtaining the hierarchical imaging parameters, wherein the hierarchical imaging parameters are the hierarchical features with the best imaging compensation optimization evaluation results.

[0019] The present application also provides a hunting camera imaging quality optimization system based on multi-mode data fusion, including:

[0020] A multimodal data feature description library establishment module, the multimodal data feature description library establishment module is used to obtain the data source characteristics of the multimodal data and establish the multimodal data feature description library; an imaging component factor hierarchical decomposition module, the imaging component factor hierarchical decomposition module is used to parse the imaging component factors, perform hierarchical decomposition on the imaging component factors, and obtain the imaging level; an imaging quality dimension correlation analysis module, the imaging quality dimension correlation analysis module is used to perform imaging quality dimension correlation analysis on the imaging level, and obtain the positive quality factor and reverse quality factor of the imaging level; a compensation effect relationship establishment module, the compensation effect relationship establishment module is used to establish the compensation effect relationship between each data feature in the multimodal data feature description library and the positive quality factor and reverse quality factor; an imaging compensation optimization module, the imaging compensation optimization module is used to perform imaging compensation optimization on each data feature in the multimodal data feature description library based on the compensation effect relationship.

[0021] The hunting camera imaging quality optimization method and system based on multi-mode data fusion proposed in this application first obtains the data source characteristics of the multi-mode data, establishes a multi-mode data feature description library, then analyzes the imaging component factors, and performs hierarchical decomposition on the imaging component factors to obtain the imaging level. Next, the imaging level is subjected to an imaging quality dimension correlation analysis to obtain the positive quality factor and the negative quality factor of the imaging level, and then a compensation effect relationship between each data feature in the multi-mode data feature description library and the positive quality factor and the negative quality factor is established. Finally, based on the compensation effect relationship, imaging compensation optimization is performed on each data feature in the multi-mode data feature description library, thereby achieving the technical effect of improving the imaging quality optimization effect, as well as improving the robustness and accuracy of the imaging quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0023] Figure 1 A flowchart of a method for optimizing the imaging quality of a hunting camera based on multi-mode data fusion provided in an embodiment of the present application.

[0024] Figure 2 A schematic diagram of the structure of a hunting camera imaging quality optimization system based on multi-mode data fusion provided in an embodiment of the present application.

[0025] Explanation of the accompanying drawings: multi-modal data feature description library establishment module 10, imaging component factor hierarchical decomposition module 20, imaging quality dimension correlation analysis module 30, compensation effect relationship establishment module 40, imaging compensation optimization module 50. DETAILED DESCRIPTION

[0026] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0027] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0028] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0029] The embodiment of the present application provides a hunting camera imaging quality optimization method based on multi-mode data fusion, such as Figure 1 As shown, the method includes:

[0030] Step S100, obtain the data source characteristics of the multi-mode data, and establish a multi-mode data feature description library. Specifically, the multi-mode data refers to data from different imaging modes, such as visible light imaging data, infrared imaging data, thermal imaging data, etc. Collect data from multiple imaging modes of hunting cameras (such as visible light, infrared, thermal imaging, etc.). Perform imaging feature analysis on the data of each imaging mode, that is, the unique attributes or characteristics of each imaging mode data, such as clarity, color reproduction, noise, lack of contrast, etc. Establish a mapping relationship between imaging features and specific imaging modes (such as visible light imaging, infrared imaging, etc.), and based on the mapping relationship, map and integrate the imaging features and imaging modes to construct a multi-mode data feature description library containing imaging feature identifiers.

[0031] In a possible implementation, the step S100 of establishing a multi-mode data feature description library further includes step S110, wherein the multi-mode data includes at least visible light imaging data, infrared imaging data, and thermal imaging data, and imaging advantage features and imaging disadvantage features of the multi-mode data are analyzed respectively to obtain imaging features of each multi-mode data. Specifically, data is collected from multiple imaging modes (such as visible light, infrared, thermal imaging, 3D imaging, multi-spectral / hyperspectral imaging, etc.) of a hunting camera, and the data includes static images, video sequences, or continuous frames, etc.

[0032] For each imaging mode, analyze its imaging advantages in specific scenes or conditions. For example, visible light imaging can provide clear colors and details when there is sufficient light during the day; infrared imaging can detect the heat distribution of objects at night or in low light conditions; thermal imaging can show the temperature difference of objects; 3D imaging can provide three-dimensional structural information of objects; multispectral / hyperspectral imaging can capture more detailed spectral information for material identification, etc. At the same time, analyze the imaging disadvantages of each imaging mode. For example, visible light imaging is not effective at night or in bad weather; infrared imaging is disturbed by ambient heat; thermal imaging has a low resolution; 3D imaging is affected by lighting conditions and surface characteristics of objects; multispectral / hyperspectral imaging data processing is more complicated, etc.

[0033] Based on the above analysis, the imaging features of each imaging mode are extracted. These features can be quantitative (such as contrast, brightness, noise level, etc.) or qualitative (such as color reproduction, clarity, etc.).

[0034] Step S120, establish a mapping relationship between the imaging features and the imaging levels, and a mapping relationship between the imaging features and the multi-mode data. Specifically, a mapping relationship is established between the extracted imaging features and the different levels of the imaging system to understand how the imaging features are generated in the different levels of the imaging system. At the same time, a mapping relationship is established between the imaging features and specific imaging modes (such as visible light imaging, infrared imaging, etc.) to distinguish the data features of different imaging modes in subsequent analysis and optimization. Step S130, based on the mapping relationship, the imaging features, imaging levels and multi-mode data are mapped and integrated to construct the multi-mode data feature description library, wherein the multi-mode data feature description library has feature identifiers of imaging advantage features and imaging disadvantage features. Specifically, based on the above mapping relationship, the imaging features, imaging levels and multi-mode data are integrated to form a database containing data features of multiple imaging modes. Add identifiers to each imaging feature in the database, including imaging advantage features and imaging disadvantage features, for rapid identification and utilization of these features in subsequent analysis and optimization. The multi-mode data feature description library is optimized, such as data deduplication, data compression, index establishment, etc., to improve the speed and efficiency of data query and processing. This implementation method establishes a comprehensive and detailed multi-mode data feature description library. This database not only contains the data features of multiple imaging modes, but also establishes the mapping relationship between these features and the imaging level and multi-mode data. This enables these features to be more accurately identified and utilized in subsequent analysis and optimization, thereby improving the imaging quality of hunting cameras.

[0035] Step S200, analyzing imaging component factors, performing hierarchical decomposition on the imaging component factors, and obtaining imaging levels. Specifically, the imaging quality of the hunting camera is analyzed to identify component factors that affect the imaging quality (various factors that affect the imaging quality of the hunting camera), such as sensor performance, optical system characteristics, signal processing algorithms, etc. By performing hierarchical decomposition on these factors, different levels of the imaging system are obtained.

[0036] In a possible implementation, the imaging component factors are analyzed, and the imaging component factors are hierarchically decomposed to obtain imaging levels. Step S200 further includes step S210, in which the imaging quality factors of the hunting camera are analyzed according to the sensor performance, optical system characteristics, and environmental conditions to obtain imaging component factors. Specifically, the imaging quality of the hunting camera is analyzed to identify the main factors affecting the imaging quality, including: sensor performance, optical system characteristics, and environmental conditions. Among them, the characteristics of the sensor such as sensitivity, resolution, and noise level directly affect the clarity and detail of the image; the characteristics of the lens such as focal length, aperture, and distortion correction determine the optical quality of the image; and external environmental factors such as lighting conditions, temperature, and humidity also affect the imaging quality. By analyzing these factors, a series of imaging component factors are obtained, which are the basis for subsequent hierarchical decomposition. Step S220, the imaging component factors are decomposed according to multiple levels of physics, signal processing, space, time, spectrum, and perception, and the correlation between the imaging component factors and the multiple levels is established. Specifically, the imaging component factors are further decomposed into sub-factors of different levels to more deeply understand their impact on imaging quality. These levels include: physical level (involving physical phenomena such as light propagation, diffraction, interference, and physical properties of sensors and lenses), signal processing level (involving signal processing processes such as image acquisition, conversion, enhancement, and filtering), spatial level (involving spatial characteristics such as image resolution, contrast, and distortion), temporal level (involving temporal characteristics such as image frame rate, dynamic range, and temporal stability), spectral level (involving spectral characteristics such as image color, brightness, and spectral response), and perceptual level (involving the perception and understanding of the image by the human visual system, including subjective evaluation of contrast, brightness, color, etc.). By establishing an association relationship between the imaging component factors and these levels, it is used to understand the contribution of each imaging component factor to the imaging quality. Step S230, based on the association relationship, project the imaging component factors into the corresponding multi-levels to obtain the imaging level. Specifically, the imaging component factors are projected into their corresponding multi-levels to form imaging levels, which reflect the different components of the imaging system and the interactions between them. This implementation method decomposes the imaging component factors into sub-factors at different levels and establishes their correlation with multiple levels, which intuitively reflects the contribution of each factor to the imaging quality and provides comprehensive information support for subsequent imaging quality optimization.

[0037] Step S300, performing an imaging quality dimension correlation analysis on the imaging level to obtain a positive quality factor and a negative quality factor of the imaging level. Specifically, for each imaging level, key dimensions that affect the imaging quality, such as clarity, contrast, noise, etc., are identified. The imaging influence relationship between the key dimensions and the imaging level is fitted to identify factors that have a positive impact on the imaging quality (positive quality factor) and factors that have a negative impact on the imaging quality (negative quality factor).

[0038] In one possible implementation, the imaging quality dimension correlation analysis is performed on the imaging level to obtain the positive quality factor and the reverse quality factor of the imaging level. Step S300 further includes step S310, identifying the key dimensions that affect the imaging quality for each imaging level. Specifically, for each imaging level, relevant imaging data is collected, including but not limited to sensor output, image processing results, environmental parameters, etc. These data are preprocessed to ensure the accuracy and consistency of the data. Statistical analysis methods are used to analyze the key factors that affect the imaging quality in each imaging level. For example, the physical level includes resolution, dynamic range, noise level, etc.; the signal processing level includes filtering effects, enhancement algorithm performance, etc.; the spatial level includes distortion degree, contrast, etc.; the temporal level includes frame rate stability, exposure time, etc.; the spectral level includes color accuracy, spectral response range, etc.; the perceptual level includes the human eye's perceptual evaluation of the image, etc.

[0039] Step S320, fit the key dimension and the single-layer imaging influence relationship of the level, and identify the single-layer positive quality factor and the single-layer reverse quality factor. Specifically, for the key dimensions of each imaging level, a mathematical model is established to describe the relationship between these dimensions and the imaging quality. Parameters of the established model are estimated using the collected imaging data and key dimension information. Based on the established model and the parameter estimation results, the single-layer positive quality factor (factor that has a positive impact on the imaging quality) and the single-layer reverse quality factor (factor that has a negative impact on the imaging quality) are identified. Step S330, fit the inter-layer imaging influence relationship between each imaging level, and identify the inter-layer positive quality factor and the inter-layer reverse quality factor. Specifically, based on the interaction between different imaging levels, a cross-layer mathematical model is established to describe the imaging influence relationship between these levels. Parameters of the established cross-layer model are estimated using the collected cross-layer imaging data and inter-level relationship information. Based on the established cross-layer model and parameter estimation results, inter-layer positive quality factors (factors at one layer have a positive impact on the imaging quality of another layer) and inter-layer negative quality factors (factors at one layer have a negative impact on the imaging quality of another layer) are identified.

[0040] Step S340, based on the imaging level, fuse the quality influence relationship of the single-layer forward quality factor, the single-layer reverse quality factor, the inter-layer forward quality factor, and the inter-layer reverse quality factor to obtain the forward quality factor and the reverse quality factor of each imaging level. Specifically, the identified single-layer forward / reverse quality factor and the inter-layer forward / reverse quality factor are comprehensively considered, and the comprehensive forward quality factor and the reverse quality factor of each imaging level are obtained by weighting, summing or other fusion methods. This implementation method comprehensively obtains the factors affecting the imaging quality by identifying key dimensions, fitting the influence relationship, identifying the quality factors and fusing them, and provides more accurate and specific guidance for the optimization of imaging quality.

[0041] Step S400, establish a compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor. Specifically, match the features in the multi-mode data feature description library with the quality factors in the imaging level. Analyze how each feature affects the forward quality factor and the reverse quality factor, and establish a compensation effect relationship between them. For example, in visible light imaging mode, the camera provides a high-resolution image, but the noise level is high under low-light conditions (reverse quality factor). To solve this problem, a low-noise sensor can be used to improve the image quality under low-light conditions, and an advanced noise reduction algorithm (forward quality factor) can be applied to reduce the noise in the image. At the same time, under low-light conditions, switch to infrared or thermal imaging modes, which are not limited by light and can provide additional information (forward quality factor). By fusing these multi-mode data, a synthetic image containing both high-resolution visible light information and useful information under low-light conditions can be created.

[0042] In a possible implementation, a compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor is established, and step S400 further includes step S410, obtaining the quality influence relationship between the forward quality factor and the reverse quality factor on the imaging level, performing basic gradient conversion on the quality influence relationship, and obtaining the imaging influence gradient of each quality factor. Specifically, the specific influence data of the forward quality factor and the reverse quality factor on the imaging quality at all imaging levels are collected, including laboratory test data, field test data or simulation-based data. The collected data is cleaned and preprocessed to ensure the accuracy and consistency of the data, including steps such as removing outliers, filling missing values, and smoothing data. Next, the influence relationship of each quality factor is subjected to basic gradient conversion, that is, the influence relationship of the quality factor is standardized for subsequent gradient projection and compensation analysis. The basic gradient conversion can be performed based on different scales, for example, linear transformation, logarithmic transformation or nonlinear transformation methods can be used. After completing the basic gradient conversion, the imaging influence gradient of each quality factor is calculated, and the imaging influence gradient represents the contribution or influence of the quality factor on the imaging quality. Step S420, extracting imaging features of the multimodal data corresponding to the imaging level according to the multimodal data feature description library. Specifically, accessing the constructed multimodal data feature description library, which contains the mapping relationship between various imaging levels and multimodal data. According to the requirements of the imaging level, extracting corresponding imaging features from the multimodal data feature description library, including color, texture, shape, temperature distribution, etc.

[0043] Step S430, based on the correspondence between the imaging features and the quality factors, the imaging features of the multi-mode data of each imaging level are projected with imaging influence gradients to obtain the imaging influence gradients of the multi-mode data. Specifically, the extracted imaging features are matched with the quality factors, and the specific influence of each feature on the quality factor is determined based on the correspondence between them. The imaging influence gradient of the quality factor obtained by the basic gradient conversion is used as a reference, and the influence of the imaging features is projected onto these gradients to quantify the contribution of the imaging features to the imaging quality, and the imaging influence gradient of each imaging feature is recorded for subsequent compensation analysis. Step S440, based on the imaging influence gradient of the multi-mode data, an imaging quality compensation analysis is performed to obtain a compensation effect relationship, and the compensation effect relationship has a gradient identifier. Specifically, the imaging influence gradients of each imaging feature are analyzed to determine which features contribute the most to the imaging quality. Based on the analysis results of the imaging influence gradients, a compensation strategy is determined, including enhancing certain imaging features, suppressing other imaging features, or adjusting imaging parameters. Apply the determined compensation strategy to perform imaging quality compensation analysis on the multi-mode data, optimize the imaging quality, and make the image clearer and more accurate. The impact of compensation strategy on image quality is recorded and analyzed to obtain the compensation effect relationship, which has a gradient mark and reflects the contribution of compensation strategy to image quality. This implementation method obtains the compensation effect relationship by comprehensively analyzing the relationship between the imaging characteristics of multi-mode data and the quality factor, providing an information basis for optimizing the imaging quality of hunting cameras.

[0044] Step S500, based on the compensation effect relationship, the imaging compensation optimization is performed on each data feature in the multi-mode data feature description library. Specifically, according to the compensation effect relationship, the compensation strategy for different imaging levels and imaging features is determined, including adjusting sensor parameters, optimizing optical system design, improving signal processing algorithms, fusing multi-mode data, etc. According to the determined compensation strategy, the data features to be adjusted are selected from the multi-mode data feature description library, including color, brightness, contrast, noise level, temperature distribution, etc. According to the gradient identifier provided in the compensation effect relationship, these features are quantitatively adjusted to achieve the purpose of optimizing the imaging quality. After applying the compensation strategy and adjusting the data features, the imaging quality of the hunting camera is re-evaluated through laboratory testing, field testing or simulation-based methods, and the evaluation content includes key imaging quality indicators such as image clarity, contrast, color reproduction, noise level, etc. According to the evaluation results of the imaging quality, the compensation strategy and data features are further adjusted to perform iterative optimization. This process needs to be repeated many times until satisfactory imaging quality is achieved. The embodiment of the present application collects data from different imaging modes, performs imaging feature analysis on the data of each imaging mode, establishes a mapping relationship between imaging features and imaging modes, constructs a multi-mode data feature description library containing imaging feature identifiers, analyzes the imaging quality of the hunting camera, identifies the component factors that affect the imaging quality, obtains different levels of the imaging system by hierarchical decomposition of these factors, identifies the key dimensions that affect the imaging quality for each imaging level, fits the imaging influence relationship between the key dimensions and the imaging level, identifies the positive quality factor and the reverse quality factor, matches the features in the multi-mode data feature description library with the quality factors in the imaging level, analyzes how each feature affects the positive quality factor and the reverse quality factor, establishes a compensation effect relationship between them, determines compensation strategies for different imaging levels and imaging features based on the compensation effect relationship, and achieves the technical effect of improving the imaging quality optimization effect and improving the robustness and accuracy of the imaging quality by applying the compensation strategy and other technical means.

[0045] In a possible implementation, based on the compensation effect relationship, imaging compensation optimization is performed on each data feature in the multi-mode data feature description library, and step S500 further includes step S510, based on the gradient identifier, feature gradient segmentation is performed on each data feature in the multi-mode data feature description library to construct multi-scale features. Specifically, the gradient identifier of each data feature is extracted from the compensation effect relationship, and these gradient identifiers reflect the degree of influence of the data feature on the imaging quality. According to the size of the gradient identifier, the data feature is divided into different gradient intervals. For example, the feature can be divided into intervals such as high-impact gradient, medium-impact gradient and low-impact gradient. In each gradient interval, the data feature is further subdivided to construct multi-scale features, including analyzing the change trend and stability of the feature at different scales, so as to more finely describe the influence of the feature on the imaging quality. Step S520, according to the compensation effect relationship corresponding to the gradient identifier, the multi-scale feature is subjected to scale feature mapping interaction, and imaging compensation optimization is performed according to the multi-scale interaction relationship to obtain hierarchical imaging parameters. Specifically, according to the compensation effect relationship corresponding to the gradient identifier, the multi-scale features are mapped and interacted, including analyzing the interaction between features at different scales and their influence on the imaging quality. Analyze the interaction between features of different scales, determine which features have synergistic effects in optimizing imaging quality and which features may conflict, and formulate imaging compensation optimization strategies based on multi-scale interaction relationships, including adjusting sensor parameters, optimizing optical system design, improving signal processing algorithms, fusing multi-mode data, etc., to optimize imaging quality. Through the implementation of the optimization strategy, the optimized hierarchical imaging parameters are obtained, which reflect the best contribution of data features to imaging quality at different imaging levels. This implementation method finely analyzes the impact of data features on imaging quality through feature gradient segmentation and multi-scale feature construction, improves the accuracy of compensation optimization, and ensures the accuracy and reliability of the optimization results.

[0046] In a possible implementation, the multi-scale features are subjected to scale feature mapping interaction, and imaging compensation optimization is performed according to the multi-scale interaction relationship to obtain hierarchical imaging parameters. Step S520 further includes step S521, sorting according to the parameter size relationship of the multi-scale features to determine the scale adjacency relationship. Specifically, from the multi-modal data feature description library, the features of all scales are extracted according to the constructed multi-scale features. The features of each scale are sorted according to their parameter values ​​(such as gradient size, influence degree, etc.) to determine the relative importance between the features and the adjacency relationship between them. According to the sorting result, the features of adjacent scales are paired to form scale adjacency relationships, and these relationships are used for subsequent feature interaction processes. Step S522, according to the scale adjacency relationship, adjacent scale feature interactions are performed in sequence, wherein the adjacent scale feature interactions include: inputting the features of two adjacent scales into the feature enhancement layer through the attention mechanism to perform feature enhancement to obtain a first enhanced feature and a second enhanced feature; performing inner product calculation and one softmax calculation on the first enhanced feature and the second enhanced feature to obtain the corresponding relationship between the two adjacent scale features, and performing two scale feature fusion based on the corresponding relationship to obtain a fusion feature. Specifically, the features of two adjacent scales are input into the feature enhancement layer, which enhances the features through the attention mechanism (a neural network mechanism that can automatically learn and focus on the important parts of the input data) to highlight the important features and suppress the minor features, thereby obtaining the first enhanced feature and the second enhanced feature. The first enhanced feature refers to the feature of the first scale after being processed by the feature enhancement layer; the second enhanced feature refers to the feature of the second scale after being processed by the feature enhancement layer. The first enhanced feature and the second enhanced feature are inner-producted, that is, the corresponding elements of the two vectors are multiplied and summed to obtain the similarity between them. Then, the similarity is converted into a probability distribution through the softmax function (a function that converts a real vector into a probability distribution) to represent the correspondence between the two scale features. Based on the correspondence, the features of the two scales are fused, and the fusion process can adopt weighted average, maximum value selection or other fusion strategies to obtain the fused features.

[0047] Step S523, complete the interaction of features of all adjacent scales according to the scale adjacency relationship, perform imaging compensation optimization evaluation on the fused features, and obtain the hierarchical imaging parameters, which are the features of each level with the best imaging compensation optimization evaluation results. Specifically, according to the scale adjacency relationship, the features of adjacent scales are interacted in turn until the features of all scales have completed the interaction. The fused features are evaluated for imaging compensation optimization, and the evaluation process can use laboratory tests, field tests, or simulation-based methods to evaluate the degree of improvement in imaging quality. According to the imaging compensation optimization evaluation results, the best features of each level are selected as hierarchical imaging parameters. These parameters reflect the best contribution of data features to imaging quality at different imaging levels, and are used to guide the imaging quality optimization process of hunting cameras. This implementation method makes full use of the complementarity between features of different scales through the interaction and fusion of adjacent scale features, and improves the overall optimization effect of imaging quality.

[0048] In the above, refer to Figure 1 The method for optimizing the imaging quality of a hunting camera based on multi-mode data fusion according to an embodiment of the present invention is described in detail. Figure 2 A hunting camera imaging quality optimization system based on multi-mode data fusion according to an embodiment of the present invention is described.

[0049] The hunting camera imaging quality optimization system based on multi-mode data fusion according to the embodiment of the present invention is used to solve the technical problems of limited optimization effect, insufficient robustness and accuracy of imaging quality of existing hunting cameras, so as to achieve the technical effect of improving imaging quality optimization effect, as well as improving robustness and accuracy of imaging quality. The hunting camera imaging quality optimization system based on multi-mode data fusion includes: a multi-mode data feature description library establishment module 10, an imaging component factor hierarchical decomposition module 20, an imaging quality dimension correlation analysis module 30, a compensation effect relationship establishment module 40, and an imaging compensation optimization module 50.

[0050] The multimodal data feature description library establishment module 10 is used to obtain the data source characteristics of the multimodal data and establish the multimodal data feature description library; the imaging component factor hierarchical decomposition module 20 is used to parse the imaging component factors, perform hierarchical decomposition on the imaging component factors, and obtain the imaging level; the imaging quality dimension correlation analysis module 30 is used to perform imaging quality dimension correlation analysis on the imaging level, and obtain the positive quality factor and reverse quality factor of the imaging level; the compensation effect relationship establishment module 40 is used to establish the compensation effect relationship between each data feature in the multimodal data feature description library and the positive quality factor and reverse quality factor; the imaging compensation optimization module 50 is used to perform imaging compensation optimization on each data feature in the multimodal data feature description library based on the compensation effect relationship.

[0051] The specific configuration of the multi-modal data feature description library establishment module 10 will be described in detail below. As described above, a multi-modal data feature description library is established, and the multi-modal data feature description library establishment module 10 may further include: an imaging feature analysis unit is used for the multi-modal data including at least visible light imaging data, infrared imaging data, and thermal imaging data, respectively, to analyze the imaging advantage features and imaging disadvantage features of the multi-modal data to obtain the imaging features of each multi-modal data; a mapping relationship establishment unit is used to establish the mapping relationship between the imaging features and the imaging level, and the mapping relationship between the imaging features and the multi-modal data; a multi-modal data feature description library construction unit is used to map and integrate the imaging features, imaging levels and multi-modal data based on the mapping relationship, and construct the multi-modal data feature description library, wherein the multi-modal data feature description library has feature identifiers of imaging advantage features and imaging disadvantage features.

[0052] The specific configuration of the imaging component factor hierarchical decomposition module 20 will be described in detail below. As described above, the imaging component factors are analyzed, the imaging component factors are hierarchically decomposed, and the imaging level is obtained. The imaging component factor hierarchical decomposition module 20 may further include: an imaging quality factor analysis unit for analyzing the imaging quality factor of the hunting camera according to the sensor performance, optical system characteristics, and environmental conditions to obtain the imaging component factors; an association relationship establishment unit for decomposing the imaging component factors according to the physical, signal processing, space, time, spectrum, and perception multiple levels to establish the association relationship between the imaging component factors and the multiple levels; an imaging level acquisition unit for projecting the imaging component factors into the corresponding multiple levels based on the association relationship to obtain the imaging level.

[0053] The specific configuration of the imaging quality dimension correlation analysis module 30 will be described in detail below. As described above, the imaging quality dimension correlation analysis is performed on the imaging level to obtain the positive quality factor and the negative quality factor of the imaging level. The imaging quality dimension correlation analysis module 30 may further include: a key dimension identification unit for identifying the key dimension affecting the imaging quality for each imaging level; a single-layer imaging influence relationship fitting unit for fitting the single-layer imaging influence relationship between the key dimension and the level, identifying the single-layer positive quality factor and the single-layer negative quality factor; an inter-layer imaging influence relationship fitting unit for fitting the inter-layer imaging influence relationship between each imaging level, identifying the inter-layer positive quality factor and the inter-layer negative quality factor; a quality influence relationship fusion unit for performing quality influence relationship fusion on the single-layer positive quality factor, the single-layer negative quality factor, the inter-layer positive quality factor, and the inter-layer negative quality factor based on the imaging level, and obtaining the positive quality factor and the negative quality factor of each imaging level.

[0054] The specific configuration of the compensation effect relationship establishment module 40 will be described in detail below. As described above, the compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor is established, and the compensation effect relationship establishment module 40 may further include: a basic gradient conversion unit is used to obtain the quality influence relationship of the forward quality factor and the reverse quality factor on the imaging level, and perform basic gradient conversion on the quality influence relationship to obtain the imaging influence gradient of each quality factor; an imaging feature extraction unit is used to extract the imaging features of the multi-mode data corresponding to the imaging level according to the multi-mode data feature description library; an imaging influence gradient projection unit is used to perform imaging influence gradient projection on the imaging features of the multi-mode data of each imaging level based on the corresponding relationship between the imaging features and the quality factor to obtain the imaging influence gradient of the multi-mode data; an imaging quality compensation analysis unit is used to perform imaging quality compensation analysis based on the imaging influence gradient of the multi-mode data to obtain the compensation effect relationship, and the compensation effect relationship has a gradient identifier.

[0055] The specific configuration of the imaging compensation optimization module 50 will be described in detail below. As described above, based on the compensation effect relationship, the imaging compensation optimization is performed on each data feature in the multi-mode data feature description library, and the imaging compensation optimization module 50 may further include: a feature gradient segmentation unit is used to perform feature gradient segmentation on each data feature in the multi-mode data feature description library based on the gradient identifier to construct a multi-scale feature; a scale feature mapping interaction unit is used to perform scale feature mapping interaction on the multi-scale feature according to the compensation effect relationship corresponding to the gradient identifier, perform imaging compensation optimization according to the multi-scale interaction relationship, and obtain hierarchical imaging parameters.

[0056] Among them, scale feature mapping interaction is performed on the multi-scale features, imaging compensation optimization is performed according to the multi-scale interaction relationship, and hierarchical imaging parameters are obtained. The scale feature mapping interaction unit may further include: a scale adjacency relationship acquisition subunit is used to sort according to the parameter size relationship of the multi-scale features to determine the scale adjacency relationship; an adjacent scale feature interaction subunit is used to perform adjacent scale feature interaction in sequence according to the scale adjacency relationship, wherein the adjacent scale feature interaction includes: inputting the features of two adjacent scales into the feature enhancement layer through the attention mechanism to perform feature enhancement to obtain a first enhanced feature and a second enhanced feature; performing inner product calculation and a softmax calculation on the first enhanced feature and the second enhanced feature to obtain a corresponding relationship between two adjacent scale features, and fusing the two scale features based on the corresponding relationship to obtain a fused feature; the hierarchical imaging parameter acquisition unit is used to complete the feature interaction of all adjacent scales according to the scale adjacency relationship, perform imaging compensation optimization evaluation on the fused features, and obtain the hierarchical imaging parameters, and the hierarchical imaging parameters are the hierarchical features with the best imaging compensation optimization evaluation results.

[0057] The hunting camera imaging quality optimization system based on multi-mode data fusion provided in the embodiment of the present invention can execute the hunting camera imaging quality optimization method based on multi-mode data fusion provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0058] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0059] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A hunting camera imaging quality optimization method based on multi-mode data fusion, characterized in that: The method comprises: Obtain the data source characteristics of multi-modal data and establish a multi-modal data feature description library; Analyzing imaging component factors, performing hierarchical decomposition on the imaging component factors to obtain imaging levels; Performing an imaging quality dimension correlation analysis on the imaging level to obtain a positive quality factor and a negative quality factor of the imaging level; Establishing a compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor; Based on the compensation effect relationship, performing imaging compensation optimization on each data feature in the multi-mode data feature description library; Performing an imaging quality dimension correlation analysis on the imaging level to obtain a positive quality factor and a negative quality factor of the imaging level includes: Identify the key dimensions that affect image quality at each imaging level; Fitting the relationship between the key dimension and the single-layer imaging influence at the level, and identifying the single-layer positive quality factor and the single-layer reverse quality factor; Fit the inter-layer imaging influence relationship between each imaging level, and identify the inter-layer positive quality factor and inter-layer negative quality factor; Based on the imaging level, the quality influence relationship of the single-layer forward quality factor, the single-layer reverse quality factor, the inter-layer forward quality factor, and the inter-layer reverse quality factor are fused to obtain the forward quality factor and the reverse quality factor of each imaging level; Establishing a compensation effect relationship between each data feature in the multi-mode data feature description library and the forward quality factor and the reverse quality factor, including: Obtaining the quality influence relationship between the forward quality factor and the reverse quality factor on the imaging level, performing basic gradient conversion on the quality influence relationship, and obtaining the imaging influence gradient of each quality factor; Extracting imaging features of the multimodal data corresponding to the imaging level according to the multimodal data feature description library; Based on the corresponding relationship between imaging features and quality factors, the imaging features of multi-mode data at each imaging level are projected with imaging influence gradients to obtain the imaging influence gradients of multi-mode data; An imaging quality compensation analysis is performed based on the imaging influence gradient of the multi-mode data to obtain a compensation effect relationship, wherein the compensation effect relationship has a gradient identifier.

2. The hunting camera imaging quality optimization method based on multi-mode data fusion as claimed in claim 1, characterized in that: The establishing of a multi-mode data feature description library comprises: The multi-mode data at least includes visible light imaging data, infrared imaging data, and thermal imaging data, and imaging advantage characteristics and imaging disadvantage characteristics of the multi-mode data are analyzed respectively to obtain imaging characteristics of each multi-mode data; Establishing a mapping relationship between the imaging feature and the imaging level, and a mapping relationship between the imaging feature and the multi-mode data; Based on the mapping relationship, imaging features, imaging levels and multimodal data are mapped and integrated to construct the multimodal data feature description library, wherein the multimodal data feature description library has feature identifiers of imaging advantage features and imaging disadvantage features.

3. The hunting camera imaging quality optimization method based on multi-mode data fusion as claimed in claim 2, characterized in that: The analyzing imaging component factors and performing hierarchical decomposition on the imaging component factors to obtain imaging levels include: The imaging quality factors of hunting cameras are analyzed according to sensor performance, optical system characteristics, and environmental conditions to obtain imaging component factors; Decomposing the imaging component factors according to multiple levels of physics, signal processing, space, time, spectrum, and perception, and establishing correlations between the imaging component factors and the multiple levels; Based on the association relationship, the imaging component factors are projected into corresponding multi-levels to obtain the imaging level.

4. The hunting camera imaging quality optimization method based on multi-mode data fusion as claimed in claim 1, characterized in that: Based on the compensation effect relationship, imaging compensation optimization is performed on each data feature in the multi-mode data feature description library, including: Based on the gradient identifier, feature gradient segmentation is performed on each data feature in the multimodal data feature description library to construct multi-scale features; According to the compensation effect relationship corresponding to the gradient identifier, scale feature mapping interaction is performed on the multi-scale features, and imaging compensation optimization is performed according to the multi-scale interaction relationship to obtain hierarchical imaging parameters.

5. The hunting camera imaging quality optimization method based on multi-mode data fusion as claimed in claim 4, characterized in that: Performing scale feature mapping interaction on the multi-scale features, performing imaging compensation optimization according to the multi-scale interaction relationship, and obtaining hierarchical imaging parameters, including: Sort by the size relationship of the parameters of the multi-scale features to determine the scale adjacency relationship; According to the scale adjacency relationship, adjacent scale feature interactions are performed in sequence, wherein the adjacent scale feature interactions include: inputting the features of two adjacent scales into a feature enhancement layer for feature enhancement through an attention mechanism to obtain a first enhanced feature and a second enhanced feature; performing an inner product calculation and a softmax calculation on the first enhanced feature and the second enhanced feature to obtain a corresponding relationship between the two adjacent scale features, and fusing the two scale features based on the corresponding relationship to obtain a fused feature; The feature interactions of all adjacent scales are completed according to the scale adjacency relationship, and the imaging compensation optimization evaluation is performed on the fusion features to obtain the hierarchical imaging parameters, which are the hierarchical features with the best imaging compensation optimization evaluation results.

6. Hunting camera imaging quality optimization system based on multi-mode data fusion, characterized in that: The system is used to implement the hunting camera imaging quality optimization method based on multi-mode data fusion according to any one of claims 1 to 5, and the system comprises: A multi-mode data feature description library establishment module, the multi-mode data feature description library establishment module is used to obtain data source features of multi-mode data and establish a multi-mode data feature description library; An imaging component factor hierarchical decomposition module, the imaging component factor hierarchical decomposition module is used to analyze the imaging component factors, perform hierarchical decomposition on the imaging component factors, and obtain imaging levels; An imaging quality dimension correlation analysis module, the imaging quality dimension correlation analysis module is used to perform imaging quality dimension correlation analysis on the imaging level to obtain a positive quality factor and a negative quality factor of the imaging level; A compensation effect relationship establishment module, the compensation effect relationship establishment module is used to establish a compensation effect relationship between each data feature in the multimodal data feature description library and the forward quality factor and the reverse quality factor; An imaging compensation optimization module is used to perform imaging compensation optimization on each data feature in the multi-mode data feature description library based on the compensation effect relationship.

Images