A method for fast reconstruction of human target infrared three-dimensional model

Abstract

The application discloses a kind of human target infrared three-dimensional model fast reconstruction method, belong to model reconstruction algorithm field.The method is first based on infrared physics to the heat transfer process on human surface is physically modeled, and utilizes the principle of graphics to generate the human infrared three-dimensional feature image of target human body under different conditions.Functions are 1) based on measured data, establish the infrared radiation database of simulation target environment.2) the heat transfer process on human surface is physically modeled using finite difference method, realizes three-dimensional human infrared model drawing.3) using computer iterative calculation generates the infrared feature image of human target in different time, place, climate condition.4) based on the principle of Planck spectrum energy distribution, long-wave infrared simulation image of human body is converted into mid-wave simulation image.Through hash algorithm, the similarity and reliability of human infrared simulation imaging result are evaluated, and the simulation similarity is higher than 70%.

Description

Technical Field

[0001] This invention relates to the field of model reconstruction algorithms, and in particular to a method for rapid reconstruction of a human target infrared 3D model. Background Technology

[0002] Any object with a temperature above absolute zero emits infrared radiation. In low-light conditions, such as at night, or when targets are camouflaged or concealed, infrared imagery obtained through thermal imaging can clearly reveal target information that is difficult or impossible to detect under visible light. Therefore, infrared identification and detection technology has been widely applied and plays a crucial role in fields such as mechanical and electronic fault diagnosis, and national defense.

[0003] Targets often possess diverse and non-cooperative characteristics. Given the complexity of their background conditions, existing infrared characteristic experiments suffer from long research cycles, high costs, and poor accuracy. Furthermore, it is difficult to capture the infrared characteristics of targets under special conditions such as dense fog and rainfall. With the development of computer graphics, computer vision, and other disciplines, and the deepening of basic research, infrared simulation and target modeling using computer technology have become widely used in various fields due to their advantages of speed, repeatability, and intuitive visualization. Compared to infrared characteristic experiments conducted in real-world environments, computer simulation not only saves experimental costs but also allows for the rapid, repeated, and consistent generation of infrared feature images of targets and backgrounds under various harsh natural conditions that are normally difficult to obtain, making it of significant application value.

[0004] Infrared simulation refers to the use of a physical heat transfer model and a computer to simulate the temperature field distribution on the surface of an object, thereby calculating the zero-line-of-sight infrared radiation value of the object's surface. This allows for the simulation of infrared feature images captured by actual infrared imaging equipment in real-world environments. Furthermore, it is necessary to consider the effects of various external factors on the object's infrared radiation value, such as various weather conditions and equipment errors, to further calculate the infrared radiation value reaching the receiving device after attenuation. This value is then converted into the grayscale values ​​of each pixel in the simulated image, resulting in an infrared feature grayscale image of the target and background.

[0005] The mainstream research method is to render and generate infrared feature images in real time based on open-source graphics rendering engines such as OpenGL. Images generated using this method have good controllability, but the development cycle is long, the difficulty is high, and the versatility is relatively poor. Summary of the Invention

[0006] To overcome the aforementioned technical problems, this invention comprehensively applies the basic principles or algorithms of multiple fields such as infrared physics, heat transfer, computer graphics, and computer vision, and designs a method for rapid reconstruction of a human target infrared three-dimensional model. It mainly includes iteratively solving the human body surface temperature field based on the explicit finite difference method, and an infrared band conversion model based on the Planck spectrum energy distribution principle.

[0007] The technical solution adopted in this invention is as follows:

[0008] A method for rapid reconstruction of a human target infrared 3D model includes the following steps:

[0009] Step 1: Based on measured data, establish a target environment database, including spontaneous radiation, background radiation, and solar radiation data of the target environment;

[0010] Step 2: Solve the temperature field distribution on the human body surface based on the explicit finite difference method to obtain the relationship between the temperature of each node on the human body surface and time.

[0011] Step 3: Based on the target environment database data and material properties of the human body surface obtained in Step 1, obtain the total radiation change of the human body surface and the temperature change of each node of the human body surface within any time period; Based on the relationship between the temperature of each node of the human body surface and time obtained in Step 2 and the temperature field distribution of the human body surface at the beginning of the time period, iteratively solve the temperature field distribution of the human body surface at the end of the time period to obtain the human body long-wave infrared simulation image.

[0012] Step 4: Based on the range of environmental parameters in the target environment database, repeat step 3 to generate simulated long-wave infrared images of the human body under different time, location, and climate conditions.

[0013] Step 5: Based on the Planck spectrum energy distribution principle, the long-wave infrared simulation image of the human body is converted into a mid-wave infrared simulation image of the human body, which serves as the final 3D model reconstruction result.

[0014] Furthermore, the relationship between the temperature of each node on the human body surface and time, as described in step 2, is as follows:

[0015]

[0016] Where c is the temperature of a certain node on the human body at a certain moment; Δx is the grid distance, that is, the distance between two nodes; x k Let β be the location of the k-th node, β be the thermal diffusivity of a node on the human body, and Δt be the time step. s This refers to the current time.

[0017] Furthermore, in step 3, the relationship between the total radiation change and the temperature change at various nodes on the human body surface is as follows:

[0018] ΔQ=CmΔT

[0019] Where ΔQ is the total radiation change, ΔT is the temperature change at each node on the human body surface, m is the mass of the human body surface, and C is the specific heat capacity of the human body surface.

[0020] Furthermore, step 5 includes:

[0021] 5.1) Using Planck's principle of spectral energy distribution, calculate the maximum brightness L in the original long-wave infrared band. max and minimum brightness L min :

[0022]

[0023]

[0024] Among them, T max and T min These represent the upper and lower bounds of temperature in the infrared image, respectively; ε represents emissivity; Δλ is the bandwidth of the long-wave infrared band; λ c C1 and C2 are the center wavelengths of the long-wave infrared band; C1 and C2 are constants, C1 = 1.191 × 10⁻⁶. 4 Unit: Wμm 4 / cm 2 sr, C2 = 1.428 × 10 4 Unit: μmK;

[0025] 5.2) Calculate the brightness L0 of any point in the original band based on its gray level G0:

[0026]

[0027] 5.3) Calculate the temperature T at this point based on the brightness:

[0028]

[0029] 5.4) Calculate the brightness L1 of this point in the target's medium-wave band:

[0030]

[0031] Where, Δλ ′ λ represents the mid-infrared band width of the target. c ′ The center wavelength of the target in the mid-infrared band;

[0032] 5.5) Convert the brightness distribution to a grayscale distribution:

[0033]

[0034] Wherein, G1 is the converted grayscale distribution, which is the mid-wave infrared simulation image of the human body.

[0035] The beneficial effects of this invention are:

[0036] This invention, based on the fundamental theories of infrared physics and heat transfer, comprehensively considers the spontaneous emission, sky radiation, solar radiation, and background radiation of the target, and physically models the heat transfer process of the human body. It enables real-time rendering and simulation of the human infrared model, and can generate infrared feature images of human targets under different times, locations, and climatic conditions. Based on the fundamental theories of infrared physics and heat transfer, this invention establishes an infrared band imaging conversion method, which can easily convert the simulated long-wave infrared model image into a mid-wave infrared model image, thereby further expanding the human infrared feature image sample library.

[0037] Infrared imaging simulation technology was initially developed to meet the need to further improve the performance of infrared imaging seekers. This invention generates infrared feature images of the target and background based on the physical model of infrared radiation and heat transfer of the target and background. It provides seeker design with infrared images that simultaneously include the target and background, which plays a very important role in improving various performance parameters of the seeker and enhancing its ability to detect, identify and track targets under various complex conditions.

[0038] Furthermore, the infrared 3D model rapid reconstruction method proposed in this invention is fast, simple and easy to use, and is applicable to various locations, times, atmospheric conditions and target persons, and has strong versatility. Attached Figure Description

[0039] Figure 1 This is a schematic diagram illustrating the process of solving human body temperature distribution data;

[0040] Figure 2 This is the original three-dimensional human body model used in this invention;

[0041] Figure 3 This is a demonstration of long-wave infrared human body simulation effects;

[0042] Figure 4 This is a comparison between the mid-wave and long-wave simulated images after band conversion. The left image shows the long-wave infrared simulation result, and the right image shows the mid-wave infrared simulation result. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments. The accompanying drawings are merely illustrative diagrams of the present invention. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, or in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0044] In this embodiment, based on the principles of infrared physics and heat transfer, a method for rapid reconstruction of a human target infrared 3D model is proposed, which mainly includes the following steps:

[0045] 1) Based on measured data, establish a database of spontaneous radiation, background radiation, and solar radiation of the target environment, and input it into the simulation system as simulation parameters.

[0046] In this step, data such as temperature, sky background radiation, and solar radiation at various times over several days at the simulated target location can be obtained through actual measurement. For example, timers, wind speed sensors, temperature and humidity sensors, as well as sky radiometers and solar photometers can be installed at the target location to measure each parameter at specific intervals. The parameters are then stored in a CSV file that is easy for computer programs to read over several days as environmental parameters for subsequent calculations. Table 1 shows some of the infrared simulation measurement data collected in this embodiment.

[0047] Table 1 shows some infrared simulation and measurement data.

[0048]

[0049] 2) Solve the temperature field distribution on the human body surface based on the explicit finite difference method.

[0050] In this step, consider the heat conduction equation in one dimension:

[0051]

[0052] Where c is the surface temperature of a certain node on the human body at a certain time, t is the time, x is the location of the node, and β represents the thermal diffusivity of the location of a certain node on the human body.

[0053] Its first derivative forward difference form is:

[0054]

[0055] Where f and f ′ Let x represent a function of arbitrary form and its derivative with respect to time. k This represents the k-th node;

[0056] Assuming the spatial step size (grid size) is fixed, let the grid scale be Δx = x. k+2 -x k+1 =x k+1 -x k When Δx is sufficiently small, then:

[0057]

[0058] Further, we can conclude that:

[0059]

[0060] Among them, t s Representing the current time point, Δt = t s+1 -t s Indicates the time step;

[0061] For any time point t s , in x k At the node, the following formula exists:

[0062]

[0063] Further obtain t s+1 Temperature of various nodes on the human body at different times and t s The relationship between the surface temperature of various nodes in the human body at any given time:

[0064]

[0065] From the above equation, we can see that solving for t... s+1 When calculating the temperature values ​​at various points on the human body surface at any given time, it is necessary to first solve for t. s The temperature values ​​of each node on the human body surface at time t=0 are obtained. Thus, by iteratively solving the values ​​of each node inside the body at time t=0, c(0,x), and the values ​​of the nodes on the boundary, c(t,0) and c(t,1), the temperature values ​​of each node on the human body at any time can be obtained.

[0066] 3) Using the radiation data (spontaneous radiation of the target environment, background radiation of the sky, and solar radiation) collected in (1), combined with the material properties of each material on the human body surface, the total radiation change ΔQ of the human body surface over any time period can be calculated. Then, using the relationship ΔQ=CmΔT (where C is the specific heat capacity of the human body surface and m is the mass of the surface layer, the value of which is determined by the manually divided surface thickness), the temperature change ΔT at each node of the surface layer can be calculated. The t derived in (2) s+1 Temperature of various nodes on the human body at different times and t s The relationship between body temperature at various points in time can be determined by iterative calculations using a computer, thus yielding the body temperature distribution data at the end of that time period. The program implementation process is as follows: Figure 1As shown:

[0067] a. Input weather data and material parameters;

[0068] b. Initialize the number of surface layers, internal temperature, time and space steps, and total iteration time;

[0069] c. Determine if the current time has exceeded the total iteration time;

[0070] If not, calculate the energy levels of solar irradiance, background radiation, and reflected radiation based on physical properties and weather data, and then calculate the surface temperature change; iterate using the finite difference method to calculate the temperature of each layer, and then return to step c.

[0071] If so, iterate once more, and the result calculated at this point is the final result.

[0072] In this step, the target environment database, under the conditions to be solved, is the database parameter established in step 1) based on the measured data. It comprehensively considers the influence and specific values ​​of factors such as solar radiation, sky background radiation, reflected radiation and human body self-radiation on the energy changes of each node on the human body surface. Using the deterministic relationship between energy change and temperature change, the surface temperature value of each node is iteratively calculated. Figure 2 This is the original three-dimensional human body model used in this invention.

[0073] 4) Based on the target environment database, change the range of input parameters to generate infrared feature images of human targets under different time, location and climate conditions, i.e. human long-wave infrared simulation images as the result of three-dimensional model reconstruction.

[0074] In this step, by extracting portions of the environmental database representing specified locations, time periods, or atmospheric conditions, and using these portions as input parameters for simulation calculations, infrared feature images of human targets under various specified conditions can be obtained; in this embodiment, Figure 3 This is a demonstration of the simulation effect of long-wave infrared human body.

[0075] 5) Based on the Planck spectrum energy distribution principle, the long-wave infrared simulation image of the human body is converted into a medium-wave simulation image.

[0076] When electromagnetic waves pass through the air, they are affected by air reflection, absorption, and scattering, causing their energy to attenuate. The amount of attenuation varies depending on the wavelength of the electromagnetic wave, and those bands with higher transmittance are called atmospheric windows. Infrared light can be divided into three bands according to different wavelengths: 8–12 μm is the long-wave band, 3–5 μm is the medium-wave band, and 1.9–2.9 μm is the short-wave band.

[0077] The output voltage V of the infrared detector det It can be represented as:

[0078]

[0079] Where, τ amb (λ) represents the atmospheric spectral transmittance at an infrared wavelength of λ, τ opt (λ) represents the optical spectral transmittance of the detector at an infrared wavelength of λ, and S(λ) represents the spectral responsivity of the detector at an infrared wavelength of λ. λ1 and λ2 are the upper and lower limits of the band, respectively. p L(T) represents the area of ​​the target captured by the infrared detector, Ω represents the solid angle of the detector at the target, L(T) represents the radiance, and T represents the temperature.

[0080] Assume τ amb (λ), τ opt If S(λ) and S(λ) are a steady system, then V det ∝L(T). For a two-dimensional infrared detector, the grayscale of the infrared image is also related to the output voltage V of the infrared detector. det Therefore, the radiance L(T) is also proportional to the gray level.

[0081] The radiance of the target object in the given wavelength band λ1~λ2 is expressed as:

[0082]

[0083] The following approximation can be made to the expression for radiance:

[0084]

[0085] We can obtain:

[0086]

[0087] Where Δλ=λ2-λ1 is the bandwidth, Where C1 and C2 are constants, and C1 = 1.191 × 10⁻⁶, the center wavelength is given. 4 [Wμm 4 / cm 2 [sr], C2 = 1.428 × 10 4 [μmK], ε(λ) represents the spectral emissivity of the object.

[0088] Based on the above principles, if the temperature and emissivity of the object corresponding to at least two pixel gray levels in an image are known, an infrared image in a certain band can be converted to any band. However, bands outside the atmospheric window are greatly affected by air reflection, absorption, and scattering, and are rarely used in infrared imaging; and compared to the other two bands, the transmittance of infrared light in the short-wave infrared band by water molecules and carbon dioxide molecules in the air is irregular (the transmittance is approximately 1 in the mid-wave and long-wave bands), resulting in a larger error in the conversion result. Therefore, this invention only considers the conversion between the mid-wave and long-wave bands. The conversion steps are as follows:

[0089] 5.1) Using Planck's principle of spectral energy distribution, calculate the maximum brightness L in the original long-wave infrared band (8–12 μm). max and minimum brightness L min :

[0090]

[0091]

[0092] Among them, T max and T min These represent the upper and lower bounds of temperature in the infrared image, respectively; ε represents emissivity; Δλ is the bandwidth of the long-wave infrared band; λ c It is the center wavelength of the long-wave infrared band;

[0093] 5.2) Calculate the brightness L0 of any point in the original band based on its gray level G0:

[0094]

[0095] 5.3) Calculate the temperature T at this point based on the brightness:

[0096]

[0097] 5.4) Calculate the brightness L1 of this point in the target band (3-5 μm):

[0098]

[0099] Where, Δλ ′ λ represents the target bandwidth (2μm in this case). c ′ This indicates the center wavelength of the target in the mid-infrared band (4μm in this case);

[0100] 5.5) Convert the brightness distribution to a grayscale distribution:

[0101]

[0102] Figure 4 The image shows a comparison between the mid-wave and long-wave simulated images after band conversion. The left image is the long-wave infrared simulation result, and the right image is the mid-wave infrared simulation result.

[0103] In one specific embodiment of the present invention, a hash algorithm is used to evaluate the credibility of simulated long-wave infrared images of the human body.

[0104] This embodiment uses the difference hashing algorithm commonly used in the field of image processing to compare the similarity between the infrared simulation results and the original infrared camera images. The implementation steps are as follows:

[0105] The resolution of the two images being compared for similarity is scaled down to 8x9 pixels.

[0106] Convert the scaled image to a grayscale image;

[0107] Calculate the average gray level of each pixel in the image;

[0108] Each pixel is compared with the next pixel in the same row. If the gray level is greater than the next pixel, it is recorded as 1; otherwise, it is recorded as 0. This generates a 64-bit hash value containing image structure information.

[0109] Compare the hash values ​​of the infrared simulation results with those of the original infrared camera images, and give the similarity between the two as a percentage.

[0110] Ultimately, calculations showed that the simulation similarity of this embodiment under these conditions was 71.88%.

[0111] The above examples are merely specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments and many variations are possible. All variations that can be directly derived or conceived by those skilled in the art from the disclosure of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for rapid reconstruction of a human target infrared 3D model, characterized in that, Includes the following steps: Step 1: Based on measured data, establish a target environment database, including spontaneous radiation, background radiation, and solar radiation data of the target environment; Step 2: Solve the temperature field distribution on the human body surface using the explicit finite difference method to obtain the relationship between the temperature at each node on the human body surface and time: ； in, The temperature of a certain point on the human body at a certain moment; This is the grid distance, i.e., the distance between two nodes; For the first The location of each node The thermal diffusivity at a certain point on the human body; For time step, This refers to the current time point; Step 3: Based on the target environment database data and the material properties of various materials on the human body surface obtained in Step 1, calculate the total radiation change of the human body surface over any given time period. According to the change in total radiation Temperature changes at various points on the human body surface Relationship This yields the temperature changes at various nodes on the human body surface; among which, The mass of the surface layer of the human body. Specific heat capacity of the human body surface area; Based on the relationship between the temperature of each node on the human body surface and time obtained in step 2, and the temperature field distribution of the human body surface at the beginning of the time period, the temperature field distribution of the human body surface at the end of the time period is solved iteratively to obtain the human body long-wave infrared simulation image. Step 4: Based on the range of environmental parameters in the target environment database, repeat step 3 to generate simulated long-wave infrared images of the human body under different time, location, and climate conditions. Step 5: Based on the Planck spectrum energy distribution principle, the long-wave infrared simulation image of the human body is converted into a mid-wave infrared simulation image of the human body, which serves as the final 3D model reconstruction result.

2. The method for rapid reconstruction of a human target infrared three-dimensional model according to claim 1, characterized in that, Step 5 includes: 5.1) Calculate the maximum brightness in the original long-wave infrared band using Planck's principle of spectral energy distribution. and minimum brightness : ； ； in, and These represent the upper and lower bounds of temperature in the infrared image, respectively. Indicates emissivity, For the long-wave infrared band width, It is the center wavelength of the long-wave infrared band; , It is a constant. ,unit , ,unit ; 5.2) The gray level of any point Calculate the brightness of this point in the original wavelength band. : ； 5.3) Calculate the temperature at that point based on the brightness. : ； 5.4) Calculate the brightness of this point in the target's medium-wave band. : ； in, The target's mid-infrared band width, The center wavelength of the target in the mid-infrared band; 5.5) Convert the brightness distribution to a grayscale distribution: ； in, The converted grayscale distribution yields the mid-wave infrared simulated image of the human body.

Images