WO2020107321A1 - 一种基于Retinex的微光图像增强方法及其装置 - Google Patents
一种基于Retinex的微光图像增强方法及其装置 Download PDFInfo
- Publication number
- WO2020107321A1 WO2020107321A1 PCT/CN2018/118174 CN2018118174W WO2020107321A1 WO 2020107321 A1 WO2020107321 A1 WO 2020107321A1 CN 2018118174 W CN2018118174 W CN 2018118174W WO 2020107321 A1 WO2020107321 A1 WO 2020107321A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- retinex
- minimum
- pixel value
- brightness adjustment
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 238000004364 calculation method Methods 0.000 claims abstract description 9
- 230000003044 adaptive effect Effects 0.000 claims description 13
- 230000000694 effects Effects 0.000 abstract description 6
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
Definitions
- the invention relates to the technical field of image processing, and in particular to a low-light image enhancement method and device based on Retinex.
- the Retinex enhancement method uses Gaussian smoothing function to estimate the brightness component of the original image, and uses the illumination compensation method to approximate the reflected image, which can enhance the information in the dark area of the image while maintaining the brightness of the image.
- Retinex has sharpening.
- the advantages of large dynamic range compression, constant color, and high color fidelity include the single-scale Retinex and multi-scale Retinex algorithms. Among them, the multi-scale Retinex algorithm works well, but the calculation amount is large and the processing efficiency is low.
- the technical problem to be solved by the present invention is to provide a low-light image enhancement method and device, which can improve the enhancement effect of darker areas of the image, improve the image quality, reduce the amount of calculation, and effectively improve the processing efficiency.
- a low-light image enhancement method based on Retinex includes the following steps:
- Multiscale Retinex processing is performed on the minimum pixel value map to obtain the Retinex processed image
- the minimum channel table, and the principle that the proportion of each component of each pixel is consistent with the proportion in the original color image calculate the other two channels in the original image except for the minimum pixel value After the value is obtained, the enhanced color image is obtained.
- the method further includes: performing adaptive brightness adjustment on the image processed by Retinex.
- the adaptive brightness adjustment of the Retinex processed image includes:
- pixels corresponding to the larger pixel value in the minimum pixel value map exceeding the second threshold are adjusted in brightness to obtain an image after brightness adjustment; otherwise, brightness adjustment is not performed.
- the brightness adjustment of the Retinex processed image is performed using the following formula:
- I_corret is the image after brightness adjustment
- I_retinex is the image after Retinex processing
- a is the brightness adjustment parameter.
- the calculated values of the other two channels except the minimum pixel value in each pixel point are calculated using the following formula:
- R', G', B' are the three-channel pixel values of the original color image
- R, G, B are the three-channel pixel values of the enhanced color image, respectively.
- a low-light image enhancement device based on Retinex includes:
- the minimum pixel value map acquisition module selects the minimum value of each pixel in the R, G, and B channels from the original image, and saves the channel where the minimum value is located, to obtain the minimum pixel value map and the minimum channel table;
- the Retinex processing module performs multi-scale Retinex processing on the minimum pixel value map to obtain the Retinex processed image
- Enhanced image calculation module calculates each pixel of the original color image except the minimum pixel value
- the processed values of the other two channels result in an enhanced color image.
- the color space conversion module is used to convert the original image to the R, G, B color space when the original image is not the R, G, B color space.
- the device further includes an adaptive brightness adjustment module, which is used to perform adaptive brightness adjustment on the image processed by Retinex.
- an adaptive brightness adjustment module which is used to perform adaptive brightness adjustment on the image processed by Retinex.
- the adaptive brightness adjustment module includes:
- the judging unit is used to judge whether the proportion of smaller pixel values in the minimum pixel value map is greater than the predetermined first threshold according to the statistical result of the statistical unit, and whether the number of larger pixel values in the minimum pixel value map is greater than the predetermined The second threshold; if it is, it is determined that there is a highlighted area, and the judgment result is passed to the adjustment unit; otherwise, no brightness adjustment is performed;
- the adjusting unit is configured to perform brightness adjustment on pixels corresponding to larger pixel values in the minimum pixel value map in the Retinex-processed image corresponding to the larger pixel value in the Retinex-processed image according to the determination result of the determination unit to obtain a brightness-adjusted image.
- the method and device provided by the present invention utilize the characteristics of the Retinex algorithm that has a better effect on dark areas.
- multi-scale Retinex enhancement processing By performing multi-scale Retinex enhancement processing on the minimum pixel value map, the enhancement effect of darker areas of the image can be improved and the image quality can be improved; Do multi-scale Retinex enhancement processing instead of doing three multi-scale Retinex enhancement processing on the three channels of R, G, and B, reducing the amount of calculation by two thirds, effectively improving the processing efficiency.
- the over-enhancement phenomenon caused by the Retinex processing of the highlighted area is solved to improve the image quality.
- FIG. 1 is a flowchart of a low-light image enhancement method provided by an embodiment of the present invention
- FIG. 2 is a flowchart of a low-light image enhancement method provided by a preferred embodiment of the present invention
- FIG. 3 is a schematic structural diagram of a low-light image enhancement device according to a preferred embodiment of the present invention.
- FIG. 1 is a flowchart of a low-light image enhancement method provided by an embodiment of the present invention
- the color images processed in the embodiment of the present invention are R, G, and B color spaces. If the images are in other spaces, they need to be converted into R, G, and B spaces.
- the minimum pixel value refers to the minimum value of a pixel in the three channels, which is defined as:
- I_min min(R,G,B)(1)
- R, G, and B are the three color channels of the image
- I_min is the minimum value of the pixels in the three color channels of R, G, and B
- the channel where the minimum pixel value is located is the minimum channel table.
- the processed image data I_Retinex can be calculated according to formula (2):
- * represents the convolution operation
- K represents the number of Gaussian scales
- Wk is the weighted primer corresponding to the kth scale
- K generally takes 3
- Wk takes 0.3
- Gk represents the kth Gaussian function
- Ck is a Gaussian scale constant, which determines the estimation of the incident component (that is, the illumination part), which determines the final enhancement effect.
- Ck generally chooses three scales of small, medium, and large.
- ⁇ k is a normalization factor such that:
- the minimum channel table calculates three of each pixel of the original image
- the processed value of the other two channels except the minimum pixel value in the channel is the enhanced color image.
- FIG. 2 is a flowchart of a low-light image enhancement method provided by a preferred embodiment of the present invention. The method includes:
- step (b) According to the statistical results, determine whether the proportion of smaller pixel values in the minimum pixel value map is greater than the predetermined first threshold, and whether the number of larger pixel values in the minimum pixel value map is greater than the preset second Threshold, if yes, proceed to step (c) for brightness adjustment, otherwise no brightness adjustment;
- the first threshold is determined by experimental results, generally 1/3
- the second threshold is also determined by the experimental results, generally 30.
- a is the brightness correction parameter, and its optimal value is determined by the experimental results, generally 1.3.
- the minimum channel table and the principle that the proportion of each component of each pixel is consistent with the proportion of the original input color image, the three channels of each pixel of the original image are calculated.
- the processed value of the other two channels except the minimum pixel value is the enhanced color image.
- the calculated values of the other two channels except the minimum pixel value in each pixel are calculated using the following formula:
- R', G', B' are the three-channel pixel values of the original color image
- R, G, B are the three-channel pixel values of the enhanced color image, respectively.
- FIG. 3 is a schematic structural diagram of a low-light image enhancement device according to an embodiment of the present invention.
- the device includes:
- the minimum pixel value map acquisition module 20 selects the minimum value of each pixel in the three channels of R, G, and B from the original image, and saves the channel where the minimum value is located, to obtain the minimum pixel value map and the minimum channel table;
- the Retinex processing module 40 performs multi-scale Retinex processing on the minimum pixel value map to obtain the Retinex processed image
- the enhanced image calculation module 60 calculates the processed values of the other two channels of each pixel except the minimum pixel value according to the Retinex processed image, the minimum channel table and the proportion of each component of each pixel. Color image.
- the enhanced image calculation module calculates three pixels for each pixel of the original image according to the principle that the Retinex processed image, the minimum channel table, and the proportion of each component of each pixel are consistent with the ratio of the original input color image.
- the processed value of the other two channels except the minimum pixel value in each channel is the enhanced color image.
- the device further includes a color space conversion module 10, which is used to convert the original image into the R, G, and B color space when the original image is not in the R, G, and B color space.
- a color space conversion module 10 which is used to convert the original image into the R, G, and B color space when the original image is not in the R, G, and B color space.
- the device further includes an adaptive brightness adjustment module 50, which is used to perform adaptive brightness adjustment on the image processed by Retinex.
- the adaptive brightness adjustment module includes:
- the statistical unit 501 is used to count the histogram of the minimum pixel value map
- the judging unit 502 is used to judge whether the proportion of smaller pixel values in the minimum pixel value map is greater than a predetermined first threshold according to the statistical result of the statistical unit, and whether the number of larger pixel values in the minimum pixel value map is greater than the predetermined The second threshold of; if it is, it is determined that there is a highlight area, and the judgment result is passed to the adjustment unit; otherwise, the brightness adjustment is not performed;
- the adjusting unit 503 is configured to perform brightness adjustment on pixels corresponding to the larger pixel value in the minimum pixel value map in the minimum pixel value image in the Retinex-processed image according to the judgment result of the judgment unit.
- the method and device provided by the present invention utilize the characteristics of the Retinex algorithm to enhance the dark area better.
- multi-scale Retinex enhancement processing on the minimum pixel value map, the image enhancement effect can be enhanced; Instead of doing three multi-scale Retinex enhancement processing on the three channels of R, G and B, the amount of calculation is reduced by nearly two-thirds and the processing efficiency is improved.
- the over-enhancement phenomenon caused by the Retinex processing of the highlighted area is solved to improve the image quality.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
一种基于Retinex的微光图像增强方法和装置，属于图像处理技术领域。该方法包括：从原始彩色图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小像素值所在的通道，得到最小像素值图及最小通道表（S102）；对最小像素值图进行多尺度Retinex处理，得到Retinex处理后的图像(S104)；根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例，计算原始图像每个像素点中除最小像素值外的其它两个通道处理后的值，即得到了增强后的彩色图像(S106)。本方法能够提高图像较暗区域的增强效果，提高了图像质量；同时大大减少了运算量，提高处理效率。
Description
本发明涉及图像处理技术领域，尤其涉及一种基于Retinex的微光图像增强方法和装置。
Retinex增强方法使用高斯平滑函数估计原图像的亮度分量，应用光照补偿方法逼近反射图像，能在保持图像亮度的同时增强图像暗处的信息，与其他的图像增强算法相比，Retinex具有锐化、动态范围压缩大、颜色恒常、色彩保真度高等优点，其中包括单尺度Retinex、多尺度Retinex算法，其中，多尺度Retinex算法效果较好，但运算量大，处理效率低。针对此问题，许多学者提出了只在亮度空间做多尺度Retinex处理、颜色分量保持不变的改进算法，主要包括HSI空间、HSV空间、YUV空间增强，其中HSI空间、HSV空间的多尺度Retinex增强结果会出现颜色过亮现象，YUV空间的多尺度Retinex增强结果颜色偏灰。
发明内容
本发明要解决的技术问题是提供一种微光图像增强方法和装置，以能够提高图像较暗区域的增强效果，提高图像质量，同时减少运算量，有效提高处理效率。
本发明解决上述技术问题所采用的技术方案如下：
根据本发明的一个方面，提供的一种基于Retinex的微光图像增强方法包括以下步骤：
从原始彩色图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小值所在的通道，得到最小像素值图及最小通道表；
当所述原始彩色图像不是R、G、B色彩空间时，将所述原始彩色图像转换为R、G、B色彩空间；
对最小像素值图进行多尺度Retinex处理，得到Retinex处理后的图像；
根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例 与原始彩色图像中比例保持一致的原则，计算原始图像每个像素点中除最小像素值外的另外两个通道处理后的值，得到增强后的彩色图像。
优选地，该方法还包括：对Retinex处理后的图像进行自适应亮度调整。
优选地，对Retinex处理后的图像进行自适应亮度调整包括：
统计最小像素值图的直方图；
如果最小像素值图中较小像素值所占的比例大于预定的第一阈值，且最小像素值图中较大像素值的个数大于预定的第二阈值，则判定存在高亮区域，对Retinex处理后的图像中对应最小像素值图中较大像素值超过第二阈值的像素进行亮度调整，得到亮度调整后的图像；否则不进行亮度调整。
优选地，对Retinex处理后的图像进行亮度调整采用下述公式进行：
I_correct＝I_retinex(a·I_retinex+a)
其中，I_corret为亮度调整后的图像，I_retinex为Retinex处理后的图像，a是亮度调整参数。
优选地，计算每个像素点中除最小像素值外的其它两个通道处理后的值采用以下公式进行：
<math><mrow><mfrac><msup><mi>R</mi><mo>&prime；</mo></msup><mi>R</mi></mfrac><mo>＝</mo><mfrac><msup><mi>G</mi><mo>&prime；</mo></msup><mi>G</mi></mfrac><mo>＝</mo><mfrac><msup><mi>B</mi><mo>&prime；</mo></msup><mi>B</mi></mfrac></mrow></math>
其中，R'、G'、B'分别是原始彩色图像的三通道像素值，R、G、B分别是增强后的彩色图像的三通道像素值。
根据本发明的另一个方面，提供的一种基于Retinex的微光图像增强装置包括：
最小像素值图获取模块，从原始图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小值所在的通道，得到最小像素值图及最小通道表；
Retinex处理模块，对最小像素值图进行多尺度Retinex处理，得到Retinex 处理后的图像；
增强图像计算模块，根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例与原始彩色图像中比例保持一致的原则，计算原始彩色图像每个像素点除最小像素值外的其它两个通道处理后的值，得到增强后的彩色图像。
色彩空间转换模块，用于当原始图像不是R、G、B色彩空间时，将原始图像转换为R、G、B色彩空间。
优选地，该述装置还包括自适应亮度调整模块，用于对Retinex处理后的图像进行自适应亮度调整。
优选地，自适应亮度调整模块包括：
统计单元，用于统计最小像素值图的直方图；
判断单元，用于根据统计单元的统计结果判断最小像素值图中较小像素值所占的比例是否大于预定的第一阈值，且最小像素值图中较大像素值的个数是否大于预定的第二阈值；如果是，则判定存在高亮区域，并将判断结果传给调整单元；否则不进行亮度调整；
调整单元，用于根据所述判断单元的判断结果对Retinex处理后的图像中对应最小像素值图中较大像素值超过第二阈值的像素进行亮度调整，得到亮度调整后的图像。
本发明提供的方法和装置，利用Retinex算法对暗区域效果较好的特性，通过在最小像素值图上做多尺度Retinex增强处理，能够提高图像较暗区域的增强效果，提高图像质量；同时，做一次多尺度Retinex增强处理替代在R、G、B三个通道上做三次多尺度Retinex增强处理，运算量减少了三分之二，有效提高了处理效率。此外，还通过对Retinex增强后的图像进行自适应亮度调整，解决高亮区域经过Retinex处理后易造成的过增强现象，提高图像质量。
为了更清楚地说明本发明实施例的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下， 还可以根据这些附图获得其他的附图。
图1为本发明实施例提供的一种微光图像增强方法流程图；
图2为本发明优选实施例提供的一种微光图像增强方法流程图；
图3为本发明优选实施例提供的一种微光图像增强装置结构示意图。
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。
以下结合附图对本发明进行详细描述。
如图1为本发明实施例提供的一种微光图像增强方法流程图；
S102、从原始彩色图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小像素值所在的通道，得到最小像素值图及最小通道表；
需要说明的是，本发明实施例处理的彩色图像是R、G、B色彩空间，若为其它空间图像，则需要先其转换成R、G、B空间。
具体来说，最小像素值是指某一像素点在三个通道中的最小值，定义为：
I_min＝min(R,G,B)(1)
其中，R、G、B即为图像的三个颜色通道，I_min是像素点在R、G、B三个颜色通道的最小值，该最小像素值所在的通道即为最小通道表。
S104、对最小像素值图进行多尺度Retinex处理，得到Retinex处理后的图像；
具体来说，可以根据公式(2)计算得到处理后的图像数据I_Retinex：
<math><mrow><mi>I</mi><mo>_</mo><mi>r</mi><mi>e</mi><mi>t</mi><mi>i</mi><mi>n</mi><mi>e</mi><mi>x</mi><mo>＝</mo><mi>exp</mi><mo>(</mo><mrow><munderover><mo>&Sigma；</mo><mrow><mi>k</mi><mo>＝</mo><mn>1</mn></mrow><mrow><mi>k</mi><mo>＝</mo><mi>K</mi></mrow></munderover><msub><mi>W</mi><mi>k</mi></msub><mrow><mo>(</mo><mrow><mi>log</mi><mi></mi><mi>I</mi><mo>_</mo><mi>m</mi><mi>i</mi><mi>n</mi><mo>-</mo> <mi>l</mi><mi>o</mi><mi>g</mi><mrow><mo>(</mo><mrow><msub><mi>G</mi><mi>k</mi></msub><mo>*</mo><mi>I</mi><mo>_</mo><mi>m</mi><mi>i</mi><mi>n</mi></mrow><mo>)</mo></mrow></mrow><mo>)</mo></mrow></mrow><mo>)</mo><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></math>
其中，*表示卷积操作，K表示高斯尺度个数，Wk是对应第k个尺度的权重引子，本实施例中一般K取3，Wk取0.3，Gk表示第k个高斯函数，其二维表达式为：
<math><mrow><msub><mi>G</mi><mi>k</mi></msub><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow><mo>＝</mo><msub><mi>&lambda；</mi><mi>k</mi></msub><mo>&CenterDot；</mo><msup><mi>e</mi><mrow><mo>-</mo><mfrac><mrow><msup><mi>x</mi><mn>2</mn></msup><mo>+</mo><msup><mi>y</mi><mn>2</mn></msup></mrow><mrow><msup><msub><mi>c</mi><mi>k</mi></msub><mn>2</mn></msup></mrow></mfrac></mrow></msup><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></math>
其中，Ck是高斯尺度常量，决定了对入射分量(即光照部分)的估计，即决定了最终的增强效果，Ck一般选择小、中、大三个尺度。λk是归一化因子，使得：
∫∫Gk(x,y)dxdy＝1(4)
S106、根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例，计算原始图像每个像素点中除最小像素值外的另外两个通道处理后的值，得到增强后的彩色图像；
具体来说，本步骤中根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例与原始输入彩色图像中的比例保持一致的原则，计算原始图像每个像素点的三个通道中除最小像素值外的其它两个通道处理后的值，即 得到了增强后的彩色图像。
如图2为本发明优选实施例提供的一种微光图像增强方法流程图，该方法包括：
S202、计算原始彩色图像的最小像素值图和最小通道表；
S204、对最小像素值图数据进行多尺度Retinex处理；
S206、对Retinex处理后的图像进行自适应亮度调整；
当图像中过暗区域较多时，高亮区域经过Retinex处理后易造成过增强现象，因此，需要对高亮区域进行亮度调整，得到校正后的图像。优选的步骤具体如下：
(a)统计最小像素值图的直方图；
(b)根据统计结果判断是否满足最小像素值图中较小像素值所占的比例是否大于预定的第一阈值，且最小像素值图中较大像素值的个数是否大于预设的第二阈值，如果是，则进步骤(c)进行亮度调整，否则不进行亮度调整；
具体来说，当最小像素值图中较小像素在整幅最小像素值图中所占的比例大于第一阈值(第一阈值由实验结果决定，一般取1/3)时，说明过暗区域较多，再判断最小像素值图中较大像素值的个数是否大于第二阈值(第二阈值同样由实验结果决定，一般取30)，如果是，说明最小像素值图中存在高亮区域，进入步骤(b)对高亮区域进行亮度调整，否则，不进行亮度调整。
(c)根据公式(5)对Retinex处理后的图像中对应最小像素值图中较大像素值超过第二阈值的像素进行亮度调整，得到亮度调整后的图像I_correct：
I_correct＝I_retinex(a·I_retinex+a)(5)
其中，a是亮度校正参数，由实验结果决定其最佳取值，一般取1.3。
S208、计算增强后的图像。
具体来说，根据亮度调整后的图像、最小通道表以及每个像素点各分量所占比例与原始输入彩色图像中的比例保持一致的原则，计算原始图像每个像素点的三个通道中除最小像素值外的其它两个通道处理后的值，即得到了增强后的彩色图像。
优选地，计算每个像素点中除最小像素值外的其它两个通道处理后的值采 用以下公式进行：
<math><mrow><mfrac><msup><mi>R</mi><mo>&prime；</mo></msup><mi>R</mi></mfrac><mo>＝</mo><mfrac><msup><mi>G</mi><mo>&prime；</mo></msup><mi>G</mi></mfrac><mo>＝</mo><mfrac><msup><mi>B</mi><mo>&prime；</mo></msup><mi>B</mi></mfrac><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>6</mn><mo>)</mo></mrow></mrow></math>
其中，R'、G'、B'分别是原始彩色图像的三通道像素值，R、G、B分别是增强后的彩色图像的三通道像素值。
如图3为本发明实施例提供的一种微光图像增强装置结构示意图，该装置包括：
最小像素值图获取模块20，从原始图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小值所在的通道，得到最小像素值图及最小通道表；
Retinex处理模块40，对最小像素值图进行多尺度Retinex处理，得到Retinex处理后的图像；
增强图像计算模块60，根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例，计算每个像素点除最小像素值外的其它两个通道处理后的值，得到增强后的彩色图像。
具体来说，增强图像计算模块根据Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例与原始输入彩色图像中的比例保持一致的原则，计算原始图像每个像素点的三个通道中除最小像素值外的其它两个通道处理后的值，即得到增强后的彩色图像。
作为本装置的一种优选方案，上述装置还包括色彩空间转换模块10，用于当原始图像不是R、G、B色彩空间时，将原始图像转换为R、G、B色彩空间。
作为本装置的一种优选方案，上述装置还包括自适应亮度调整模块50，用于对Retinex处理后的图像进行自适应亮度调整。自适应亮度调整模块包括：
统计单元501，用于统计最小像素值图的直方图；
判断单元502，用于根据统计单元的统计结果判断最小像素值图中较小像素值所占的比例是否大于预定的第一阈值，且最小像素值图中较大像素值的个数是否大于预定的第二阈值；如果是，则判定存在高亮区域，并将判断结果传给调整单元；否则不进行亮度调整；
调整单元503，用于根据所述判断单元的判断结果对Retinex处理后的图像中对应最小像素值图中较大像素值超过第二阈值的像素进行亮度调整。
本发明提供的方法和装置，利用Retinex算法对暗区域增强效果较好的特性，通过在最小像素值图上做多尺度Retinex增强处理，能够提高图像增强效果；同时，做一次多尺度Retinex增强处理替代在R、G、B三个通道上做三次多尺度Retinex增强处理，运算量减少了近三分之二，提高了处理效率。此外，还通过对Retinex增强后的图像进行自适应亮度调整，解决高亮区域经过Retinex处理后易造成的过增强现象，提高图像质量。
以上参照附图说明了本发明的优选实施例，并非因此局限本发明的权利范围。本领域技术人员不脱离本发明的范围和实质，可以有多种变型方案实现本发明，比如作为一个实施例的特征可用于另一实施例而得到又一实施例。凡在运用本发明的技术构思之内所作的任何修改、等同替换和改进，均应在本发明的权利范围之内。
Claims (4)
- 一种基于Retinex的微光图像增强方法，其特征在于，包括：从原始彩色图像中选取每一个像素点在R、G、B三通道中的最小值，并保存该最小值所在的通道，得到最小像素值图及最小通道表；当所述原始彩色图像不是R、G、B色彩空间时，将所述原始彩色图像转换为R、G、B色彩空间；对所述最小像素值图进行多尺度Retinex处理，得到Retinex处理后的图像；依据所述Retinex处理后的图像、最小通道表以及每个像素点各分量所占比例与原始彩色图像中比例保持一致的原则，计算原始彩色图像每个像素点中除最小像素值外的另外两个通道处理后的值，得到增强后的彩色图像。
- 根据权利要求1所述的微光图像增强方法，其特征在于，所述方法还包括：对所述Retinex处理后的图像进行自适应亮度调整，对所述Retinex处理后的图像进行自适应亮度调整包括：统计最小像素值图的直方图；如果最小像素值图中较小像素值所占的比例大于预定的第一阈值，且最小像素值图中较大像素值的个数大于预定的第二阈值，则判定存在高亮区域，对Retinex处理后的图像中对应最小像素值图中较大像素值超过第二阈值的像素进行亮度调整，得到亮度调整后的图像；否则不进行亮度调整。
- 根据权利要求2所述的微光图像增强方法，其特征在于，所述对Retinex处理后的图像进行亮度调整采用下述公式进行：I_correct＝I_retinex(a·I_retinex+a)其中，I_corret为亮度调整后的图像，I_retinex为Retinex处理后的图像，a是亮度调整参数。
- 根据权利要求3所述的微光图像增强方法，其特征在于，所述计算每个像素点中除最小像素值外的其它两个通道处理后的值采用以下公式进行：<math><mrow><mfrac><msup><mi>R</mi><mo>&prime；</mo></msup><mi>R</mi></mfrac><mo>＝</mo><mfrac><msup><mi>G</mi><mo>&prime；</mo></msup><mi>G</mi></mfrac><mo>＝</mo><mfrac><msup><mi>B</mi><mo>&prime；</mo></msup><mi>B</mi></mfrac>
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2018/118174 WO2020107321A1 (zh) | 2018-11-29 | 2018-11-29 | 一种基于Retinex的微光图像增强方法及其装置 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2018/118174 WO2020107321A1 (zh) | 2018-11-29 | 2018-11-29 | 一种基于Retinex的微光图像增强方法及其装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020107321A1 true WO2020107321A1 (zh) | 2020-06-04 |
Family
ID=70852622
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2018/118174 WO2020107321A1 (zh) | 2018-11-29 | 2018-11-29 | 一种基于Retinex的微光图像增强方法及其装置 |
Country Status (1)
| Country | Link |
|---|---|
| WO (1) | WO2020107321A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112488945A (zh) * | 2020-12-04 | 2021-03-12 | 珠海普生医疗科技有限公司 | 一种基于局部收敛的内窥镜图像增强方法 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103295206B (zh) * | 2013-06-25 | 2016-05-25 | 安科智慧城市技术（中国）有限公司 | 一种基于Retinex的微光图像增强方法和装置 |
| CN106997584A (zh) * | 2017-04-12 | 2017-08-01 | 哈尔滨理工大学 | 一种雾霾天气图像增强方法 |
| CN107067018A (zh) * | 2016-12-09 | 2017-08-18 | 南京理工大学 | 一种基于随机Hough变换和SVM的带电作业机器人螺栓识别方法 |
| CN107705268A (zh) * | 2017-10-20 | 2018-02-16 | 天津工业大学 | 一种基于改进的Retinex与Welsh近红外图像增强与彩色化算法 |
-
2018
- 2018-11-29 WO PCT/CN2018/118174 patent/WO2020107321A1/zh active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103295206B (zh) * | 2013-06-25 | 2016-05-25 | 安科智慧城市技术（中国）有限公司 | 一种基于Retinex的微光图像增强方法和装置 |
| CN107067018A (zh) * | 2016-12-09 | 2017-08-18 | 南京理工大学 | 一种基于随机Hough变换和SVM的带电作业机器人螺栓识别方法 |
| CN106997584A (zh) * | 2017-04-12 | 2017-08-01 | 哈尔滨理工大学 | 一种雾霾天气图像增强方法 |
| CN107705268A (zh) * | 2017-10-20 | 2018-02-16 | 天津工业大学 | 一种基于改进的Retinex与Welsh近红外图像增强与彩色化算法 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112488945A (zh) * | 2020-12-04 | 2021-03-12 | 珠海普生医疗科技有限公司 | 一种基于局部收敛的内窥镜图像增强方法 |
| CN112488945B (zh) * | 2020-12-04 | 2024-04-26 | 珠海普生医疗科技有限公司 | 一种基于局部收敛的内窥镜图像增强方法 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11127122B2 (en) | Image enhancement method and system | |
| CN104240194B (zh) | 一种基于抛物线函数的低照度图像增强算法 | |
| CN102789635B (zh) | 一种图像增强方法和装置 | |
| KR100621414B1 (ko) | 채도 적응적인 영상 향상장치 및 그 방법 | |
| CN103295206B (zh) | 一种基于Retinex的微光图像增强方法和装置 | |
| CN109146826A (zh) | 一种图像增强方法及装置 | |
| KR20040044556A (ko) | 화상 처리 방법, 장치 및 프로그램 | |
| US8406559B2 (en) | Method and system for enhancing image sharpness based on local features of image | |
| US20200320672A1 (en) | Improved inverse tone mapping method and corresponding device | |
| CN108133463B (zh) | 一种直方图均衡化图像的降噪方法及系统 | |
| CN108305232B (zh) | 一种单帧高动态范围图像生成方法 | |
| CN111340732B (zh) | 一种低照度视频图像增强方法及装置 | |
| CN107292834B (zh) | 红外图像细节增强方法 | |
| CN101478689B (zh) | 一种基于色域映射的图像光照校正系统 | |
| CN111968065A (zh) | 一种亮度不均匀图像的自适应增强方法 | |
| WO2020107308A1 (zh) | 一种基于Retinex的微光图像快速增强方法及其装置 | |
| CN112541869A (zh) | 一种基于matlab实现的Retinex图像去雾方法 | |
| WO2020107321A1 (zh) | 一种基于Retinex的微光图像增强方法及其装置 | |
| CN113450272B (zh) | 一种基于正弦曲线变化的图像增强方法及其应用 | |
| CN110111280A (zh) | 一种多尺度梯度域引导滤波的低照度图像增强算法 | |
| CN111028181B (zh) | 一种图像增强处理方法、装置、设备及存储介质 | |
| Yang et al. | Improved retinex image enhancement algorithm based on bilateral filtering | |
| CN104392423A (zh) | 基于实时均衡红外图像细节增强算法 | |
| WO2017104291A1 (ja) | 画像処理装置及び画像処理方法 | |
| CN110365949B (zh) | 一种图像偏色的校正方法、装置及电子设备 |
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: 18941674 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: 18941674 Country of ref document: EP Kind code of ref document: A1 |