WO2020224518A1 - 基于局部中值直方图的自适应红外图像去条纹算法 - Google Patents

基于局部中值直方图的自适应红外图像去条纹算法 Download PDF

Info

Publication number
WO2020224518A1
WO2020224518A1 PCT/CN2020/088056 CN2020088056W WO2020224518A1 WO 2020224518 A1 WO2020224518 A1 WO 2020224518A1 CN 2020088056 W CN2020088056 W CN 2020088056W WO 2020224518 A1 WO2020224518 A1 WO 2020224518A1
Authority
WO
WIPO (PCT)
Prior art keywords
window
column
pixels
gray value
pixel
Prior art date
Application number
PCT/CN2020/088056
Other languages
English (en)
French (fr)
Inventor
隋修宝
陈扬
陈钱
顾国华
王利平
蔡思聪
朱亮亮
于雪莲
蔡钰珏
张文辉
Original Assignee
南京理工大学
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 南京理工大学 filed Critical 南京理工大学
Publication of WO2020224518A1 publication Critical patent/WO2020224518A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/40—Image enhancement or restoration using histogram techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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

Definitions

  • the invention belongs to the field of infrared image non-uniformity correction, and specifically relates to an adaptive infrared image de-streaking algorithm based on a local median histogram.
  • infrared images have been widely used in industrial, medical, and military fields for detection under low visibility.
  • the infrared imaging system has uniformly radiated infrared light, and the gray value of each pixel on the acquired digital image should be exactly the same.
  • the photosensitive elements (pixels) on the detector are often accompanied by uneven impurity concentration, unequal thickness, and the effective photosensitive area cannot be absolutely averaged.
  • the photoelectric conversion efficiency is different, and the imaging of the uniform radiation scene is not uniform.
  • the difference between the channels of the image data readout circuit will result in fixed fringe noise in the image. This requires us to perform non-uniformity correction of the image, so that the image can get a better visual effect.
  • Commonly used infrared image non-uniformity correction techniques mainly include calibration method and scene method.
  • the calibration method mainly includes two-point correction method, multi-point correction method, etc.; scene method mainly includes time-domain high-pass filtering method, constant statistical method, neural network method, etc. But these two methods have great limitations in application.
  • the above method can be used on infrared images to calculate the middle histogram of a certain column based on the cumulative histogram of adjacent columns, and replace the middle histogram with the middle histogram. Cumulative histogram. In this way, the information of each column in a single image is applied to other columns to correct the non-uniformity of a single image.
  • this method uses the premise, the histograms between image columns are not much different. When there are complex scenes in the image, there may be unsatisfactory conditions such as poor correction effect or even banding effect and distortion.
  • the purpose of the present invention is to provide a single-frame infrared image non-uniformity correction method based on a median histogram for removing fringe noise in the infrared image.
  • Step 1 Collect an original infrared image o(i,j) with a pixel number of M ⁇ N, where i ⁇ 1,...,M ⁇ , j ⁇ 1,...,N ⁇ , i represents the number of pixels on the image
  • i represents the number of pixels on the image
  • the position of the column, j represents the position of the row of pixels on the image
  • M represents the number of columns of the collected image
  • N represents the number of rows of the collected image
  • Step 2 For the x-th column of the original infrared image o(i,j), a sliding window of size A ⁇ B is constructed with the x-th column as the center column, where A is the number of columns of the window, and is an odd number, B is the number of rows in the window, calculate the complexity of the scene in the window ⁇ :
  • k represents the pixel gray value
  • L represents the maximum gray value
  • p(k) represents the number of pixels with gray value k in the window, the larger the ⁇ , the more complex the scene
  • Step 3 Move the window up and down pixel by pixel, calculate the scene complexity of all windows with the xth column as the center column, and compare them to find the window with the least scene complexity with the xth column as the center column;
  • Step 4 Perform median histogram equalization on the x-th column in the window with the smallest scene complexity with the x-th column as the center column to obtain the corrected pixel gray value d(x, j);
  • Step 5 Subtract the average value of the corrected gray value of the pixel in the center column of the window from the original gray value of the pixel in the center column of the window to obtain the stripe value of the x-th column, that is, the correction parameter S x of the x-th column:
  • Step 6 The gray value of all pixels in the xth column of the original infrared image o(i,j) is subtracted from the correction parameter S x of the xth column, and the resulting gray value is the final output pixel gray of the xth column Degree value
  • Step 7. Perform steps 2) to 6) for each column of the original infrared image o(i,j) to correct all columns to remove the fringe noise on the original infrared image o(i,j).
  • the present invention has the following remarkable advantages:
  • the single frame image can be processed, which avoids the problem of the scene method requiring multiple frames of images to converge and the problem of ghosting.
  • Fig. 1 is a flowchart of an adaptive infrared image de-streaking algorithm based on a local median histogram of the present invention.
  • Figure 2 is an infrared image of a real scene with fringe noise and its effect after processing with different methods.
  • Figure (a) is an infrared image without fringe noise processing;
  • Figure (b) is a traditional median histogram In the image processed by the equalization algorithm, it can be seen that the image at the box is distorted;
  • Figure (c) is the image processed by the adaptive infrared image de-streaking algorithm based on the local median histogram of the present invention.
  • the traditional infrared image de-streaking algorithm based on median equalization is based on the following principle: one column of a single image contains enough histogram information, plus continuous imaging, the change of adjacent columns is very small. The closer the columns are to the current column, the more similar their histograms will be. Compared with other columns that are farther away, they account for a larger proportion in the calculation. Therefore, for an image with fringe noise, the cumulative histogram of the adjacent column of the current column is weighted by the Gaussian formula to obtain the intermediate histogram of the current column. Then the corrected gray value is obtained by inverting the middle histogram. But this method is only suitable for scenes where the histograms of adjacent columns have little difference. For those where the histograms of adjacent columns have large differences, as shown in Figure 2(a), new stripes will be introduced, as shown in Figure 2(b). .
  • this paper proposes an improved method based on the traditional infrared image de-streaking algorithm based on median equalization.
  • a sliding window the area where the scene change between the current column and the adjacent column is the smallest is found.
  • Perform median histogram equalization in the window and then apply the calculated fringe value of the current column in the window to the entire column, then the entire column of the current column can be removed, as shown in Figure 2(c).
  • an adaptive infrared image de-streaking algorithm based on local median histograms includes the following steps:
  • Step 1 Collect an original infrared image o(i,j) with a pixel number of M ⁇ N, where i ⁇ 1,...,M ⁇ , j ⁇ 1,...,N ⁇ , i represents the number of pixels on the image
  • i represents the number of pixels on the image
  • j represents the position of the row of pixels on the image
  • M represents the number of columns of the captured image
  • N represents the number of rows of the captured image.
  • Step 2 For the x-th column of the original infrared image o(i,j), a sliding window of size A ⁇ B is constructed with the x-th column as the center column, where A is the number of columns of the window, and is an odd number, B is the number of rows in the window, calculate the complexity of the scene in the window ⁇ :
  • k represents the pixel gray value
  • L represents the maximum gray value
  • p(k) represents the number of pixels with a gray value of k in the window, the larger the ⁇ , the more complex the scene.
  • Step 3 Move the window up and down pixel by pixel, calculate the scene complexity of all windows with the xth column as the center column, and compare them to find the window with the least scene complexity with the xth column as the center column.
  • moving the window up and down pixel by pixel refers to only moving up or down one pixel at a time until all the pixels in this column are traversed.
  • Step 4 In the window with the smallest scene complexity with the xth column as the center column, perform median histogram equalization on the xth column to obtain the corrected pixel gray value d(x, j).
  • the specific steps are as follows:
  • B is the number of rows in the window, that is, the total number of pixels in each column of the window;
  • k represents the gray value,
  • H i (k) represents the number of pixels with gray value k in the i-th column of the window;
  • l is the gray value
  • H i (l) represents the number of pixels with gray value less than or equal to l in the i-th column of the window
  • A is the number of columns in the window
  • n is the distance from other columns in the window to the center column
  • g(n) is the Gaussian weight function:
  • ⁇ 2 is the variance of the window
  • Step 5 Subtract the average value of the corrected gray value of the pixel in the center column of the window from the original gray value of the pixel in the center column of the window to obtain the stripe value of the x-th column, that is, the correction parameter S x of the x-th column:
  • Is the mean value of the corrected gray value of the pixels in the center column of the window and is the ratio of the sum of the gray values of all pixels in the center column of the window after correction to the number of pixels in the center column of the window
  • Step 6 The gray value of all pixels in the xth column of the original infrared image o(i,j) is subtracted from the correction parameter S x of the xth column, and the resulting gray value is the final output pixel gray of the xth column Degree value
  • Step 7. Perform steps 2) to 6) for each column of the original infrared image o(i,j) to correct all columns to remove the fringe noise on the original infrared image o(i,j).

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Picture Signal Circuits (AREA)

Abstract

本发明公开了一种基于局部中值直方图的自适应红外图像去条纹算法，通过计算滑动窗口内场景复杂度，找出原始红外图像中每一列与相邻列场景变化最小的窗口。计算该窗口所有列的累积直方图，得到窗口中心列的中值直方图，根据中值直方图计算出窗口中心列的各个像素校正后的灰度值，用原始灰度值均值减去校正后的灰度值均值就可以得到该列的条纹值。将整列都减去条纹值即为该列最终输出的灰度值。本发明通过滑动窗口找到图像中场景变化不明显的区域，校正效果不受场景影响，可以消除传统中值直方图均衡去条纹算法中图像场景变化剧烈时引入新的噪声的问题。

Description

基于局部中值直方图的自适应红外图像去条纹算法 技术领域
本发明属于红外图像非均匀性校正领域，具体涉及一种基于局部中值直方图的自适应红外图像去条纹算法。
背景技术
目前，红外图像已被广泛应用于工业、医学和军事等领域来进行低可视度下的侦测。在理想情况下，红外成像系统对均匀辐射的红外光，获取的数字图像上每个像素点的灰度值应该完全一样。但实际上，受限于固态电子的制造工艺，探测器上的光敏元件(像元)往往伴随着杂质浓度不均，厚度不等，有效光敏面积做不到绝对平均等问题，像元之间的光电转换效率各不相同，对均匀辐射的景物的成像不均匀。另外，图像数据读出电路各通道之间的差，会导致图像出现呈列分布的固定条纹噪声。这就要求我们图像进行非均匀性校正，使图像得到更好的视觉效果。
常用的红外图像非均匀校正技术主要有定标法和场景法两种。定标法主要有两点校正法、多点校正法等；场景法主要有时域高通滤波法、恒定统计法、神经网络法等。但这两种方法在应用时都有很大的局限性。
近年来，国内外学者开始关注在在静态场景中，或者说在单帧图像中去除红外条纹非均匀性的方法。
Tendero和Gilles研究在单帧图像内去除图像上的非均匀性，提出了一种利用中值直方图均衡化的去条纹非均匀性校正算法。中值直方图最初是用来修正各相机中传感器增益之间的差异。若两幅图像的累计直方图分别是H 1和H 2，则其中间直方图的计算公式如下：
Figure PCTCN2020088056-appb-000001
由于红外图像固定条纹噪声一般来说并不是孤立的，所以可以将上述方法用在红外图像上，根据相邻列的累计直方图计算某一列的中间直方图，用中间直方图来替代该列的累计直方图。通过这种方法使单幅图像上每一列的信息作用到其他列上，以进行单幅图像的非均匀性校正。该方法使用前提时图像列与列之间直 方图相差不大。在图像中有复杂场景时，可能会出现校正效果不佳甚至出现带状效应，发生畸变等不理想的情况。
发明内容
本发明的目的在于提供一种基于中值直方图的单帧红外图像非均匀性校正方法，用于去除红外图像中的条纹噪声。
实现本发明目的的技术解决方案为：一种基于局部中值直方图的自适应红外图像去条纹算法，实现步骤如下：
步骤1、采集一幅像素数为M×N的原始红外图像o(i,j)，其中i∈{1,…,M}，j∈{1,…,N}，i表示图像上像素的列所在位置，j表示图像上像素的行所在位置，M表示采集图像的列数，N表示采集图像的行数；
步骤2、对于上述原始红外图像o(i,j)的第x列，以第x列为中心列构造一个大小为A×B的滑动窗口，其中，A为窗口的列数，且为奇数，B为窗口的行数，计算窗口内场景复杂度μ：
Figure PCTCN2020088056-appb-000002
其中，k表示像素灰度值，L表示最大灰度值，
Figure PCTCN2020088056-appb-000003
表示窗口中所有像素灰度值均值，p(k)表示窗口中灰度值为k的像素的个数，μ越大，表示场景越复杂；
步骤3、逐像素地上下移动窗口，计算以第x列为中心列的所有窗口的场景复杂度，并进行比较，找出以第x列为中心列的场景复杂度最小的窗口；
步骤4、在以第x列为中心列的场景复杂度最小的窗口中，对第x列进行中值直方图均衡，得到校正后的像素灰度值d(x,j)；
步骤5、用窗口内中心列的像素原始灰度值均值减去窗口内中心列的像素校正后的灰度值均值，得到第x列的条纹值，即第x列的校正参数S x：
Figure PCTCN2020088056-appb-000004
其中，
Figure PCTCN2020088056-appb-000005
为窗口内中心列的像素校正后的灰度值均值，
Figure PCTCN2020088056-appb-000006
为窗口内中心列的像素的原始灰度值均值；
步骤6、将原始红外图像o(i,j)第x列的所有像素的灰度值都减去第x列的校正参数S x，得到的灰度值即为最终输出的第x列像素灰度值；
步骤7、对原始红外图像o(i,j)每列都进行步骤2)-步骤6)的操作，即可对所有列进行校正，去除原始红外图像o(i,j)上的条纹噪声。
本发明与现有技术相比，其显著的优点为：
(1)减少了定标法的工作量，避免多次反复标定。
(2)对单帧图像即可进行处理，避免了场景法需要多帧图像才能收敛的问题以及鬼影问题。
(3)与普通的中值直方图均衡非均匀性校正算法相比，剔除了复杂场景的影响，避免了图像发生畸变。
附图说明
图1是本发明基于局部中值直方图的自适应红外图像去条纹算法的流程图。
图2是具有条纹噪声的真实场景的红外图像和用不同方法处理后的效果图，其中图(a)为未对条纹噪声进行过处理的红外图像；图(b)为经过传统中值直方图均衡算法处理的图像，可以看到方框处图像发生了畸变；图(c)为经过本发明的基于局部中值直方图的自适应红外图像去条纹算法处理后的图像。
具体实施方式
下面结合附图进一步详细说明。
传统的基于中值均衡的红外图像去条纹算法基于以下原理：单幅图像的一列包含了足够的直方图信息，加上连续成像，相邻列的变化非常小。和当前列相邻越近的列，它们的直方图就越相似，计算的时候相比于其他离得远的列来说，占得比重更大。因此对于有条纹噪声的图像，对当前列的相邻列的累积直方图采用高斯公式加权，来得到当前列的中间直方图。再通过对中间直方图求逆得到校正后的灰度值。但这种方法只适用于相邻列直方图差别不大的场景，对于相邻列直方图差别较的大的，如图2(a)，就会引入新的条纹，如图2(b)。
因此，本文在传统的基于中值均衡的红外图像去条纹算法的基础上，提出了一种改进方法，通过滑动窗口，找出当前列与相邻列场景变化最小的区域。在窗口内进行中值直方图均衡，再将窗口内当前列计算出的条纹值，应用到整列，就可去除当前列整列的条纹，如图2(c)。
结合图1，一种基于局部中值直方图的自适应红外图像去条纹算法，包括以下步骤：
步骤1、采集一幅像素数为M×N的原始红外图像o(i,j)，其中i∈{1,…,M}，j∈{1,…,N}，i表示图像上像素的列所在位置，j表示图像上像素的行所在位置，M表示采集图像的列数，N表示采集图像的行数。
步骤2、对于上述原始红外图像o(i,j)的第x列，以第x列为中心列构造一个大小为A×B的滑动窗口，其中，A为窗口的列数，且为奇数，B为窗口的行数，计算窗口内场景复杂度μ：
Figure PCTCN2020088056-appb-000007
其中，k表示像素灰度值，L表示最大灰度值，
Figure PCTCN2020088056-appb-000008
表示窗口中所有像素灰度值均值，p(k)表示窗口中灰度值为k的像素的个数，μ越大，表示场景越复杂。
步骤3、逐像素地上下移动窗口，计算以第x列为中心列的所有窗口的场景复杂度，并进行比较，找出以第x列为中心列的场景复杂度最小的窗口。其中，逐像素地上下移动窗口指每次仅上移或下移一个像素，直到历遍本列的所有像素。
步骤4、在以第x列为中心列的场景复杂度最小的窗口中，对第x列进行中值直方图均衡，得到校正后的像素灰度值d(x,j)，具体步骤如下：
4-1)计算窗口中每一列的统计直方图：
Figure PCTCN2020088056-appb-000009
其中，B为窗口的行数，即窗口中每一列的总像素个数；k表示灰度值，o(i,j)＝k表示窗口中(i,j)位置的像素灰度值为k，h i(k)表示窗口中第i列中灰度值为k的像素的个数；
4-2)根据窗口中每一列的统计直方图计算累积直方图：
Figure PCTCN2020088056-appb-000010
其中，l为灰度值，H i(l)表示窗口中第i列中灰度值小于等于l的像素的个数；
4-3)对窗口中每一列的累积直方图求逆，得到
Figure PCTCN2020088056-appb-000011
4-4)对窗口内每一列的
Figure PCTCN2020088056-appb-000012
进行高斯加权，得到窗口内中心列的中值直方图：
Figure PCTCN2020088056-appb-000013
其中，A为窗口的列数，n为窗口中其他列到中心列的距离，g(n)为高斯权重函数：
Figure PCTCN2020088056-appb-000014
σ 2为窗口的方差；
4-5)窗口内中心列的像素校正后的灰度值为：
Figure PCTCN2020088056-appb-000015
步骤5、用窗口内中心列的像素原始灰度值均值减去窗口内中心列的像素校正后的灰度值均值，得到第x列的条纹值，即第x列的校正参数S x：
Figure PCTCN2020088056-appb-000016
其中，
Figure PCTCN2020088056-appb-000017
是窗口内中心列的像素的原始灰度值均值，为校正前窗口内中心列所有像素的灰度值之和与窗口内中心列像素个数的比值，
Figure PCTCN2020088056-appb-000018
为窗口内中心列的像素校正后的灰度值均值，为校正后窗口内中心列所有像素的灰度值之和与窗口内中心列像素个数的比值；
步骤6、将原始红外图像o(i,j)第x列的所有像素的灰度值都减去第x列的校正参数S x，得到的灰度值即为最终输出的第x列像素灰度值；
步骤7、对原始红外图像o(i,j)每列都进行步骤2)-步骤6)的操作，即可对所有列进行校正，去除原始红外图像o(i,j)上的条纹噪声。

Claims (4)

  1. 一种基于局部中值直方图的自适应红外图像去条纹算法，其特征在于，包括以下步骤：
    步骤1、采集一幅像素数为M×N的原始红外图像o(i,j)，其中i∈{1,…,M}，j∈{1,…,N}，i表示图像上像素的列所在位置，j表示图像上像素的行所在位置，M表示采集图像的列数，N表示采集图像的行数；
    步骤2、对于上述原始红外图像o(i,j)的第x列，以第x列为中心列构造一个大小为A×B的滑动窗口，其中，A为窗口的列数，且为奇数，B为窗口的行数，计算窗口内场景复杂度μ：
    Figure PCTCN2020088056-appb-100001
    其中，k表示像素灰度值，L表示最大灰度值，
    Figure PCTCN2020088056-appb-100002
    表示窗口中所有像素灰度值均值，p(k)表示窗口中灰度值为k的像素的个数，μ越大，表示场景越复杂；
    步骤3、逐像素地上下移动窗口，计算以第x列为中心列的所有窗口的场景复杂度，并进行比较，找出以第x列为中心列的场景复杂度最小的窗口；
    步骤4、在以第x列为中心列的场景复杂度最小的窗口中，对第x列进行中值直方图均衡，得到校正后的像素灰度值d(x,j)；
    步骤5、用窗口内中心列的像素原始灰度值均值减去窗口内中心列的像素校正后的灰度值均值，得到第x列的条纹值，即第x列的校正参数S x：
    Figure PCTCN2020088056-appb-100003
    其中，
    Figure PCTCN2020088056-appb-100004
    为窗口内中心列的像素校正后的灰度值均值，
    Figure PCTCN2020088056-appb-100005
    为窗口内中心列的像素的原始灰度值均值；
    步骤6、将原始红外图像o(i,j)第x列的所有像素的灰度值都减去第x列的校正参数S x，得到的灰度值即为最终输出的第x列像素灰度值；
    步骤7、对原始红外图像o(i,j)每列都进行步骤2)-步骤6)的操作，即可对所有列进行校正，去除原始红外图像o(i,j)上的条纹噪声。
  2. 根据权利要求1所述的基于局部中值直方图的自适应红外图像去条纹算法，其特征在于：上述步骤3中，逐像素地上下移动窗口指每次仅上移或下移一个像素，直到历遍本列的所有像素。
  3. 根据权利要求1所述的基于局部中值直方图的自适应红外图像去条纹算法，其特征在于：上述步骤4中，在以第x列为中心列的场景复杂度最小的窗口中，对第x列进行中值直方图均衡，具体步骤为：
    4-1)计算窗口中每一列的统计直方图：
    h i(k)＝∑ B1 {o(i，j)＝k}
    其中，B为窗口的行数，即窗口中每一列的总像素个数；k表示灰度值，o(i,j)＝k表示窗口中(i,j)位置的像素灰度值为k，h i(k)表示窗口中第i列中灰度值为k的像素的个数；
    4-2)根据窗口中每一列的统计直方图计算累积直方图：
    Figure PCTCN2020088056-appb-100006
    其中，l为灰度值，H i(l)表示窗口中第i列中灰度值小于等于l的像素的个数；
    4-3)对窗口中每一列的累积直方图求逆，得到
    Figure PCTCN2020088056-appb-100007
    4-4)对窗口内每一列的
    Figure PCTCN2020088056-appb-100008
    进行高斯加权，得到窗口内中心列的中值直方图：
    Figure PCTCN2020088056-appb-100009
    其中，A为窗口的列数，n为窗口中其他列到中心列的距离，g(n)为高斯权重函数：
    Figure PCTCN2020088056-appb-100010
    σ 2为窗口的方差；
    4-5)窗口内中心列的像素校正后的灰度值为：
    Figure PCTCN2020088056-appb-100011
  4. 根据权利要求1所述的基于局部中值直方图的自适应红外图像去条纹算法，其特征在于：步骤5中，窗口内中心列的像素原始灰度值均值
    Figure PCTCN2020088056-appb-100012
    为校正前窗口内中心列所有像素的灰度值之和与窗口内中心列像素个数的比值；窗口内中心列的像素校正后的灰度值均值
    Figure PCTCN2020088056-appb-100013
    为校正后窗口内中心列所有像素的灰度值之和与窗口内中心列像素个数的比值。
PCT/CN2020/088056 2019-05-06 2020-04-30 基于局部中值直方图的自适应红外图像去条纹算法 WO2020224518A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910372187.0A CN110211056B (zh) 2019-05-06 2019-05-06 基于局部中值直方图的自适应红外图像去条纹算法
CN201910372187.0 2019-05-06

Publications (1)

Publication Number Publication Date
WO2020224518A1 true WO2020224518A1 (zh) 2020-11-12

Family

ID=67785514

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/088056 WO2020224518A1 (zh) 2019-05-06 2020-04-30 基于局部中值直方图的自适应红外图像去条纹算法

Country Status (2)

Country Link
CN (1) CN110211056B (zh)
WO (1) WO2020224518A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112601068A (zh) * 2020-12-15 2021-04-02 济南浪潮高新科技投资发展有限公司 视频数据增广方法、装置及计算机可读介质
CN113793282A (zh) * 2021-09-17 2021-12-14 中国科学院长春光学精密机械与物理研究所 一种空间相机传函测试图像莫尔条纹模糊去除方法
CN114235149A (zh) * 2021-11-30 2022-03-25 中国科学院合肥物质科学研究院 一种基于ccd反射成像法的激光测量系统及其方法
CN114742732A (zh) * 2022-04-19 2022-07-12 武汉博宇光电系统有限责任公司 一种基于细节丰富度的红外图像增强方法

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110211056B (zh) * 2019-05-06 2022-08-12 南京理工大学 基于局部中值直方图的自适应红外图像去条纹算法
CN110796621B (zh) * 2019-10-29 2022-06-07 浙江大华技术股份有限公司 红外图像去横纹处理方法、处理设备和存储装置
CN112837250B (zh) * 2021-01-27 2023-03-10 武汉华中数控股份有限公司 一种基于广义直方图均衡的红外图像自适应增强方法
CN113610733B (zh) * 2021-08-10 2024-04-05 国网浙江省电力有限公司电力科学研究院 图像处理方法及装置
CN115375545B (zh) * 2022-08-22 2023-08-18 杭州微影软件有限公司 一种图像校正方法及装置
CN117147631A (zh) * 2023-10-30 2023-12-01 南京沃唐光电科技有限公司 一种基于区域环境分析的气体智能检测预警方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103076096A (zh) * 2013-01-07 2013-05-01 南京理工大学 基于中值直方图均衡的红外非均匀性校正算法
CN108830808A (zh) * 2018-06-01 2018-11-16 北京航空航天大学 基于相似线窗口均值补偿的星上红外图像条纹噪声去除方法
CN110211056A (zh) * 2019-05-06 2019-09-06 南京理工大学 基于局部中值直方图的自适应红外图像去条纹算法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102413283B (zh) * 2011-10-25 2013-08-14 广州飒特红外股份有限公司 红外热图数字信号处理系统及方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103076096A (zh) * 2013-01-07 2013-05-01 南京理工大学 基于中值直方图均衡的红外非均匀性校正算法
CN108830808A (zh) * 2018-06-01 2018-11-16 北京航空航天大学 基于相似线窗口均值补偿的星上红外图像条纹噪声去除方法
CN110211056A (zh) * 2019-05-06 2019-09-06 南京理工大学 基于局部中值直方图的自适应红外图像去条纹算法

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112601068A (zh) * 2020-12-15 2021-04-02 济南浪潮高新科技投资发展有限公司 视频数据增广方法、装置及计算机可读介质
CN113793282A (zh) * 2021-09-17 2021-12-14 中国科学院长春光学精密机械与物理研究所 一种空间相机传函测试图像莫尔条纹模糊去除方法
CN113793282B (zh) * 2021-09-17 2023-10-24 中国科学院长春光学精密机械与物理研究所 一种空间相机传函测试图像莫尔条纹模糊去除方法
CN114235149A (zh) * 2021-11-30 2022-03-25 中国科学院合肥物质科学研究院 一种基于ccd反射成像法的激光测量系统及其方法
CN114235149B (zh) * 2021-11-30 2024-02-27 中国科学院合肥物质科学研究院 一种基于ccd反射成像法的激光测量系统及其方法
CN114742732A (zh) * 2022-04-19 2022-07-12 武汉博宇光电系统有限责任公司 一种基于细节丰富度的红外图像增强方法
CN114742732B (zh) * 2022-04-19 2024-05-28 武汉博宇光电系统有限责任公司 一种基于细节丰富度的红外图像增强方法

Also Published As

Publication number Publication date
CN110211056A (zh) 2019-09-06
CN110211056B (zh) 2022-08-12

Similar Documents

Publication Publication Date Title
WO2020224518A1 (zh) 基于局部中值直方图的自适应红外图像去条纹算法
Cao et al. Effective strip noise removal for low-textured infrared images based on 1-D guided filtering
Wang et al. Enhancing low light videos by exploring high sensitivity camera noise
CN109741267B (zh) 基于三边滤波和神经网络的红外图像非均匀性校正方法
CN111986120A (zh) 一种基于帧累加和多尺度Retinex的低光照图像增强优化方法
CN111292257B (zh) 一种基于Retinex的暗视觉环境下图像增强方法
CN110246105B (zh) 一种基于实际相机噪声建模的视频去噪方法
Tendero et al. ADMIRE: a locally adaptive single-image, non-uniformity correction and denoising algorithm: application to uncooled IR camera
WO2017047494A1 (ja) 画像処理装置
CN106803235B (zh) 基于各向异性时空域全变分非均匀性校正的方法
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
CN102968776A (zh) 一种线性与非线性滤波器相结合的非均匀性校正方法
CN110782403A (zh) 一种红外图像非均匀性校正方法
Lu Stripe non-uniformity correction of infrared images using parameter estimation
Rong et al. Infrared fix pattern noise reduction method based on shearlet transform
CN113432723B (zh) 用于弱化杂散辐射的图像处理方法、系统及计算机系统
Li et al. Non-uniformity correction of infrared images based on improved CNN with long-short connections
CN114519676A (zh) 基于Bayer格式的raw图像的去噪装置及方法
CN114627011A (zh) 改进双边滤波与多帧平均联合的红外序列图像降噪方法
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
Zeng et al. High dynamic range infrared image compression and denoising
CN111932478A (zh) 一种非制冷红外焦平面的自适应非均匀校正方法
Lee et al. Dual-branch structured de-striping convolution network using parametric noise model
CN105005967A (zh) 时空滤波相结合的红外成像非均匀性校正方法及其装置
CN107610072A (zh) 一种基于梯度引导滤波的微光视频图像自适应降噪方法

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: 20802383

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20802383

Country of ref document: EP

Kind code of ref document: A1