A method and system for infrared image optimization based on data analysis

Abstract

The present invention relates to an infrared image optimization method and system based on data analysis, and belongs to the field of infrared image optimization technology. First, several scene type databases are set to store historical infrared images and corresponding on-site images; by graying and edge recognition of historical on-site images, a historical edge line positioning map is obtained, and alignment and adaptation with historical infrared images are performed; edge factor analysis is performed on historical infrared images, edge factor reference segments and their interval characteristics are determined, and edge factor reference segment feature data sets are constructed; based on the relative position relationship of edge lines in the historical edge line positioning map, an association relationship between edge factor reference segment feature data sets is established; when optimizing real-time infrared images, the corresponding edge factor reference segment feature data sets are called to improve and supplement the intervals between edge factors. Through the above technical scheme, the present invention realizes the supplement of edges in infrared images and improves the overall quality of images.

Description

Technical Field

[0001] The invention relates to an infrared image optimization method and system based on data analysis, belonging to the technical field of infrared image optimization. Background Art

[0002] In the field of infrared image processing, infrared images, as a special type of image, are widely used in many industries such as military, aerospace, security, medicine and electricity. However, due to the characteristics of infrared imaging systems and the influence of the external environment, infrared images often have blurred edges and unclear details. This problem directly affects the accuracy and efficiency of images in subsequent analysis, target detection and recognition applications. Summary of the invention

[0003] The purpose of the present invention is to solve the deficiencies of the above-mentioned prior art and to provide an infrared image optimization method and system based on data analysis. The present invention realizes the supplementation of edges in infrared images and improves the overall quality of the images.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] A method for optimizing infrared images based on data analysis, which is special in that it includes the following steps:

[0006] Step S100, for the scene to which the infrared image is applied, several scene type databases are set, and the historical infrared images and historical on-site images are stored in the corresponding scene type database in the form of historical image groups;

[0007] Step S200, graying the historical scene images in the historical image group to obtain a historical scene grayscale image, and using edge recognition technology to identify the edges in the historical scene grayscale image, and perform edge delineation, projecting the edge delineation result onto the positioning template image to obtain a historical edge line positioning image, and aligning and adapting the historical edge line positioning image with the historical infrared images in the historical image group;

[0008] Step S300, performing edge factor analysis on the historical infrared image, determining the edge factor reference segment in the historical infrared image, associating the edge factor reference segment with the corresponding edge line in the historical edge line positioning map, and determining the interval characteristics between the edge factor reference segments based on the shape characteristics of the edge lines, recording the edge factor reference segments and the interval characteristics between them as an edge factor reference segment feature data group, and constructing the edge factor reference segment feature data group as an edge factor reference segment feature data set;

[0009] Step S400, based on the relative positional relationship between adjacent edge lines in the historical edge line positioning map, the relative positional relationship between edge factor reference segment feature data groups is recorded, and an association relationship between edge factor reference segment feature data groups is established;

[0010] Step S500, in the process of optimizing the real-time infrared image, the application scenario of the real-time infrared image is determined, and the corresponding edge factor reference segment feature data set is called based on the application scenario, and then the edge factor analysis is performed on the real-time infrared image, and based on the analysis results, the intervals between the edge factor reference segments are improved and supplemented.

[0011] Preferably, in step S200, the method for identifying edges in the historical scene grayscale image includes:

[0012] Applying edge detection operators to judge the gradient values ​​of pixel points in the historical scene grayscale image, and determining edge points in the historical scene grayscale image;

[0013] Based on the position and gradient direction of each edge point, the edge points are connected to form the edge in the historical scene grayscale image.

[0014] Preferably, in step S300, the method of performing edge factor analysis on the historical infrared image includes:

[0015] Grayscale the historical infrared image to obtain a historical infrared grayscale image, and use edge recognition technology to determine the edge factor segment in the historical infrared grayscale image and mark it in the historical infrared grayscale image;

[0016] The historical infrared grayscale image is compared with the historical edge positioning image, and the edge lines in the historical edge positioning image are mapped to the historical infrared grayscale image. The edge factor segments that do not overlap with the edge lines are eliminated, and the remaining edge factor segments are identified as edge factor reference segments.

[0017] Preferably, in step S300, the method for determining the interval characteristics between edge factor reference segments based on the shape characteristics of the edge line includes:

[0018] An image coordinate system is constructed for the historical infrared image, and based on a preset interval, a number of edge detection points are set on each edge factor reference segment, the point coordinates of each edge detection point are determined, and the average point coordinates of all point coordinates corresponding to the edge factor reference segment are calculated;

[0019] The average point coordinates of the edge factor reference segments mapped to the same edge line are associated, and the vector between each two adjacent average point coordinates is calculated, and the vectors are combined and recorded as the edge factor reference segment vector group;

[0020] Determine the interval reference segment between the edge factor reference segments on the edge line, and based on the preset interval, set a number of interval detection points on the interval reference segment, and determine the point coordinates of each interval detection point, and calculate the vector between each adjacent interval detection point, and combine the vectors to obtain a vector group in the interval reference segment;

[0021] Based on the position of the edge factor reference segment relative to the edge line, the vector group between the edge factor reference segments and the vector group in the interval reference segment are associated in sequence to obtain the edge factor reference segment feature data group.

[0022] Preferably, in step S500, the method for performing edge factor analysis on the real-time infrared image includes:

[0023] Grayscale processing is performed on the real-time infrared image to obtain a real-time infrared grayscale image, and edge recognition technology is used to determine the edge factor reference segment in the real-time infrared image;

[0024] Dynamic block scanning is performed on the edge factor reference segment in the real-time infrared image, including gradually changing the position of the preset block in the real-time infrared image, and performing relative position analysis on the edge factor reference segment in the block, and searching the analysis results in the edge factor reference segment feature data set to determine a number of matching edge factor reference segment feature data groups.

[0025] Preferably, the method of retrieving the analysis result in the edge factor reference segment feature data set includes:

[0026] A number of edge detection points are set for the edge factor reference segment in the block, and the point coordinates of each edge detection point are determined, and the average point coordinates of the edge factor reference segment in the block are determined, and the vectors between the average point coordinates of each adjacent edge factor reference segment are calculated, and the vectors are combined to obtain a vector group between edge factor reference segments;

[0027] Based on the vector group between edge factor reference segments in the block, the relative directions between edge factor reference segments in the block are determined, and the relative directions between edge factor reference segments are combined into a relative direction sequence;

[0028] The relative direction sequence between edge factor reference segments in the block is used as a retrieval label and retrieved in the edge factor reference segment feature dataset.

[0029] Preferably, the method for performing edge factor analysis on the real-time infrared image comprises:

[0030] The edge factor reference segment feature data groups retrieved and screened are respectively compared with the corresponding edge factor reference segment vector groups in the block to form several comparison groups, and the primary matching parameters of each comparison group are respectively determined. If the primary matching parameters are greater than or equal to the preset values, the edge factor reference segment vector groups in the comparison group are identified as valid edge factor reference segment vector groups, otherwise the edge factor reference segment feature data groups in the comparison group are replaced;

[0031] Based on the primary matching parameters of each comparison group and the relative position relationship between the effective edge factor reference segment vector groups, the secondary matching parameters of all comparison groups are determined. If the secondary matching parameters are greater than or equal to the preset value, the edge factor reference feature data groups in all comparison groups are identified as matching edge factor reference feature data groups.

[0032] Based on the determined matching edge factor reference feature data set, the intervals between edge factor reference segments in the real-time infrared image are supplemented, including using the vector group in the interval reference segment in the matching edge factor reference feature data set as the interval supplement basis.

[0033] Preferably, the method for calculating the primary matching parameter of the comparison group includes:

[0034] Perform several offset correspondences on the edge factor reference segment vector group in the block relative to the edge factor reference segment feature data group, and after each offset correspondence, determine the number of corresponding identical vectors in the edge factor reference segment vector group in the block, and calculate the ratio of the number of identical vectors to the total number of vectors;

[0035] Based on the equivalent vector ratio of the vector group between the edge factor reference segments and the total number of vectors, the first-level matching parameters of the comparison group are determined;

[0036] Among them, the expression for calculating the first-level matching parameter of the comparison group is:

[0037] ;

[0038] in, is the first-level matching parameter of the comparison group, is the matching parameter conversion coefficient, is the number of equivalent vectors, is the total number of vectors, is the total vector quantity impact adjustment coefficient, A constant for total vector quantity effect scaling.

[0039] Preferably, the method for determining the secondary matching parameters of all comparison groups comprises:

[0040] Determine the relative position relationship between the effective edge factor reference segment vector groups in the block, and the position matching number of the relative position relationship between the corresponding edge factor reference segment feature data groups, and calculate the matching ratio of the position matching number to the number of vector groups of the effective edge factor reference segment vector groups;

[0041] The first-level matching parameters are corrected based on the matching ratio to obtain the second-level matching parameters;

[0042] Among them, the expression of the secondary matching parameter is:

[0043] ;

[0044] in, is the secondary matching parameter, is the number of position matches, is the number of vector groups, R is the matching ratio impact adjustment coefficient, Adjust the constant to match the ratio effect.

[0045] The present invention also discloses an infrared image optimization system based on data analysis, comprising:

[0046] The first module is used to set a number of scene type databases according to the scenes to which the infrared images are applied, and store the historical infrared images and the historical on-site images in the corresponding scene type databases in the form of historical image groups;

[0047] The second module is used to grayscale the historical scene images in the historical image group to obtain the historical scene grayscale map, and use the edge recognition technology to identify the edges in the historical scene grayscale map, and perform edge delineation, project the edge delineation result onto the positioning template map, obtain the historical edge line positioning map, and align and adapt the historical edge line positioning map with the historical infrared images in the historical image group;

[0048] The third module is used to perform edge factor analysis on the historical infrared image, determine the edge factor reference segment in the historical infrared image, associate the edge factor reference segment with the corresponding edge line in the historical edge line positioning map, and determine the interval characteristics between the edge factor reference segments based on the shape characteristics of the edge lines, record the edge factor reference segments and the interval characteristics between each other as an edge factor reference segment feature data group, and construct the edge factor reference segment feature data group into an edge factor reference segment feature data set, based on the relative position relationship between adjacent edge lines in the historical edge line positioning map, record the relative position relationship between the edge factor reference segment feature data groups, and establish the association relationship between the edge factor reference segment feature data groups;

[0049] The fourth module is used to determine the application scenario of the real-time infrared image during the optimization process of the real-time infrared image, and call the corresponding edge factor reference segment feature data set based on the application scenario, and then perform edge factor analysis on the real-time infrared image, and based on the analysis results, improve and supplement the intervals between the edge factor reference segments.

[0050] The invention discloses an infrared image optimization method and system based on data analysis, and relates to the technical field of infrared image optimization. Firstly, a plurality of scene type databases are set to store historical infrared images and corresponding on-site images; by graying and edge recognition of the historical on-site images, a historical edge line positioning map is obtained, and the historical infrared images are aligned and adapted; edge factor analysis is performed on the historical infrared images, edge factor reference segments and their interval characteristics are determined, and an edge factor reference segment feature data set is constructed; based on the relative position relationship of edge lines in the historical edge line positioning map, an association relationship between edge factor reference segment feature data sets is established; when optimizing real-time infrared images, the corresponding edge factor reference segment feature data set is called to improve and supplement the intervals between edge factors. The invention realizes the supplement of edges in infrared images and improves the overall quality of images through the above technical scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a method step diagram of an infrared image optimization analysis method based on data analysis of the present invention. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0053] An image optimization method of this embodiment that can optimize and improve the content of infrared images, see Figure 1 , including the following steps:

[0054] Step S100, for the scenes to which the infrared images are applied, several scene type databases are set, and the historical infrared images and historical on-site images are stored in the corresponding scene type databases in the form of historical image groups.

[0055] Infrared images have various application scenarios, and the image features in different scenarios are different. Therefore, the present invention first sets up several scene type databases according to the application scenarios of infrared images. These databases are used to store historical infrared images and corresponding on-site images, providing a rich data basis for subsequent edge recognition and factor analysis. By storing images in the form of historical image groups, relevant data can be easily managed and called.

[0056] Step S200, grayscale the historical scene images in the historical image group to obtain a historical scene grayscale map, and use edge recognition technology to identify the edges in the historical scene grayscale map and perform edge delineation, project the edge delineation result onto the positioning template map to obtain a historical edge line positioning map, and align and adapt the historical edge line positioning map and the historical infrared images in the historical image group.

[0057] In order to simplify the image processing process and highlight the edge information, the present invention grayscales the on-site images in the historical image group to obtain a historical on-site grayscale image. Subsequently, edge recognition technology is used to accurately identify and delineate the edges in the grayscale image. Edge recognition technology detects edges based on brightness changes in images and is a common method in image processing. The result of edge delineation is projected onto a positioning template image to obtain a historical edge line positioning map, thereby achieving accurate extraction and positioning of edge information. Finally, the historical edge line positioning map is aligned and adapted with the corresponding historical infrared image to provide an accurate reference for subsequent edge factor analysis.

[0058] The specific steps include: image preprocessing, performing denoising, smoothing and other preprocessing operations on infrared images to improve the accuracy of edge recognition; edge detection, selecting a suitable edge detection operator to perform edge detection on the preprocessed image; according to the characteristics and requirements of the infrared image, the Canny operator or other suitable operators can be selected; edge connection and post-processing, connecting the detected edge points to form edge lines, and post-processing the edge lines (such as denoising, refinement, etc.) to obtain clearer and more accurate edge information.

[0059] In some embodiments disclosed in the present invention, the method for identifying edges in a historical scene grayscale image includes:

[0060] Step S201, applying an edge detection operator to judge the gradient values ​​of the pixel points in the historical scene grayscale image, and determining the edge points in the historical scene grayscale image.

[0061] Apply edge detection operator. Edge detection operator is an algorithm used to detect image edges in image processing. In the present invention, the edge points in the image can be determined by judging the gradient value of the pixel points in the historical scene grayscale image. The gradient value indicates the degree of brightness change of the pixel point. The larger the gradient value, the more likely the point is an edge point.

[0062] Step S202, based on the position and gradient direction of each edge point, the edge points are connected to form an edge in the historical scene grayscale image.

[0063] Step S300, perform edge factor analysis on the historical infrared image, determine the edge factor reference segment in the historical infrared image, associate the edge factor reference segment with the corresponding edge line in the historical edge line positioning map, and determine the interval characteristics between the edge factor reference segments based on the shape characteristics of the edge line, record the edge factor reference segments and the interval characteristics between them as an edge factor reference segment feature data group, and construct the edge factor reference segment feature data group into an edge factor reference segment feature data set.

[0064] Edge factor refers to various factors that affect the edge clarity of infrared images, including edge shape, brightness contrast, etc. The present invention conducts an in-depth edge factor analysis on historical infrared images and determines the edge factor reference segments in the image. These reference segments are areas with obvious edge features in the image, which are of great significance for subsequent edge improvement and supplementation. By associating the edge factor reference segments with the corresponding edge lines in the historical edge line positioning map, a corresponding relationship between the edge factor and the actual edge can be established. At the same time, based on the shape characteristics of the edge lines, the interval characteristics between the edge factor reference segments are determined, which provides important reference information for subsequent real-time image processing. These feature data are recorded as edge factor reference segment feature data groups and constructed as edge factor reference segment feature data sets, which provides strong support for subsequent optimization processing.

[0065] In some embodiments disclosed in the present invention, the method of performing edge factor analysis on historical infrared images includes:

[0066] Step S301 , grayscale the historical infrared image to obtain a historical infrared grayscale image, and use edge recognition technology to determine the edge factor segment in the historical infrared grayscale image and mark it in the historical infrared grayscale image.

[0067] Step S302, compare the historical infrared grayscale image with the historical edge positioning image, map the edge lines in the historical edge positioning image to the historical infrared grayscale image, remove the edge factor segments that do not overlap with the edge lines, and identify the remaining edge factor segments as edge factor reference segments.

[0068] In some embodiments disclosed in the present invention, a method for determining the interval characteristics between edge factor reference segments based on the shape characteristics of the edge line includes:

[0069] Step S303, constructing an image coordinate system for the historical infrared image, and setting a number of edge detection points on each edge factor reference segment based on a preset interval, determining the point coordinates of each edge detection point, and calculating the average point coordinates of all point coordinates corresponding to the edge factor reference segment;

[0070] Step S304, associating the average point coordinates of the edge factor reference segments mapped to the same edge line, calculating the vector between every two adjacent average point coordinates, and combining the vectors to be recorded as an edge factor reference segment vector group;

[0071] Step S305, determining the interval reference segment between the edge factor reference segments on the edge line, and based on the preset interval, setting a number of interval detection points on the interval reference segment, and determining the point coordinates of each interval detection point, and calculating the vector between each adjacent interval detection point, and combining the vectors to obtain a vector group in the interval reference segment;

[0072] Step S306 , based on the position of the edge factor reference segment relative to the edge line, the vector group between the edge factor reference segments and the vector group in the interval reference segment are associated in sequence to obtain the edge factor reference segment feature data group.

[0073] Step S400, based on the relative positional relationship between adjacent edge lines in the historical edge line positioning map, the relative positional relationship between edge factor reference segment feature data groups is recorded, and an association relationship between edge factor reference segment feature data groups is established.

[0074] In infrared images, there is often a specific relative position relationship between adjacent edge lines. Based on the relative position relationship between adjacent edge lines in the historical edge line positioning map, the present invention records the relative position relationship between edge factor reference segment feature data groups in detail and establishes the association relationship between them. The implementation of this step enables the missing or blurred edge information to be inferred and improved according to the position relationship of adjacent edge lines in the subsequent real-time image processing process, thereby further improving the edge clarity and overall quality of the image.

[0075] Step S500, in the process of optimizing the real-time infrared image, the application scenario of the real-time infrared image is determined, and the corresponding edge factor reference segment feature data set is called based on the application scenario, and then the edge factor analysis is performed on the real-time infrared image, and based on the analysis results, the intervals between the edge factor reference segments are improved and supplemented.

[0076] In the process of optimizing the real-time infrared image, the present invention first determines the application scenario of the image, and calls the corresponding edge factor reference segment feature data set based on the scenario. In this way, targeted optimization processing can be performed according to the image features in different scenarios. Subsequently, the real-time infrared image is subjected to edge factor analysis to identify the edge factor reference segments and their interval features in the image. Based on the analysis results, the intervals between the edge factor reference segments are improved and supplemented, that is, the missing or blurred edge information in the real-time image is inferred and supplemented based on the edge features in the historical data set. This step can significantly improve the edge clarity and overall quality of the real-time infrared image, and provide high-quality image data support for subsequent applications.

[0077] In some embodiments disclosed in the present invention, the method for performing edge factor analysis on a real-time infrared image includes:

[0078] Step S501 , grayscale processing is performed on the real-time infrared image to obtain a real-time infrared grayscale image, and edge recognition technology is used to determine the edge factor reference segment in the real-time infrared image.

[0079] Step S502, dynamically scanning the edge factor reference segments in the real-time infrared image, including gradually changing the position of the preset blocks in the real-time infrared image, and performing relative position analysis on the edge factor reference segments in the blocks, and searching the analysis results in the edge factor reference segment feature data set to determine a number of matching edge factor reference segment feature data groups.

[0080] In some embodiments disclosed in the present invention, a method for retrieving the analysis result in an edge factor reference segment feature data set includes:

[0081] Step S5021, set a number of edge detection points for the edge factor reference segment in the block, determine the point coordinates of each edge detection point, and determine the average point coordinates of the edge factor reference segment in the block, calculate the vector between the average point coordinates of each adjacent edge factor reference segment, and combine the vectors to obtain a vector group between edge factor reference segments.

[0082] Step S5022: Based on the vector group between edge factor reference segments in the block, determine the relative directions between edge factor reference segments in the block, and combine the relative directions between edge factor reference segments into a relative direction sequence.

[0083] Step S5023: The relative direction sequence between the edge factor reference segments in the block is used as a search tag, and search is performed in the edge factor reference segment feature data set.

[0084] In some embodiments disclosed in the present invention, the method for performing edge factor analysis on a real-time infrared image includes:

[0085] Step S503, the edge factor reference segment feature data groups retrieved and screened are compared with the corresponding edge factor reference segment vector groups in the block to form several comparison groups, and the first-level matching parameters of each comparison group are determined respectively. If the first-level matching parameters are greater than or equal to the preset values, the edge factor reference segment vector groups in the comparison group are identified as valid edge factor reference segment vector groups, otherwise the edge factor reference segment feature data groups in the comparison group are replaced.

[0086] In some embodiments disclosed in the present invention, the method for calculating the primary matching parameter of the comparison group includes:

[0087] Step S5031, perform several offset correspondences on the edge factor reference segment vector group within the block relative to the edge factor reference segment feature data group, and after each offset correspondence, determine the number of corresponding identical vectors in the edge factor reference segment vector group within the block, and calculate the ratio of the number of identical vectors to the total number of identical vectors.

[0088] Step S5032, based on the equivalent vector ratio of the edge factor reference segment vector group and the total number of vectors, determine the first-level matching parameter of the comparison group.

[0089] Among them, the expression for calculating the first-level matching parameter of the comparison group is:

[0090] .

[0091] in, is the first-level matching parameter of the comparison group, is the matching parameter conversion coefficient, is the number of equivalent vectors, is the total number of vectors, is the total vector quantity impact adjustment coefficient, A constant for total vector quantity effect scaling.

[0092] In some embodiments disclosed herein, the method for determining the secondary matching parameters of all comparison groups includes:

[0093] Step S5033, determine the relative position relationship between the vector groups between the effective edge factor reference segments in the block, and the number of position matches between the relative position relationship between the corresponding edge factor reference segment feature data groups, and calculate the matching ratio between the number of position matches and the number of vector groups between the effective edge factor reference segments.

[0094] Step S5034, modifying the primary matching parameter based on the matching ratio to obtain the secondary matching parameter;

[0095] Among them, the expression of the secondary matching parameter is:

[0096] .

[0097] in, is the secondary matching parameter, is the number of position matches, is the number of vector groups, R is the matching ratio impact adjustment coefficient, Adjust the constant to match the ratio effect.

[0098] Step S504, based on the first-level matching parameters of each comparison group and the relative position relationship between the effective edge factor reference segment vector groups, determine the second-level matching parameters of all comparison groups. If the second-level matching parameters are greater than or equal to the preset value, the edge factor reference feature data groups in all comparison groups are identified as matching edge factor reference feature data groups.

[0099] Step S505, based on the determined matching edge factor reference feature data set, supplement the intervals between edge factor reference segments in the real-time infrared image, including using the vector group in the interval reference segment in the matching edge factor reference feature data set as the interval supplement basis.

[0100] In some embodiments disclosed in the present invention, there is also disclosed an infrared image optimization system based on data analysis, comprising:

[0101] The first module is used to set a number of scene type databases according to the scenes used by infrared images, and store historical infrared images and historical on-site images in the corresponding scene type databases in the form of historical image groups.

[0102] The second module is used to grayscale the historical scene images in the historical image group to obtain the historical scene grayscale map, and use edge recognition technology to identify the edges in the historical scene grayscale map and perform edge delineation. The result of edge delineation is projected onto the positioning template map to obtain the historical edge line positioning map, and the historical edge line positioning map is aligned and adapted with the historical infrared images in the historical image group.

[0103] The third module is used to perform edge factor analysis on historical infrared images, determine edge factor reference segments in historical infrared images, associate the edge factor reference segments with corresponding edge lines in the historical edge line positioning map, and determine the interval characteristics between edge factor reference segments based on the shape characteristics of the edge lines. The edge factor reference segments and the interval characteristics between them are recorded as edge factor reference segment feature data groups, and the edge factor reference segment feature data groups are constructed as edge factor reference segment feature data sets. Based on the relative position relationship between adjacent edge lines in the historical edge line positioning map, the relative position relationship between edge factor reference segment feature data groups is recorded, and the association relationship between edge factor reference segment feature data groups is established.

[0104] The fourth module is used to determine the application scenario of the real-time infrared image during the optimization process of the real-time infrared image, and call the corresponding edge factor reference segment feature data set based on the application scenario, and then perform edge factor analysis on the real-time infrared image, and based on the analysis results, improve and supplement the intervals between the edge factor reference segments.

[0105] The invention discloses an infrared image optimization method and system based on data analysis, and relates to the technical field of infrared image optimization. Firstly, a plurality of scene type databases are set to store historical infrared images and corresponding on-site images; by graying and edge recognition of the historical on-site images, a historical edge line positioning map is obtained, and the historical infrared images are aligned and adapted; edge factor analysis is performed on the historical infrared images, edge factor reference segments and their interval characteristics are determined, and an edge factor reference segment feature data set is constructed; based on the relative position relationship of edge lines in the historical edge line positioning map, an association relationship between edge factor reference segment feature data sets is established; when optimizing real-time infrared images, the corresponding edge factor reference segment feature data set is called to improve and supplement the intervals between edge factors. The invention realizes the supplement of edges in infrared images and improves the overall quality of images through the above technical scheme.

[0106] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An infrared image optimization method based on data analysis, characterized in that: include: Step S100, for the scene to which the infrared image is applied, several scene type databases are set, and the historical infrared images and historical on-site images are stored in the corresponding scene type database in the form of historical image groups; Step S200, graying the historical scene images in the historical image group to obtain a historical scene grayscale map, and using edge recognition technology to identify the edges in the historical scene grayscale map, and perform edge delineation, projecting the edge delineation result onto the positioning template map to obtain a historical edge line positioning map, and aligning and adapting the historical edge line positioning map and the historical infrared images in the historical image group; Step S300, performing edge factor analysis on the historical infrared image, determining the edge factor reference segment in the historical infrared image, associating the edge factor reference segment with the corresponding edge line in the historical edge line positioning map, and determining the interval characteristics between the edge factor reference segments based on the shape characteristics of the edge lines, recording the edge factor reference segments and the interval characteristics between them as an edge factor reference segment feature data group, and constructing the edge factor reference segment feature data group as an edge factor reference segment feature data set; Step S400, based on the relative positional relationship between adjacent edge lines in the historical edge line positioning map, the relative positional relationship between edge factor reference segment feature data groups is recorded, and an association relationship between edge factor reference segment feature data groups is established; Step S500, in the process of optimizing the real-time infrared image, determining the application scenario of the real-time infrared image, and calling the corresponding edge factor reference segment feature data set based on the application scenario, then performing edge factor analysis on the real-time infrared image, and based on the analysis result, improving and supplementing the intervals between the edge factor reference segments; The method for determining the interval characteristics between edge factor reference segments based on the shape characteristics of the edge line includes: An image coordinate system is constructed for the historical infrared image, and based on a preset interval, a number of edge detection points are set on each edge factor reference segment, the point coordinates of each edge detection point are determined, and the average point coordinates of all point coordinates corresponding to the edge factor reference segment are calculated; The average point coordinates of the edge factor reference segments mapped to the same edge line are associated, and the vector between each two adjacent average point coordinates is calculated, and the vectors are combined and recorded as the edge factor reference segment vector group; Determine the interval reference segment between the edge factor reference segments on the edge line, and based on the preset interval, set a number of interval detection points on the interval reference segment, and determine the point coordinates of each interval detection point, and calculate the vector between each adjacent interval detection point, and combine the vectors to obtain a vector group in the interval reference segment; Based on the position of the edge factor reference segment relative to the edge line, the vector group between the edge factor reference segments and the vector group in the interval reference segment are associated in sequence to obtain the edge factor reference segment feature data group.

2. The infrared image optimization method based on data analysis according to claim 1, characterized in that: Methods for identifying edges in historical scene grayscale images include: Applying edge detection operators to judge the gradient values ​​of pixel points in the historical scene grayscale image, and determining edge points in the historical scene grayscale image; Based on the position and gradient direction of each edge point, the edge points are connected to form the edge in the historical scene grayscale image.

3. The infrared image optimization method based on data analysis according to claim 1 is characterized in that: The methods for edge factor analysis of historical infrared images include: Grayscale the historical infrared image to obtain a historical infrared grayscale image, and use edge recognition technology to determine the edge factor segment in the historical infrared grayscale image and mark it in the historical infrared grayscale image; The historical infrared grayscale image is compared with the historical edge positioning image, and the edge lines in the historical edge positioning image are mapped to the historical infrared grayscale image. The edge factor segments that do not overlap with the edge lines are eliminated, and the remaining edge factor segments are identified as edge factor reference segments.

4. The infrared image optimization method based on data analysis according to claim 1 is characterized in that: The methods for edge factor analysis of real-time infrared images include: Grayscale processing is performed on the real-time infrared image to obtain a real-time infrared grayscale image, and edge recognition technology is used to determine the edge factor reference segment in the real-time infrared image; Dynamic block scanning is performed on the edge factor reference segment in the real-time infrared image, including gradually changing the position of the preset block in the real-time infrared image, and performing relative position analysis on the edge factor reference segment in the block, and searching the analysis results in the edge factor reference segment feature data set to determine a number of matching edge factor reference segment feature data groups.

5. The infrared image optimization method based on data analysis according to claim 4 is characterized in that: Methods for retrieving analysis results in edge factor reference segment feature datasets include: A number of edge detection points are set for the edge factor reference segment in the block, and the point coordinates of each edge detection point are determined, and the average point coordinates of the edge factor reference segment in the block are determined, and the vectors between the average point coordinates of each adjacent edge factor reference segment are calculated, and the vectors are combined to obtain a vector group between edge factor reference segments; Based on the vector group between edge factor reference segments in the block, the relative directions between edge factor reference segments in the block are determined, and the relative directions between edge factor reference segments are combined into a relative direction sequence; The relative direction sequence between edge factor reference segments in the block is used as a retrieval label and retrieved in the edge factor reference segment feature dataset.

6. The infrared image optimization method based on data analysis according to claim 4 is characterized in that: The methods for edge factor analysis of real-time infrared images include: The edge factor reference segment feature data groups retrieved and screened are respectively compared with the corresponding edge factor reference segment vector groups in the block to form several comparison groups, and the primary matching parameters of each comparison group are respectively determined. If the primary matching parameters are greater than or equal to the preset values, the edge factor reference segment vector groups in the comparison group are identified as valid edge factor reference segment vector groups, otherwise the edge factor reference segment feature data groups in the comparison group are replaced; Based on the primary matching parameters of each comparison group and the relative position relationship between the effective edge factor reference segment vector groups, the secondary matching parameters of all comparison groups are determined. If the secondary matching parameters are greater than or equal to the preset value, the edge factor reference feature data groups in all comparison groups are identified as matching edge factor reference feature data groups. Based on the determined matching edge factor reference feature data set, the intervals between edge factor reference segments in the real-time infrared image are supplemented, including using the vector group in the interval reference segment in the matching edge factor reference feature data set as the interval supplement basis.

7. The infrared image optimization method based on data analysis according to claim 6 is characterized in that: Methods for calculating the primary matching parameters of the alignment group include: Perform several offset correspondences on the edge factor reference segment vector group in the block relative to the edge factor reference segment feature data group, and after each offset correspondence, determine the number of corresponding identical vectors in the edge factor reference segment vector group in the block, and calculate the ratio of the number of identical vectors to the total number of vectors; Based on the equivalent vector ratio of the vector group between the edge factor reference segments and the total number of vectors, the first-level matching parameters of the comparison group are determined; Among them, the expression for calculating the first-level matching parameter of the comparison group is: Among them, W1 is the first-level matching parameter of the comparison group, K1 is the matching parameter conversion coefficient, x1 is the number of equivalent vectors, x2 is the total number of vectors, K2 is the total number of vectors impact adjustment coefficient, and b is the total number of vectors impact adjustment constant.

8. The infrared image optimization method based on data analysis according to claim 6 is characterized in that: Methods for determining secondary fit parameters for all alignment groups include: Determine the relative position relationship between the effective edge factor reference segment vector groups in the block, and the position matching number of the relative position relationship between the corresponding edge factor reference segment feature data groups, and calculate the matching ratio of the position matching number to the number of vector groups of the effective edge factor reference segment vector groups; The first-level matching parameters are corrected based on the matching ratio to obtain the second-level matching parameters; Among them, the expression of the secondary matching parameter is: Among them, W2 is the secondary matching parameter, u1 is the position matching quantity, u2 is the number of vector groups, R is the matching ratio influence adjustment coefficient, and c is the matching ratio influence adjustment constant.

9. An infrared image optimization system based on data analysis, characterized in that: include: The first module is used to set a number of scene type databases according to the scenes to which the infrared images are applied, and store the historical infrared images and the historical on-site images in the corresponding scene type databases in the form of historical image groups; The second module is used to grayscale the historical scene images in the historical image group to obtain the historical scene grayscale map, and use the edge recognition technology to identify the edges in the historical scene grayscale map, and perform edge delineation, project the edge delineation result onto the positioning template map, obtain the historical edge line positioning map, and align and adapt the historical edge line positioning map with the historical infrared images in the historical image group; The third module is used to perform edge factor analysis on the historical infrared image, determine the edge factor reference segment in the historical infrared image, associate the edge factor reference segment with the corresponding edge line in the historical edge line positioning map, and determine the interval characteristics between the edge factor reference segments based on the shape characteristics of the edge lines, record the edge factor reference segments and the interval characteristics between each other as an edge factor reference segment feature data group, and construct the edge factor reference segment feature data group into an edge factor reference segment feature data set, based on the relative position relationship between adjacent edge lines in the historical edge line positioning map, record the relative position relationship between the edge factor reference segment feature data groups, and establish the association relationship between the edge factor reference segment feature data groups; An image coordinate system is constructed for the historical infrared image, and based on a preset interval, a number of edge detection points are set on each edge factor reference segment, the point coordinates of each edge detection point are determined, and the average point coordinates of all point coordinates corresponding to the edge factor reference segment are calculated; The average point coordinates of the edge factor reference segments mapped to the same edge line are associated, and the vector between each two adjacent average point coordinates is calculated, and the vectors are combined and recorded as the edge factor reference segment vector group; Determine the interval reference segment between the edge factor reference segments on the edge line, and based on the preset interval, set a number of interval detection points on the interval reference segment, and determine the point coordinates of each interval detection point, and calculate the vector between each adjacent interval detection point, and combine the vectors to obtain a vector group in the interval reference segment; Based on the position of the edge factor reference segment relative to the edge line, the vector group between the edge factor reference segments and the vector group in the interval reference segment are associated in sequence to obtain the edge factor reference segment feature data group; The fourth module is used to determine the application scenario of the real-time infrared image during the optimization process of the real-time infrared image, and call the corresponding edge factor reference segment feature data set based on the application scenario, and then perform edge factor analysis on the real-time infrared image, and based on the analysis results, improve and supplement the intervals between the edge factor reference segments.

Images