WO2017107343A1 - 一种反差约束的气动热辐射校正方法 - Google Patents

一种反差约束的气动热辐射校正方法 Download PDF

Info

Publication number
WO2017107343A1
WO2017107343A1 PCT/CN2016/079129 CN2016079129W WO2017107343A1 WO 2017107343 A1 WO2017107343 A1 WO 2017107343A1 CN 2016079129 W CN2016079129 W CN 2016079129W WO 2017107343 A1 WO2017107343 A1 WO 2017107343A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
radiation
thermal radiation
aerodynamic
aerodynamic thermal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2016/079129
Other languages
English (en)
French (fr)
Inventor
张天序
章川
侯旋
刘立
陈泉
钟奥
许明星
周雨田
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to US15/562,887 priority Critical patent/US10127641B2/en
Publication of WO2017107343A1 publication Critical patent/WO2017107343A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20172—Image enhancement details
    • G06T2207/20182—Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30181—Earth observation
    • G06T2207/30184—Infrastructure

Definitions

  • the invention belongs to the field of cross-scientific technology combining pneumatic optics, information processing and aerospace science and technology, and particularly relates to a contrast-constrained aerodynamic thermal radiation correction method, which can be applied to image pre-preparation during high-speed aircraft remote sensing, detection, navigation and guidance. During processing.
  • Pneumatic optics is an interdisciplinary combination of aerodynamics and optics.
  • High-speed aircraft can produce severe aero-optical effects when flying in the atmosphere, seriously affecting the imaging quality of optical imaging detection systems. Therefore, the aero-optical effect and its correction method are important directions for the research of this discipline, and it is also one of the main technical problems that restrict the development and application of high-speed aircraft.
  • the interaction between the optical window and the airflow creates a complex flow field. Due to the viscosity of the air, the airflow in contact with the surface of the optical window will be retarded, causing the velocity of the airflow to decrease, forming a boundary layer near the surface of the optical window. The layers with large velocity gradients in the boundary layer will generate strong friction, and the kinetic energy of the airflow will irreversibly become thermal energy, causing the temperature of the window wall to rise. The high temperature airflow will continue to transfer heat to the low temperature wall, causing strong aerodynamic heating.
  • the optical window is pneumatically heated to be in a severe aerodynamic thermal environment, producing thermal radiation noise, reducing the signal-to-noise ratio and image quality of the photodetection system.
  • the irradiance of the airflow outside the window and the irradiance of the window are superimposed with the irradiance of the background, and the imaging sensor will enter a nonlinear region or saturate, resulting in loss of effective information of the scene, and a reduction in signal-to-noise ratio and signal-to-noise ratio.
  • Its detection performance is degraded. Therefore, it is necessary to perform aerodynamic thermal radiation correction on the image acquired by the imaging sensor to improve image quality. Since the degradation model of aerodynamic thermal radiation is unknown and randomly changing, the degraded image also contains other noise, which increases the difficulty of image restoration or correction.
  • the specific application environment of the aircraft, especially the hypersonic aircraft, and the high frame rate characteristics of the imaging system require the correction algorithm to have extremely high real-time performance.
  • the present invention provides a contrast-constrained aerodynamic thermal radiation correction method, which solves the problem that the existing aerodynamic thermal radiation correction method is poor in processing strong heat radiation effects, and when processing large images.
  • the problem of low calibration efficiency can be used for aerodynamic thermal radiation correction of images acquired during high-speed aircraft remote sensing, detection, navigation and guidance.
  • the present invention provides a method for correcting aerodynamic thermal radiation, which comprises the following steps:
  • step (6) comparing the contrast change of the image block corresponding to the radiation center region with the contrast value of the image block corresponding to the non-radiation center region, and when the difference between the two is less than or equal to a predetermined threshold, the initial corrected image S 1 is taken as The final calibration result, otherwise step (6) is performed in sequence;
  • the above method further comprises the following step (7): weighting the edge of the radiation center region of the secondary corrected image S 2 and the edge of the radiation center region of the initial corrected image S 1 to eliminate the edge caused by the segmentation Effects that bring images to higher quality.
  • the original aerodynamic thermal radiation image Z and the initial corrected image S 1 are divided into equal image blocks centering on the radiation center region of the aerodynamic thermal radiation field B 1 to make the original aerodynamic heat
  • the image block of the radiation image Z corresponding to the radiation center region of the aerodynamic thermal radiation field B 1 is at the center of all image blocks of the original aeroelastic thermal radiation image Z, such that the radiation center of the initial corrected image S 1 and the aerodynamic thermal radiation field B 1 First correction image corresponding to a central region S in an image block all image blocks.
  • the above technical solution conceived by the present invention has the following beneficial effects compared with the prior art: by statistically characterizing the aerodynamic thermal radiation image under different intensities, it is found that the image with stronger aerodynamic thermal radiation effect has a contrast. The smaller the characteristics; when using the gradient fitting algorithm for thermal radiation correction, it is found that the time consumption increases exponentially with the increase of the fitting surface order and the image size.
  • the present invention can quickly and effectively aerodynamic heat. Radiation images are restored to significantly improve the signal-to-noise ratio and image quality of the image.
  • FIG. 1 is a graph showing an image contrast value obtained by adding a gradually increasing irradiance illuminance surface to an input image and calculating a change in the contrast value thereof according to the aerodynamic thermal radiation intensity;
  • 2 is a graph showing the time-consuming curve of the gradient least squares fitting correction method with the fitting order K;
  • Figure 3 is a graph showing the time-consuming variation of the gradient least squares fitting correction method with image size
  • FIG. 4 is a flow chart of a method for correcting aerodynamic thermal radiation of a contrast constraint according to an embodiment of the present invention
  • WLS weighted squared minimum
  • Figure 8 is an experimental diagram of the arc wind tunnel experiment, in which (a) is the aerodynamic thermal radiation image of the arc wind tunnel, (b) is the unconstrained calibration result, (c) is the contrast constraint correction result, and (d) is a column of pixel display. Effect chart.
  • the invention is based on the following three characteristics of thermal radiation correction: (1) the image contrast value gradually decreases with the increase of the aerodynamic thermal radiation intensity, as shown in Fig. 1; (2) when the gradient fitting correction heat radiation is performed, the time It increases exponentially with the fitting order K, as shown in Fig. 2; (3) When the gradient fitting is used to correct the thermal radiation, the correction time increases exponentially with the image size. As shown in Fig. 3, the contrast-constrained aerodynamics is proposed. Thermal radiation correction method.
  • the contrast-constrained aerodynamic thermal radiation correction method of the embodiment of the present invention includes the following steps:
  • I is the identity matrix
  • ⁇ is the smooth parameter
  • D x and D y are forward difference operators in the x-direction and the y-direction of the image coordinate system, respectively
  • W x and W y are weighted coefficient diagonal matrices of the image coordinate system x direction and y direction, respectively.
  • the value of the diagonal element of the ith row of W x is The value of the diagonal element of the ith row of W y
  • l is the logarithmic transformation of the input image Z, with Representing the gradients in the x and y directions, respectively
  • ⁇ is a constant used to control the gradient sensitivity of the filtering to the original image. The larger ⁇ , the more sensitive it is to the gradient of the original image, and the stronger the edge retention of the filtered image. To be 1.1, ⁇ is a small constant that prevents the denominator from being 0. This example is set to 0.00001.
  • FIG. 5 is a processing effect diagram of weighted squared minimum filtering, wherein FIG. 5(a) is an original image, and FIG. 5(b) is a filtered image, which is apparent from the figure, and is subjected to weighted squared minimum filtering processing. After that, most of the details of the image are filtered out. According to this characteristic of weighted square minimization, the filtering process is performed on the aero-thermal radiation degradation image.
  • Figure 6(a) is the original infrared image
  • Figure 6(b) is the aerodynamic thermal radiation degradation image
  • Figure 6(c) is The weighted square minimizes the filtered image, filtering out high frequency components (noise and detail) in the image, making the estimation of the low frequency aerodynamic field more accurate in subsequent steps.
  • the aerodynamic thermal radiation field B can be fitted with a K-order two-dimensional polynomial, using the least squares method, from the filtered image It is estimated that the aerodynamic thermal radiation field B of the original aerodynamic thermal radiation image Z is obtained.
  • the aerodynamic thermal radiation field estimated by K 2, which is called B 1 , and the estimated initial correction residual in the subsequent step (6).
  • the filtered image is used as an input image for estimating the radiation field a second time in a subsequent step. From the initial corrected image, it can be seen that most of the radiation field of the image has been filtered out, and only a strong radiation field remains in the central portion of the radiation. Therefore, the subsequent steps focus on the estimation and removal of the thermal radiation field in the central region of the radiation. .
  • the image block contrast value is calculated by the sum of the square of the difference between the gray value of each pixel of the image block and the gray values of the surrounding four neighboring pixels, divided by the total pixels of the image block.
  • ⁇ (i,j) is the near-gradation difference of adjacent pixels
  • P ⁇ (i,j) is the pixel distribution probability that the gray level difference between adjacent pixels is ⁇
  • Ctr is the image contrast value
  • the contrast value of the image block of the original aeroelastic heat radiation image Z and the image block of the initial corrected image S 1 is separately calculated using the contrast calculation formula, as shown in Tables 1 and 2 below, wherein Table 1 is the image of the original aerodynamic thermal radiation image Z. contrast value block, table 2 First corrected image contrast value S of each image block 1, the difference between the contrast value of the original radiation image aerothermal First and Z corrected image corresponding to the image block S 1 as shown in table 3.
  • the contrast of the image is smaller, and the contrast of the image block of the original aerodynamic thermal radiation image Z obtained by the analysis step (4) with respect to the corresponding image block of the initial corrected image S 1 is compared. If the value of the change in the center of the radiation is significantly smaller than the change of the contrast value of the non-radiative central region, the difference between the two is greater than the predetermined threshold T, indicating that the original aerodynamic thermal radiation image Z has a higher radiation intensity, and the non-radiative central region The correction effect reaches a higher level, and the aerodynamic thermal radiation noise remains in the center of the radiation.
  • step (6) is performed sequentially; otherwise, if the contrast value of the radiation center area changes and non-radiation The difference in the contrast of the central region is small, and the difference between the two is less than or equal to the predetermined threshold T. This indicates that the original aerodynamic thermal radiation image Z has a low radiation intensity, and the image has been corrected better, and the initial corrected image S 1 as the final calibration result;
  • FIG. 6(f) It can be seen from Fig. 6(f) that the aerodynamic thermal radiation noise has been completely removed, but the edge effect caused by the block remains, and the weighted image is shown in Fig. 6(g).
  • Figure 6(h) is a column of pixel values of the original image, the aerodynamic thermal radiation image and the contrast-constrained correction image respectively. It can be seen from the graph that after the invention provided by the invention, the aerodynamic thermal radiation effect in the image has been obtained. Better correction.
  • High order The result of direct fitting will produce a strong halo in the center of the aerodynamic radiation, affecting the image quality.
  • the quadratic correction method of the contrast constraint overcomes this problem.
  • Figure 7(c) selects the original image, aerodynamic heat.
  • Table 4 below is the time to correct the same aerodynamic thermal radiation image using the direct correction and contrast constraint correction methods, and the PSNR (peak signal to noise ratio) parameter comparison. From the table, we can clearly see that the correction time is greatly shortened and the peak value is The signal to noise ratio has increased a little.
  • Figure (8) is an example of correction of an arc wind tunnel image
  • Figure 8 (a) is aerodynamic thermal radiation image of a wind tunnel image
  • Figure 8 (c) Contrast constraint correction image

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Processing (AREA)
  • Radiation Pyrometers (AREA)

Abstract

本发明公开了一种反差约束的气动热辐射校正方法。通过统计不同强度下的气动热辐射图像的特点，发现气动热辐射效应越强的图像，其反差越小的特点；在使用梯度拟合算法进行热辐射校正时，发现其时间消耗随着拟合曲面阶数的增长和图像大小的增长均呈指数增长趋势，本发明能快速有效地对气动热辐射图像进行恢复，显著提高图像的信噪比和图像质量。

Description

一种反差约束的气动热辐射校正方法 【技术领域】
本发明属于气动光学、信息处理和航天科技相结合的交叉科学技术领域，具体涉及一种反差约束的气动热辐射校正方法，该方法可应用于高速飞行器遥感、探测、导航和制导时的图像预处理过程中。
【背景技术】
气动光学是空气动力学与光学相结合的交叉学科，高速飞行器在大气层中飞行时会产生严重的气动光学效应，严重影响光学成像探测系统的成像质量。因此，气动光学效应及其校正方法是该学科研究的重要方向，同时也是制约高速飞行器发展应用的主要技术难题之一。
带有光学成像探测系统的高速飞行器在大气层内飞行时，光学窗口与气流之间的相互作用形成复杂的流场。由于空气粘性的作用，与光学窗口表面相接触的气流将受到阻滞，使得气流速度降低，在光学窗口表面附近形成边界层。边界层内具有很大速度梯度的各层会产生强烈的摩擦，气流的动能不可逆转地变为热能，造成窗口壁面温度的升高。高温气流将不断向低温壁面传热，引起很强的气动加热。光学窗口被气动加热而处于严重的气动热环境中，产生热辐射噪声，降低光电探测系统的信噪比和图像质量。
飞行速度越大，气流在飞行器表面加热的程度就越严重。窗外气流的辐照度和窗口的辐照度与背景的辐照度迭加，成像传感器将进入非线性区或饱和，造成景物有效信息的丢失，以及信噪比、信杂比的降低，使其探测性能下降。因此，需要对成像传感器获取的图像进行气动热辐射校正，以提高图像质量。由于气动热辐射的退化模型是未知且随机变化的，退化图像还含有其他噪声，增加了图像恢复或校正的难度。另外，针对高速飞 行器，尤其是高超音速飞行器的特定应用环境，其成像系统的高帧频特性，要求校正算法必须具有极高的实时性。
【发明内容】
针对现有技术的以上缺陷或改进需求，本发明提供了一种反差约束的气动热辐射校正方法，解决现有气动热辐射校正方法在处理强热辐射效应时效果差，并且在处理大型图像时校正效率低的问题，可用于对高速飞行器遥感、探测、导航和制导时获取的图像进行气动热辐射校正。
为实现上述目的，本发明提供了一种气动热辐射校正方法，其特征在于，包括如下步骤：
(1)滤除原始气动热辐射图像Z中的噪声和细节，得到滤波图像
Figure PCTCN2016079129-appb-000001
以克服噪声在热辐射场估计过程中产生的不利影响；
(2)从滤波图像
Figure PCTCN2016079129-appb-000002
中估计得到原始气动热辐射图像Z的气动热辐射场B1，进而得到初校正图像S1＝Z-B1；
(3)求解气动热辐射场B1的辐射中心区域，根据气动热辐射场B1的辐射中心区域，将原始气动热辐射图像Z和初校正图像S1划分为对等的图像块；
(4)分别计算原始气动热辐射图像Z的图像块的反差值以及初校正图像S1的图像块的反差值，得到原始气动热辐射图像Z的图像块相对于初校正图像S1的对应图像块的反差值变化；
(5)对比辐射中心区域对应的图像块的反差值变化与非辐射中心区域对应的图像块的反差值变化，在二者的差异均小于或等于预定的阈值时，将初校正图像S1作为最终的校正结果，否则顺序执行步骤(6)；
(6)由滤波图像
Figure PCTCN2016079129-appb-000003
和气动热辐射场B1得到图像
Figure PCTCN2016079129-appb-000004
将图像
Figure PCTCN2016079129-appb-000005
对应于气动热辐射场B1的辐射中心区域的部分作为新的滤波图像
Figure PCTCN2016079129-appb-000006
从滤波图像
Figure PCTCN2016079129-appb-000007
中估计得到初校正图像S1的辐射中心区域残余的气动热辐射场B2，进而 得到二次校正图像S2＝S1-B2。
优选地，上述方法还包括如下步骤(7)：将二次校正图像S2的辐射中心区域的边缘与初校正图像S1的辐射中心区域的边缘进行加权处理，以消除分块带来的边缘效应，使图像达到更高的质量。
优选地，所述步骤(3)中，以气动热辐射场B1的辐射中心区域为中心，将原始气动热辐射图像Z和初校正图像S1划分为对等的图像块，使原始气动热辐射图像Z的与气动热辐射场B1的辐射中心区域对应的图像块在原始气动热辐射图像Z的所有图像块的中心，使初校正图像S1的与气动热辐射场B1的辐射中心区域对应的图像块在初校正图像S1的所有图像块的中心。
总体而言，通过本发明所构思的以上技术方案与现有技术相比，具有以下有益效果：通过统计不同强度下的气动热辐射图像的特点，发现气动热辐射效应越强的图像，其反差越小的特点；在使用梯度拟合算法进行热辐射校正时，发现其时间消耗随着拟合曲面阶数的增长和图像大小的增长均呈指数增长趋势，本发明能快速有效地对气动热辐射图像进行恢复，显著提高图像的信噪比和图像质量。
【附图说明】
图1是对输入图像加入逐渐增强的辐射照度曲面，并计算其反差值变化情况得到的图像反差值随气动热辐射强度变化的曲线；
图2是梯度最小二乘拟合校正方法耗时随拟合阶数K变化曲线图；
图3是梯度最小二乘拟合校正方法耗时随图像大小变化曲线图；
图4是本发明实施例的反差约束的气动热辐射校正方法流程图；
图5是加权平方最小化(WLS)滤波效果图，其中，(a)为原始图像，(b)为滤波图像；
图6是反差约束的气动热辐射校正方法示例，其中，(a)为基准图像，(b)为气动热辐射退化图像，(c)为(b)加权平方最小化滤波结果，(d)为(b)经过K＝2的情况下，梯度拟合气动热辐射曲面的初校正结果，(e) 为(d)根据K＝2拟合的气动热辐射场和初校正结果进行分块的示意图，(f)为对(e)中心区域进行二次校正的结果图，(g)为(e)进行边缘融合的结果图，(h)为(g)中取一列像素显示，以验证气动热辐射校正结果的效果图；
图7是反差约束下的气动热辐射校正与无约束的气动热辐射校正方法对比图，其中，(a)为K＝9的校正结果图，(b)为反差约束的校正图，(c)为分别取基准图像，无约束校正图像，反差约束校正图像的一列像素显示效果图；
图8是电弧风洞实验效果图，其中，(a)为电弧风洞气动热辐射图像，(b)为无约束校正结果，(c)为反差约束校正结果，(d)为取一列像素显示效果图。
【具体实施方式】
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。此外，下面所描述的本发明各个实施方式中所涉及到的技术特征只要彼此之间未构成冲突就可以相互组合。
本发明基于热辐射校正的如下三个特点：(1)图像反差值随气动热辐射强度的增大逐渐减小，如图1所示；(2)在进行梯度拟合校正热辐射时，时间随拟合阶数K呈指数增长，如图2所示；(3)在进行梯度拟合校正热辐射时，校正时间随图像大小呈近似指数增长，如图3所示，提出反差约束的气动热辐射校正方法。
如图4所示，本发明实施例的反差约束的气动热辐射校正方法包括如下步骤：
(1)由于气动热辐射场的低频特性，利用基于加权平方最小的图像平滑算法滤除原始气动热辐射图像Z中的噪声和细节，得到滤波图像
Figure PCTCN2016079129-appb-000008
以克 服噪声在热辐射场估计过程中产生的不利影响；
具体地，滤波图像
Figure PCTCN2016079129-appb-000009
表示为：
Figure PCTCN2016079129-appb-000010
其中，I为单位矩阵，λ为光滑参数，λ越大滤波后的图像越光滑，此例设为0.002，
Figure PCTCN2016079129-appb-000011
Dx和Dy分别为图像坐标系x方向和y方向的前向差分算子，Wx和Wy分别为图像坐标系x方向和y方向的加权系数对角矩阵。
具体地，Wx的第i行对角元素的值为
Figure PCTCN2016079129-appb-000012
Wy的第i行对角元素的值为
Figure PCTCN2016079129-appb-000013
其中，l为输入图像Z的对数变换，
Figure PCTCN2016079129-appb-000014
和
Figure PCTCN2016079129-appb-000015
分别代表x方向和y方向的梯度，α是用于控制滤波对原始图像梯度敏感度的常量，α越大，其对原始图像的梯度越敏感，滤波后的图像边缘保留越强，此例设为1.1，ε为防止分母为0的较小的常量，此例设为0.00001。
图5为加权平方最小化滤波的处理效果图，其中，图5(a)为原始图像，图5(b)为滤波处理后图像，从图中可以明显看出，经过加权平方最小化滤波处理后，图像的大部分细节被滤除。根据加权平方最小化的这一特性，对气动热辐射退化图像进行该滤波处理，图6(a)为原始的红外图像，图6(b)为气动热辐射退化图像，图6(c)为加权平方最小化滤波处理后图像，滤除图像中的高频成分(噪声和细节)，使后续步骤中对低频气动辐射场的估计更精确。
(2)根据气动热辐射场B能用K阶二维多项式拟合的特点，使用最小二乘法，从滤波图像
Figure PCTCN2016079129-appb-000016
中估计得到原始气动热辐射图像Z的气动热辐射场B，本步骤中我们设K＝2估计得出的气动热辐射场，称之为B1，后续步骤(6)中估计的初校正残余气动热辐射场，称之为B2。由得到的气动热辐射场B1进而得到初校正图像S1＝Z-B1；
Figure PCTCN2016079129-appb-000017
其中，(x,y)为像素坐标；a为多项式系数，C为把x和y代入上式的常数矩阵，K为二维多项式的阶数。
Figure PCTCN2016079129-appb-000018
利用最小二乘法估计，滤波图像
Figure PCTCN2016079129-appb-000019
的梯度
Figure PCTCN2016079129-appb-000020
与待估计的气动热辐射场B的梯度
Figure PCTCN2016079129-appb-000021
之差的L2范数达到最小时的解，即为求出的气动热辐射场B，如上式。
此方法在求解的过程中，消耗时间随着K的增长呈指数增长趋势，如图2所示，因此，在首次使用上述方法的过程中，设定K＝2解出气动热辐射场B1，从而得出初校正图像S1，如图6(d)所示，并得出滤波图像减去辐射场B1的图像
Figure PCTCN2016079129-appb-000022
作为滤波图像用于后续步骤中第二次估计辐射场的输入图像。从初校正图像中，能看出图像的大部分辐射场已经被滤除，只有辐射中心区域部分残留着较强的辐射场，因此，后续步骤重点对辐射中心区域进行热辐射场的估计和去除。
(3)求解气动热辐射场B1的辐射中心区域，该区域对应热辐射场图像中灰度值较大的区域，根据气动热辐射场B1的辐射中心区域，将原始气动热辐射图像Z和初校正图像S1划分为对等的图像块；
例如，使用求得的辐射中心区域对输入的气动热辐射图像如图6(b)和初校正图像S1如图6(d)，进行分块处理，以辐射中心区域为中心分为9块，分块结果如图6(e)所示。
(4)分别计算原始气动热辐射图像Z的图像块的反差值以及初校正图像S1的图像块的反差值，得到原始气动热辐射图像Z的图像块相对于初校正图像S1的对应图像块的反差值变化；
例如，图像块反差值的计算方法为：图像块每个像素灰度值与周围4个近邻像素灰度值之差的平方之和，除以图像块的总像素。
Ctr＝∑δδ(i,j)2Pd(i,j)
δ(i,j)相邻像素近灰度差，Pδ(i,j)为相邻像素间的灰度差为δ的像素分布概率，Ctr为图像反差值。
使用反差计算公式分别计算原始气动热辐射图像Z的图像块和初校正图像S1的图像块的反差值，如下表1和2所示，其中，表1为原始气动热辐射图像Z的各图像块反差值，表2为初校正图像S1的各图像块的反差值，原始气动热辐射图像Z与初校正图像S1的对应图像块的反差值之差如表3所示。
表1
41.5 26.7 83.2
43.7 18.0 70.3
100.3 37.1 22.8
表2
126.0 169.8 337.4
220.9 44.6 271.5
237.3 126.2 103.8
表3
84.4 143.1 254.2
177.1 26.5 201.1
137.0 89.1 80.9
(5)根据气动热辐射图像辐射越强其图像反差值越小的特点，对比分析步骤(4)得到的原始气动热辐射图像Z的图像块相对于初校正图像S1的对应图像块的反差值变化，若辐射中心区域的反差值变化明显小于非辐射中心区域的反差值变化，二者的差异大于预定的阈值T，则说明原始气动热辐射图像Z的辐射强度较高，非辐射中心区域校正效果达到较高水平，辐射中心区域仍残留较强的气动热辐射噪声，需要对辐射中心区域进行二次校正，顺序执行步骤(6)；反之，若辐射中心区域的反差值变化与非辐射中 心区域的反差值变化相差较小，二者的差异均小于或等于预定的阈值T，则说明原始气动热辐射图像Z的辐射强度较低，图像已经得到较好的校正，将初校正图像S1作为最终的校正结果；
从上表(c)中可以明显的看出辐射中心区域的反差值恢复情况明显低于周边区域，这表示辐射中心区域仍残留这较强的气动热辐射噪声，需对辐射中心区域进行二次校正。
(6)由步骤(1)中得出的滤波图像
Figure PCTCN2016079129-appb-000023
和步骤(2)中得出的气动热辐射场B1得到图像
Figure PCTCN2016079129-appb-000024
将图像
Figure PCTCN2016079129-appb-000025
对应于气动热辐射场B1的辐射中心区域的部分作为新的滤波图像
Figure PCTCN2016079129-appb-000026
从滤波图像
Figure PCTCN2016079129-appb-000027
中估计得到初校正图像S1的辐射中心区域残余的气动热辐射场B2，进而得到二次校正图像S2＝S1-B2；
在此处的拟合过程中，因为只对图像的中心较小的区域进行拟合处理，可以使用相对于第二步中较高的K进行拟合，而不会带来很高的时间代价，本实验中K＝5。初校正图像中心区域减去估计的中心区域气动热辐射场得到二次校正图像S2，如图6(f)所示。
(7)由于对图像的分块处理会不可避免地带来边缘效应，在该步骤中，将二次校正图像S2的辐射中心区域的边缘与初校正图像S1的辐射中心区域的边缘进行加权处理，以消除分块带来的边缘效应，使图像达到更高的质量。
从图6(f)中可以看出气动热辐射噪声已经完全去除，但残留着分块带来的边缘效应，经过加权处理后的图像如图6(g)所示。图6(h)为分别选取原始图像、气动热辐射图像和反差约束校正图像的一列像素值，从曲线图可以看出，经过本发明提供的发明处理之后，图像中的气动热辐射效应已得到较好的校正。
如图7所示为直接使用K＝9梯度拟合气动辐射场的结果，图7(a)，和反差约束二次校正的校正结果，图7(b)，从图中可以明显看出，高阶 直接拟合的结果会在气动辐射中心区域产生较强的光晕圈，影响图像质量，反差约束的二次校正方法则克服了这一问题，图7(c)为分别选取原始图像，气动热辐射图像，K＝9的高阶二维多项式拟合校正图像，反差约束校正图像的一列像素值，从中可以看出在保证效果的同时减小校正时间。下表4是分别使用直接校正和反差约束校正方法对同一气动热辐射图像进行校正的时间，PSNR(峰值信噪比)的参数对比，从表中我们可以明显看出，校正时间大幅缩短，峰值信噪比有少许增加。
表4
Figure PCTCN2016079129-appb-000028
图(8)为电弧风洞图像的校正示例，图8(a)为风洞图像的气动热辐射图像，图8(b)为K＝9的高阶二维多项式拟合校正图像，图8(c)反差约束校正图像，图8(d)为分别取电弧风洞图像的气动热辐射图像，高阶(K＝9)二维多项式拟合校正图像，反差约束校正图像的一列像素值。
本领域的技术人员容易理解，以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。

Claims (3)

  1. 一种气动热辐射校正方法，其特征在于，包括如下步骤：
    (1)滤除原始气动热辐射图像Z中的噪声和细节，得到滤波图像
    Figure PCTCN2016079129-appb-100001
    以克服噪声在热辐射场估计过程中产生的不利影响；
    (2)从滤波图像
    Figure PCTCN2016079129-appb-100002
    中估计得到原始气动热辐射图像Z的气动热辐射场B1，进而得到初校正图像S1＝Z-B1；
    (3)求解气动热辐射场B1的辐射中心区域，根据气动热辐射场B1的辐射中心区域，将原始气动热辐射图像Z和初校正图像S1划分为对等的图像块；
    (4)分别计算原始气动热辐射图像Z的图像块的反差值以及初校正图像S1的图像块的反差值，得到原始气动热辐射图像Z的图像块相对于初校正图像S1的对应图像块的反差值变化；
    (5)对比辐射中心区域对应的图像块的反差值变化与非辐射中心区域对应的图像块的反差值变化，在二者的差异均小于或等于预定的阈值时，将初校正图像S1作为最终的校正结果，否则顺序执行步骤(6)；
    (6)由滤波图像
    Figure PCTCN2016079129-appb-100003
    和气动热辐射场B1得到图像
    Figure PCTCN2016079129-appb-100004
    将图像
    Figure PCTCN2016079129-appb-100005
    对应于气动热辐射场B1的辐射中心区域的部分作为新的滤波图像
    Figure PCTCN2016079129-appb-100006
    从滤波图像
    Figure PCTCN2016079129-appb-100007
    中估计得到初校正图像S1的辐射中心区域残余的气动热辐射场B2，进而得到二次校正图像S2＝S1-B2。
  2. 如权利要求1所述的气动热辐射校正方法，其特征在于，还包括如下步骤(7)：将二次校正图像S2的辐射中心区域的边缘与初校正图像S1的辐射中心区域的边缘进行加权处理，以消除分块带来的边缘效应，使图像达到更高的质量。
  3. 如权利要求1或2所述的气动热辐射校正方法，其特征在于，所述步骤(3)中，以气动热辐射场B1的辐射中心区域为中心，将原始气动热辐 射图像Z和初校正图像S1划分为对等的图像块，使原始气动热辐射图像Z的与气动热辐射场B1的辐射中心区域对应的图像块在原始气动热辐射图像Z的所有图像块的中心，使初校正图像S1的与气动热辐射场B1的辐射中心区域对应的图像块在初校正图像S1的所有图像块的中心。
PCT/CN2016/079129 2015-12-24 2016-04-13 一种反差约束的气动热辐射校正方法 Ceased WO2017107343A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/562,887 US10127641B2 (en) 2015-12-24 2016-04-13 Contrast constrained aerothermal radiation correction method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510988503.9A CN105654430B (zh) 2015-12-24 2015-12-24 一种反差约束的气动热辐射校正方法
CN201510988503.9 2015-12-24

Publications (1)

Publication Number Publication Date
WO2017107343A1 true WO2017107343A1 (zh) 2017-06-29

Family

ID=56476780

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/079129 Ceased WO2017107343A1 (zh) 2015-12-24 2016-04-13 一种反差约束的气动热辐射校正方法

Country Status (3)

Country Link
US (1) US10127641B2 (zh)
CN (1) CN105654430B (zh)
WO (1) WO2017107343A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10521888B2 (en) * 2015-12-23 2019-12-31 Huazhong University Of Science And Technology Aerothermal radiation effect frequency domain correction method
CN108288030B (zh) * 2017-12-29 2021-05-18 华中科技大学 一种机载红外小目标检测识别实时处理系统和方法
CN111583149B (zh) * 2020-05-08 2023-02-17 西安电子科技大学 一种基于l1-l0范数最小化的气动热辐射图像自动校正方法
CN114022399B (zh) * 2022-01-07 2022-04-19 武汉工程大学 一种气动光学热辐射校正方法、电子设备及可读存储介质
CN114529481B (zh) * 2022-04-21 2022-07-26 武汉工程大学 气动光学热辐射效应校正方法、装置、设备及存储介质
CN114662369B (zh) * 2022-05-19 2022-08-09 中国飞机强度研究所 一种空天飞机复杂曲面结构大梯度极高温热强度评估方法
CN116304541B (zh) * 2023-05-25 2023-09-01 山东大学 一种基于多源地球物理参数的火山爆发分析方法及系统
CN119090724B (zh) * 2024-11-05 2025-03-25 中国人民解放军火箭军工程大学 一种遥感图像相对辐射校正方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101261176A (zh) * 2008-04-03 2008-09-10 华中科技大学 基于序列图像校正的气动光学传输效应测评方法与装置
CN103093425A (zh) * 2012-12-31 2013-05-08 华中科技大学 一种气动热辐射指纹库的建立方法及其应用
CN105118037A (zh) * 2015-09-02 2015-12-02 华中科技大学 一种气动热辐射图像自动校正方法
CN105160687A (zh) * 2015-10-21 2015-12-16 华中科技大学 一种气动光学热辐射噪声校正方法与系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102081737B (zh) * 2010-12-30 2012-10-10 华中科技大学 一种气动热辐射图像多尺度建模方法及其应用
US9927289B2 (en) * 2015-10-23 2018-03-27 Raytheon Company Polarization filtering for window self-emission due to aero-thermal heating

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101261176A (zh) * 2008-04-03 2008-09-10 华中科技大学 基于序列图像校正的气动光学传输效应测评方法与装置
CN103093425A (zh) * 2012-12-31 2013-05-08 华中科技大学 一种气动热辐射指纹库的建立方法及其应用
CN105118037A (zh) * 2015-09-02 2015-12-02 华中科技大学 一种气动热辐射图像自动校正方法
CN105160687A (zh) * 2015-10-21 2015-12-16 华中科技大学 一种气动光学热辐射噪声校正方法与系统

Also Published As

Publication number Publication date
CN105654430A (zh) 2016-06-08
CN105654430B (zh) 2017-04-19
US20180286017A1 (en) 2018-10-04
US10127641B2 (en) 2018-11-13

Similar Documents

Publication Publication Date Title
WO2017107343A1 (zh) 一种反差约束的气动热辐射校正方法
CN100535618C (zh) 一种红外焦平面阵列非均匀性自适应校正方法
CN105160687B (zh) 一种气动光学热辐射噪声校正方法与系统
CN109272520B (zh) 一种联合运动指导与边缘检测的自适应红外焦平面非均匀校正方法
CN105118037B (zh) 一种气动热辐射图像自动校正方法
CN105654432B (zh) 一种气动热辐射效应的频域校正方法
Liu et al. Optics Temperature-Dependent Nonuniformity Correction Via $\ell _ {0} $-Regularized Prior for Airborne Infrared Imaging Systems
WO2019183843A1 (zh) 基于帧间配准和自适应步长的红外图像非均匀性校正方法
CN110147816B (zh) 一种彩色深度图像的获取方法、设备、计算机存储介质
CN109993174B (zh) 一种基于噪声抑制的空中目标检测方法及系统
CN115752757A (zh) 探测器和内黑体分离的红外扫描相机在轨非均匀校正方法
CN104574412A (zh) 一种非均匀云雾条件下的遥感图像去雾方法
WO2019127059A1 (zh) 基于引导滤波和高通滤波的红外图像非均匀性校正方法
CN106056557B (zh) 一种基于改进大气散射模型的单幅图像快速去雾方法
CN104809706B (zh) 一种基于图像颜色平缓变化先验的单透镜计算成像方法
Liu et al. Correction of aeroheating-induced intensity nonuniformity in infrared images
CN103093425B (zh) 一种气动热辐射指纹库的建立方法及其应用
CN111583149A (zh) 一种基于l1-l0范数最小化的气动热辐射图像自动校正方法
CN106709876B (zh) 一种基于暗像元原理的光学遥感图像去雾方法
CN103985089B (zh) 结合权重边缘分析与帧内迭代的图像条纹校正方法
US10521888B2 (en) Aerothermal radiation effect frequency domain correction method
CN111076815B (zh) 一种高光谱图像非均匀性校正方法
CN119418063B (zh) 高超飞行器热辐射效应曲面回归在线校正方法及系统
CN102081737B (zh) 一种气动热辐射图像多尺度建模方法及其应用
Crawford et al. Quantitative boundary-layer transition measurements using IR thermography

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16877157

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 15562887

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16877157

Country of ref document: EP

Kind code of ref document: A1