Infrared camera simulation data source generation method and system

Abstract

The invention provides an infrared camera simulation data source generation method and system, and the method comprises the steps: collecting a ground object background and a target image, building a background and target image alternative library and a gray variance lookup table, and facilitating the rapid adjustment of the gray range of the background and the target during subsequent processing; after the scene is established, selecting a target image and a background image, and superposing the target image and the background image by using a linear spectrum hybrid model; and after noise and blind flash elements are added by applying a simple model, an infrared camera analog data code stream is generated according to a camera data format. The invention provides an infrared camera simulation data source rapid generation method which can rapidly generate a large number of infrared camera simulation data sources based on an established background and target image alternative library. A data source generated by the system can be set as required, the requirements of various application scenes can be met, and a simulation data source generated by the system can be applied to camera index demonstration; a simulation data source generated by the system can meet the requirement of a mass data set of an information processing algorithm.

Description

Technical Field

[0001] The present invention relates to the field of infrared camera technology, and in particular to a method and system for generating a simulated data source for an infrared camera, and more particularly to a method and system for generating a simulated data source for an infrared camera that can generate data quickly and in large quantities and can be set as needed to meet the needs of various application scenarios. Background Art

[0002] The rapid generation method of infrared camera simulation data source has a wide range of application value and is a hot research topic in scientific research and military fields in various countries. Compared with the actual test method, the application of imaging simulation technology is not limited to specific experimental environments and equipment, and rich and comprehensive infrared simulation images can be obtained at a very low cost. Many scholars have conducted a lot of research work and formed a series of models for the design, training and evaluation of actual infrared imaging systems. Among them, the representative infrared imaging systems are Vega Prime IR / NVG Sensor Module, CAMEO-SIM, SE-WorkBench-IR, Mantis ViXsen, ShipIR / NTCS. Most of the research on these systems generates infrared simulation images by physically modeling the scene and calculating the surface infrared radiation and reflection. There are very few methods for rapid simulation by establishing a database. Therefore, it is necessary to propose a method for rapid generation of infrared camera simulation data source to improve the speed of infrared camera simulation imaging and supplement the measured data. Summary of the invention

[0003] In view of the defects in the prior art, an object of the present invention is to provide a method and system for generating a simulated data source for an infrared camera.

[0004] A method for quickly generating a simulated data source for an infrared camera provided by the present invention comprises:

[0005] Step S1: establishing a candidate library of ground object background images, wherein the candidate library includes grayscale images of each candidate image and radiation intensity images after radiation calibration; classifying the ground objects to determine the ground object category represented by each pixel in the candidate image;

[0006] Step S2: Establish a lookup table of the mean and standard deviation of the radiation intensity of various types of ground objects under various detection band conditions;

[0007] Step S3: After performing target detection, positioning and extraction on the measured data, a target image candidate library is established; the target image candidate library includes grayscale image slices, radiation intensity image slices and target motion trajectory information of each typical target;

[0008] Step S4: Establish a lookup table of the mean and standard deviation of the radiation intensity of each typical target in each detection band;

[0009] Step S5: setting a scene, selecting background and target data in the background and target image candidate library, and adjusting the grayscale values ​​of the target and background images according to the set band information;

[0010] Step S6: According to the target motion trajectory in the set scene, the target and the background are superimposed to generate a detection image;

[0011] Step S7: simplifying the image noise and blind flash element model according to the working mode and characteristics of the infrared camera, and adding noise, blind pixels and flash pixels to the detection image;

[0012] Step S8: Generate a data code stream according to the camera data format.

[0013] Preferably, the candidate database of ground object background images in step S1 contains matching ground object classification data, and the ground object categories are divided into green vegetation, desert, ice and snow, water, bare land, city, high-level cloud, middle-level cloud, low-level cloud, and vertical cloud; the ground object background image comes from a public satellite remote sensing data source.

[0014] Preferably, the target images in the target image candidate library in step S3 include various typical categories of aircraft and ship targets, which are obtained by performing target detection, positioning and extraction on measured remote sensing data.

[0015] Preferably, the background gray value adjustment method in step S5 includes: converting the background radiation intensity in the lookup table into an image gray value according to the simulation band and camera parameter conditions, and then adjusting the gray of the ground object background image according to the gray mean and standard deviation of each type of ground object, and the adjustment method is:

[0016]

[0017] Where, L B 'Adjusted background gray value, L B is the background grayscale value before adjustment, is the background grayscale mean before adjustment, σ L To adjust the standard deviation of the foreground and background grayscale, σ I is the grayscale standard deviation of the objects in the lookup table, is the mean grayscale value of the objects in the lookup table.

[0018] Preferably, the target grayscale value adjustment method in step S5 includes: converting the corresponding target radiation intensity in the lookup table into an image grayscale value according to the simulation band and camera parameter conditions, and adjusting the grayscale of the ground object background image according to the grayscale mean and standard deviation of the target.

[0019] Preferably, the target and background image superposition method in step S6 includes: using a linear mixed model to process the boundary area between the target and the background, and superimposing the target and the background. The linear spectral mixed model represents the radiation measurement value at the boundary between the target and the background as a linear combination of the target and the background with the pixel ratio occupied by each pixel as the weight coefficient:

[0020] L(λ)=L tar (λ)S tar (λ)+L bg (λ)S bg (λ)

[0021] Where L(λ) represents the gray value of the superimposed image, L tar (λ) and L bg (λ) are the gray values ​​corresponding to the target and background respectively; S tar (λ) and S bg (λ) are the percentages of the target and background occupied by the current pixel area.

[0022] Preferably, the simplified noise and blind flash element model in step S7 includes: the noise is divided into stripe noise and global Gaussian noise, which is simplified to be randomly distributed with a certain mean and variance; the grayscale value fluctuation range of the blind element takes the lowest 5% and the highest 5% of the camera grayscale dynamic range, and the grayscale value of the flash element fluctuates randomly, and the fluctuation range is consistent with the camera grayscale dynamic range.

[0023] A system for quickly generating a simulated data source for an infrared camera provided by the present invention comprises:

[0024] Module M1: Establishing a candidate library of ground object background images, the candidate library of ground object background images includes grayscale images of each candidate image and radiation intensity images after radiation calibration; classifying the ground objects and determining the ground object category represented by each pixel in the candidate image;

[0025] Module M2: Establish a lookup table of the mean and standard deviation of the radiation intensity of various types of objects under various detection band conditions;

[0026] Module M3: After target detection, positioning and extraction of measured data, a target image candidate library is established; the target image candidate library includes grayscale image slices, radiation intensity image slices and target motion trajectory information of each typical target;

[0027] Module M4: Establish a lookup table of the mean and standard deviation of the radiation intensity of each typical target in each detection band;

[0028] Module M5: setting the scene, selecting the background and target data in the background and target image candidate library, and adjusting the grayscale values ​​of the target and background images according to the set band information;

[0029] Module M6: According to the target motion trajectory in the set scene, the target and the background are superimposed to generate a detection image;

[0030] Module M7: Simplify the image noise and blind flash element model according to the working mode and characteristics of the infrared camera, and add noise, blind pixels and flash pixels to the detected image;

[0031] Module M8: Generates data code stream according to the camera data format.

[0032] Preferably, the optional database of ground object background images in the module M1 contains matching ground object classification data, and the ground object categories are divided into green vegetation, desert, ice and snow, water, bare land, city, high-level clouds, middle-level clouds, low-level clouds, and vertical clouds; the ground object background images come from public satellite remote sensing data sources.

[0033] Preferably, the target images in the target image candidate library in the module M3 include various typical categories of aircraft and ship targets, which are obtained by performing target detection, positioning and extraction on measured remote sensing data.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] 1. The present invention proposes a method for quickly generating an infrared camera simulation data source, which can quickly generate a large number of infrared camera simulation data sources based on an established background and target image candidate library.

[0036] 2. The data source generated by the present invention can be set as needed to meet the needs of various application scenarios.

[0037] 3. The simulated data source generated by the present invention can be applied to camera indicator demonstration.

[0038] 4. The simulated data source generated by the present invention can support the needs of massive data sets for information processing algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0040] Figure 1 A flow chart of a method for quickly generating a simulated data source for a mid-infrared camera of the present invention;

[0041] Figure 2 Schematic diagram of the linear spectral mixing model in the present invention;

[0042] Figure 3a-3b It is a comparison diagram of the effects of the embodiments of the present invention; wherein, Figure 3a For example, a low clutter background Figure 3b This is an example of a high clutter background. DETAILED DESCRIPTION

[0043] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0044] The present invention discloses a method for quickly generating a simulated data source of an infrared camera, referring to Figure 1 and Figure 2 As shown in the figure, based on the method of establishing a background and target image candidate library and a grayscale variance lookup table, the grayscale range of the background and the target can be quickly adjusted to generate infrared image data of the band of interest. It can provide a large amount of data support for target detection and recognition in infrared images, support the demonstration of camera design indicators, and effectively improve efficiency. Specifically, it includes the following steps:

[0045] Step S1: Establish a candidate library of ground object background images. Based on various open-source satellite optical and infrared remote sensing data, a candidate library of ground object background images under different resolution conditions is established. The library contains grayscale images of each candidate image and radiation intensity images after radiation calibration. Classify the ground objects and determine the ground object category represented by each pixel in the candidate image.

[0046] In a preferred embodiment, the feature background images in the feature background image candidate library in step S1 store matching feature classification data, and the feature categories are divided into ten features: green vegetation, desert, ice and snow, water, bare land, city, high-level cloud, middle-level cloud, low-level cloud, and vertical cloud. The feature background images are sourced from public satellite remote sensing data sources such as Landsat, Worldview, and Fengyun satellites.

[0047] Step S2: Establish a lookup table of the mean and standard deviation of the radiation intensity of various types of ground objects under various detection band conditions;

[0048] Step S3: Establish a target image candidate library. After performing target detection, positioning, and extraction on the measured data, establish a target data candidate library. The library contains grayscale image slices, radiation intensity image slices, and target motion trajectory information of each typical target;

[0049] Specifically, the target images in the target image candidate library in step S3 include targets of various typical categories, such as airplanes and ships, which are all obtained by performing target detection, positioning, and extraction on measured remote sensing data.

[0050] Step S4: Establish a lookup table of the mean and standard deviation of the radiation intensity of each typical target in each detection band;

[0051] Step S5: setting a scene, selecting background and target data in the background and target image candidate library, and adjusting the grayscale values ​​of the target and background images according to the set band information;

[0052] Specifically, the background gray value adjustment method in step S5 includes: according to the simulation band and camera parameter conditions, the background radiation intensity in the lookup table is converted into an image gray value, and then the gray value of the ground object background image is adjusted according to the gray value mean and standard deviation of each type of ground object. The adjustment method is:

[0053]

[0054] Where, L B 'Adjusted background gray value, L B is the background grayscale value before adjustment, is the background grayscale mean before adjustment, σ L To adjust the standard deviation of the foreground and background grayscale, σ I is the grayscale standard deviation of the objects in the lookup table, is the mean grayscale value of the objects in the lookup table.

[0055] The target grayscale value adjustment method in step S5 includes: according to the simulation band and camera parameter conditions, the corresponding target radiation intensity in the lookup table is converted into an image grayscale value, and then the grayscale of the ground object background image is adjusted according to the grayscale mean and standard deviation of the target. The formula of the target grayscale adjustment method is the same as that of the background.

[0056] Step S6: According to the target motion trajectory in the set scene, the target and the background are superimposed to generate a detection image.

[0057] Specifically, the target and background image superposition method in step S6 includes: using a linear mixed model to process the boundary area between the target and the background, and superimposing the target and the background. The linear spectral mixed model represents the radiation measurement value at the boundary between the target and the background as a linear combination of the target and the background with the pixel ratio occupied by them as the weight coefficient:

[0058] L(λ)=L tar (λ)S tar (λ)+L bg (λ)S bg (λ)

[0059] Where L(λ) represents the gray value of the superimposed image, L tar (λ) and L bg (λ) are the gray values ​​corresponding to the target and background respectively; S tar (λ) and S bg (λ) are the percentages of the target and background occupied by the current pixel area.

[0060] Step S7: simplifying the image noise and blind flash element model according to the working mode and characteristics of the infrared camera, and adding noise and blind flash elements to the detection image;

[0061] Specifically, the simplified noise and blind flash element model in step S7 includes: the noise is divided into stripe noise and global Gaussian noise, which is simplified to be randomly distributed with a certain mean and variance. The gray value fluctuation range of the blind element takes the lowest 5% and the highest 5% of the camera gray dynamic range, and the gray value of the flash element fluctuates randomly, and the fluctuation range is consistent with the camera gray dynamic range.

[0062] Step S8: Generate a data code stream according to the camera data format.

[0063] The present invention will be further described below in conjunction with the specific drawings.

[0064] Reference Figure 3a-3b As shown, this embodiment uses the public remote sensing data of Fengyun-4 to establish a background candidate library, adjusts its grayscale mean and variance, adds stripe noise, random noise and blind flash elements, and generates infrared camera simulation data. Low clutter background and high clutter background areas are selected for generation.

[0065] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0066] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for quickly generating a simulated data source for an infrared camera, characterized in that: include: Step S1: establishing a candidate library of ground object background images, wherein the candidate library of ground object background images includes grayscale images of each candidate image and radiation intensity images after radiation calibration; Classify the ground objects and determine the ground object category represented by each pixel in the candidate image; Step S2: Establish a lookup table of the mean and standard deviation of the radiation intensity of various types of ground objects under various detection band conditions; Step S3: After performing target detection, positioning and extraction on the measured data, a target image candidate library is established; the target image candidate library includes grayscale image slices, radiation intensity image slices and target motion trajectory information of each typical target; Step S4: Establish a lookup table of the mean and standard deviation of the radiation intensity of each typical target in each detection band; Step S5: setting a scene, selecting background and target data in the background and target image candidate library, and adjusting the grayscale values ​​of the target and background images according to the set band information; Step S6: According to the target motion trajectory in the set scene, the target and the background are superimposed to generate a detection image; Step S7: simplifying the image noise and blind flash element model according to the working mode and characteristics of the infrared camera, and adding noise, blind pixels and flash pixels to the detection image; Step S8: Generate a data code stream according to the camera data format.

2. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: The candidate database of ground object background images in step S1 contains matching ground object classification data, and the ground object categories are divided into green vegetation, desert, ice and snow, water, bare land, city, high-level cloud, middle-level cloud, low-level cloud, and vertical cloud; The ground feature background images come from public satellite remote sensing data sources.

3. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: The target images in the target image candidate library in step S3 include various typical categories of aircraft and ship targets, which are obtained by performing target detection, positioning and extraction on the measured remote sensing data.

4. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: The background gray value adjustment method in step S5 includes: converting the background radiation intensity in the lookup table into an image gray value according to the simulation band and camera parameter conditions, and then adjusting the gray of the ground object background image according to the gray mean and standard deviation of each type of ground object, and the adjustment method is: Where, L B 'Adjusted background gray value, L B is the background grayscale value before adjustment, is the background grayscale mean before adjustment, σ L To adjust the standard deviation of the foreground and background grayscale, σ I is the grayscale standard deviation of the objects in the lookup table, is the mean grayscale value of the objects in the lookup table.

5. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: The target grayscale value adjustment method in step S5 includes: converting the corresponding target radiation intensity in the lookup table into an image grayscale value according to the simulation band and camera parameter conditions, and adjusting the grayscale of the ground object background image according to the target grayscale mean and standard deviation.

6. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: The target and background image superposition method in step S6 includes: using a linear mixed model to process the boundary area between the target and the background, and superimposing the target and the background. The linear spectral mixed model represents the radiation measurement value at the boundary between the target and the background as a linear combination of the target and the background with the pixel ratio occupied by the target and the background as the weight coefficient: L(λ)=L tar (λ)S tar (λ)+L bg (λ)S bg (l) Where L(λ) represents the gray value of the superimposed image, L tar (λ) and L bg (λ) are the gray values ​​corresponding to the target and background respectively; S tar (λ) and S bg (λ) are the percentages of the target and background occupied by the current pixel area.

7. The method for quickly generating an infrared camera simulation data source according to claim 1, characterized in that: include: The simplified noise and blind flash element model in step S7 includes: the noise is divided into stripe noise and global Gaussian noise, which is simplified to be randomly distributed with a certain mean and variance; the grayscale value fluctuation range of the blind element takes the lowest 5% and the highest 5% of the camera grayscale dynamic range, and the grayscale value of the flash element fluctuates randomly, and the fluctuation range is consistent with the camera grayscale dynamic range.

8. A rapid generation system for infrared camera simulation data source, characterized in that: include: Module M1: Establishing a candidate library of ground object background images, the candidate library of ground object background images includes grayscale images of each candidate image and radiation intensity images after radiation calibration; classifying the ground objects and determining the ground object category represented by each pixel in the candidate image; Module M2: Establish a lookup table of the mean and standard deviation of the radiation intensity of various types of objects under various detection band conditions; Module M3: After target detection, positioning and extraction of measured data, a target image candidate library is established; the target image candidate library includes grayscale image slices, radiation intensity image slices and target motion trajectory information of each typical target; Module M4: Establish a lookup table of the mean and standard deviation of the radiation intensity of each typical target in each detection band; Module M5: setting the scene, selecting the background and target data in the background and target image candidate library, and adjusting the grayscale values ​​of the target and background images according to the set band information; Module M6: According to the target motion trajectory in the set scene, the target and the background are superimposed to generate a detection image; Module M7: Simplify the image noise and blind flash element model according to the working mode and characteristics of the infrared camera, and add noise, blind pixels and flash pixels to the detected image; Module M8: Generates data code stream according to the camera data format.

9. The infrared camera simulation data source rapid generation system according to claim 8, characterized in that: The object background image candidate library in the module M1 contains matching object classification data, and the object categories are divided into green vegetation, desert, ice and snow, water, bare land, city, high-level cloud, middle-level cloud, low-level cloud, and vertical cloud; The ground feature background images come from public satellite remote sensing data sources.

10. The infrared camera simulation data source rapid generation system according to claim 8, characterized in that: The target images in the target image candidate library in the module M3 include various typical categories of aircraft and ship targets, which are obtained by performing target detection, positioning and extraction on the measured remote sensing data.

Images